ee nt celle Gir aif, Ae tattle totes Bb ast ag GS El tem & awh tod 6620088 Oa. 2 eens be! tenes S . ~- ww Vv Soran patteny Capo Soot en * Saropere $58 -> Jane eann en Paths mann Bee NA) ea Te Tk ae be eee eee 7S : Caen Se
Serer eae oss ‘ = aba ee : > SUTTER SS ee a me ES Spain etencs 5 Lika laren a Sores en TE ae ee 5 SLAC ene s ~ cC A * “ . . ” . *
ee
ene
Coen aan
nee 9 : re we oe ae ae
Sere SS ew weer eens Lnetcbenecereeech ional
sa oo 8 cae eh
Cie Ce rrr PIT er te
Se ee eee “ omni ~ s ee beer ewes ee Se 8 TR wy ee et ee ee a te tome . S Pee cerns ee A Sr 5 he ee rrr. ee ee LE ener)
ere ree ree rn ere ere rere rireuieai
rere ana
ey A me ah
ae erry
Aaah ee Ak Se hee NN Ae ae NNT TA A A A ee A AA RR eee ere ee et ee eer et er er ee es NN A EO OE ee ee eS nt ee eS eee eee rr eran ee ee eee
os ee Ae as A EN ene ee NENT NA mee Ane themn NA ee et om me mm ee ee ee, Ad pte meth eh A ee eet ee hin tee mm eee te Rech nt, mee ay th ee See eS Re ae ny lett Pc ee re eer ern fee erere ent ere kes
A teh tam mch ae a ~ = rs —. ——
a en cers
a ae Snes
eres : Sree = ———
ee
ee
THOMAS LINCOLN CASEY ~ LIBRARY 1925
or
% Sad pa GIN Ls
eA
Wd WdsAtn Ean ie i Me bait ae ey eae b, ahr Bau Pate i oll fing a sot cy in ae ; cin ) Ban 2) i iy SHATTER
pans Gay i
HO Noaa a
JOURNAL
OF
The Academy of Natural Sciences
OF
PHILADELPHIA.
VOLUME XII, SECOND SERIES.
PHILADELPHIA : PRINTED FOR THE ACADEMY. 1902-1904.
PUBLICATION COMMITTEE.
Henry Sxryner, M. D. Partie P. Catvert, Pu. D. Henry A. Pirssry, Sc. D. 3 WiITMER STONE.
Epwarp J. Nouan, M.D.
The President, SamurL G. Dixon, M. D., ex-officio.
Epitor, Epwarp J. Noxan, M.D.
1 ART: I.—Karyokinesis and Cytokinesis in the Maturation, Fertilization and Cleavage of Crepidula and other Gasteropoda. By Edwin G. Conklin, Ph.D.
2 ART. Il.—Certain Aboriginal Remains of the Northwest Florida Coast.
SARE
CONTENTS.
PARTY I.
PART II.
Clarence B. Moore
PART III.
. 11].—Certain Aboriginal Remains of the Florida Central West Coast.
B. Moore
. 1V.—Certain Aboriginal Mounds of the Apalachicola River.
PART IV.
. V.—A Collection of Fishes from Sumatra. By Henry W. Fowler.
XX VIL).
1 Extra copies printed for the author, 2 Extra copies printed for the author, 8 Extra copies printed for the author,
4 Extra copies printed for the author,
November 15, 1902.
October 22, 1902.
September 15, 1903. June 10, 1904.
(Plates I-VI)
Part IT. By
By Clarence
By Clarence B. Moore
(Plates VII-
123
495
“ Bit ey)
py ara nian
¥ alii
siffeett
ius
rt fhe 7 the
y
INDEX TO SPECIES, ETC.,
REFERRED TO OR DESCRIBED IN
VOLUME XII.
New Species, ete., are printed in Small Capitals, Synonyms in Italics.
Acanthocystis, 65.
Acanthurus, 544.
ACTINICOLA, 533
Actinospherium, 18, 37, 43, 56, 59, 65, 68, 69, 103, 112.
olis, 56, 59, 61. papillosa, 6.
JETHALOPERCA, 522.
ALEPES GLABRA, 507. melanoptera, 507. SCITULA, 509.
Allolobophora, 11, 22, 29.
Amphioxus, 28.
Amphiprion, 533. ephippium, 533. percula, 533.
Anabantide, 530.
Anabas scandens, 530.
Anchoyia commersonnii, 501. encrasicholoides, 501. yalenciennesi, 501.
ANEMURA, 527.
Anguilla bengalensis, 500. bicolor, 500.
Anguillide, 500.
Anthias johnii, 524.
Aphanostoma, 98.
APOGON EVANIDUS, 518. frenatus, 519. hyalosoma, 518. melanorhijnchos, 519. novemfasciatus, 519. vittiger, 519.
Apogonide, 518.
Apolectus niger, 513.
. Aprion typus, 527.
Arbacia, 98. Archamia bleekeri, 519. Arenicola, 25, 29, 54.
Ascaris, 20; 27, 35, 51, 52, 61, Ga GE), val Ves megalocephala, 104.
Aspisurus, 544
Asterias, 52, 66.
Axolotl, 20.
Balistapus undulatus, 547.
Balistide, 546.
Barbodes binotatus, 500. fasciatus, 500. schwanenfeldii, 500.
Belonide, 501.
BENNETTIA, 524.
Blenniide, 552.
Bodianus, 522. aurantius, 522. INDELEBILIS, 521. miniatus, 522.
Togaa, 522.
Branchipus, 20.
Butis gymnopomus, 550.
Cesio cerulaureus, 528. ERYTHROCHILURUS, 528. lunaris, 529.
Callonymide, 550.
Callionymus sagitta, 550.
Callista gigantea, 396.
| Campeloma lima, 481. | Canis occidentalis, 228.
Carangide, 506, 518. Carangoides oblongus, 513. GIBBER, 512. malabaricus, 512, 513. Caranx djedaba, 509. forstert, 510. kuhli, 510. mate, 510. macrurus, 509. | malam, 509. |
Caranx megalaspis, 510. sem, 512. SEMISOMNUs, 510.
Carcharhinus menisorrah, 499.
Cardium, 396.
Cerebratulus, 20, 29, 35, 36,
54, 66.
Cerithium 397.
Chetodon canaliculatus, 546. trifasciatus, 544. vagabundus, 544.
Cheetodontide, 544.
| Cheetopterus, 29, 62, 66, 78. | Cheilinus diagramma, 539.
enneacanthus, 539. trilobatus, 539.
| Chiloscyllium indicum, 499. | Chirocentride, 501.
Chirocentrus dorab, 501.
Chlarias batrachus, 499. liacanthus, 500. OLIVACEUS, 499.
Chlariidae, 499.
| Choerops schoenleinii, 535.
Chromis cinerascens, 533. Chrysiptera unimaculata, 535. modesta, 535. Cirrhimurzena chinensis, 501. Citula armata, 513. atropos, 513. Clarias fuscus, 499, 500. Clupeide, 501. Cobitidide, 500. Cottus seaber, 550.
| Crepidula, 10--113.
adunea, 6. conyexa, 6. fornicata, 6. plana, 6, 7, 26.
Ctenodon ctenodon, 545.
Ctenolabrus, 98.
Cucullanus, 50.
Cyclocheilichthys siaja, 500.
Cyclops, 35, 52.
CynoGLossus 0s, 556. sumatranus, 556.
Cyprinus carpio, 500.
Cyprinide, 500.
Dascyllus aruanus, 533. trimaculatus, 533.
Dasyatiide, 499.
Dasyatis russellii, 499.
Dentex blochii, 527. mesoprion, 527. notatus, 527. tolu, 527.
Denticine, 527.
DEVEXIMENTUM, 517. insidiator, 517.
Diagramma affine, 527.
Diaulula, 16, 17, 56, 57, 61, 69. |
Dosinia discus, 482.
Dussumieride, 501.
Dussumieria elopsoides, 501.
Dytiscus, 38.
Echeneidide, 552.
Echinus, 56, 61, 115.
Elagatis bipinnulatus, 507.
Eleotridine, 550.
Eleotris fusca, 550.
Engraulidide, 501.
EnToMACRODUS CALURUS, 555. LEOPARDUS, 554.
Ephippide, 544.
Ephippus argus, 544.
Epinephelus dermochirus, 524. | heniochus, 522, 524. | horridus, 524. lanceolatus, 524. maculatus, 524. megachir, 524. preopercularis, 524. | sexfasciatus, 524.
Equula blochi, 517. lineolata, 514. | splendens, 516.
‘Kteline, 527.
| Gerres filamentosus, 530. Gerride, 530. | Glycymeris americana, 228.
| Gobiide, 550.
| Gymnocranius lethrinoides,
| Halicheeres, 537.
Equutitss, 513. Erythrodon, 546.
INDEX.
FUBLEEKERIA, 516. EUELATICHTHYS, 527. EUTHERAPON, 527. EurHYOPTEROMA, 527. Evenchelys macrurus, 501. Eyoplites decemlineatus, 524. Fasciolaria, 394, 415, 417. Felis concolor, 228. Fulgur, 56, 78, 160, 220, 252. | 365, 396, 417, 446, 474.
carica, 6.
perversum, 198, 212, | 225, 258, 268, 394, 396, 415, 426, 486, | 44%. |
pugilis, 196, 198.
Galeocerdo tigrinus, 499. Galeorhinide, 499. Gazza minuta, 517. tapeinosoma, 517. Genyoroge, 526. Georgiana vivipara, 481. Germo germon, 506.
Glyphisodon leucogaster, 535. saxatilis, 535. septemfasciatus, 535.
Glyptothorax platygonides,
500. platypogon, 500.
Gobiine, 551.
Gobius sumatranus, 552. VENUSTULUS, 551.
GrRAMMOPLITES, 550. scaber, 550.
527.
Gymnothorax fimbriatus, 501. |
flavimarginatus, 501. | Hemulide, 527.
ANNULATUS, 535. euttatus, 537. ! hartzfeldiu, 527. leparensis, 537.
Halicheres miniatus, 535. nigriscens, 537. poecila, 537.
Haminea, 56, 61. solitaria, 6
Hampala macrolepidota, 500.
Harpochirus longimanus, 544.
Harpodon nehereus, 501.
| Harpurrpm, 544.
Harpurus, 544. GNOPHODES, 544 Helix, 70. Hemipteronotus baldwini, 541. LIOGENYS, 539. Hemiramphide, 501. Hemiramphus far, 501. Hemiscylliidae, 499. Heniochus macrolepidotus, 544. Hippocampus kamphylotrache- los, 502. kuda, 502. pasxtops, DOL. Holocenthridx, 504. Holocenthrus, 504. albo-ruber, 506. AUREORUBER, 504. melanospilos, 506. radjabau, 528. Hymenophysa hymenophysa, 500. Hyporhamphus neglectus, 501
Hypselobagrus micracanthus,
500. Tlarches orbis, 544. Tlarchidee, 544.
| Jlisha brachysoma, 501.
hoeyvenii, 501. Tllyonassa, 56, 78.
| Julis dorsale, 537.
(Halicheeres) harloffi, 537. miniatus, 530 semifasciatus, 539. (Julis) urostigma, 539. Kyphoside, 530. Kyphosus lembus, 530. Labeobarbus douronensis, 500. tambra, 500.
Labride, 535. Labrus vittatus, 550. Lactarip® 517. Lactarius, 518. lactarius, 518. Lagocephalus albo-plumbeus, 547. lunaris, 547. Leiognathide, 513. Leiognathus, 517. blochii, 517. bindoides, 516. edentulus, 517. sprLotus, 516. splendens, 516. VERMICULATUS, 513. VIRGATUS, 515 Leiurus stellaris, 546. Lepidaplois mesothorax, 535. LETHRINELLA, 529. Lethrinus, 529. miniatus, 529. opercularis, 529. ornatus, 529. Timax, 13, 22, 23; 56, 59, 60, 79, 104. Limnea, 29. Lobotes, 522. Lutianide, 522, 524. Lutianus, 524. argentimaculatus, 526. biguttatus, 524. cerulo-punctatus, 525. chirtah, 526. decussatus, 525. FURVICAUDATUS, 529. johnii, 525. lepisurus, 520. lineatus, 525. lioglossus, 525. lunulatus, 525. madras, 524. malabaricus, 527. nouleny, 524. percula, 533. roseus, 525. russellii, 525. seb 527. vaigiensis, 520.
INDEX.
Lutianus vitta, 524. Macrobdella, 64. Macrognathide, 501. Macropteronotus fuscus, 499. Malacanthide, 549. MALACANTHUS URICHTHYS, 549. Marginella, 446. apicina, 169, 190. Mastacembelus unicolor, 501. Megalaspis rottleri, 506. Melongena corona, 394. Mene maculata, 517. Menidae, 517. Monoceros, 544. Monopteridz, 500. Monopterus albus, 500. Mugilide, 502. Mullide, 530. Murenide, 501. Murex flavescens, 482. Myripristis murdjan, 504. Mystacoleucus padangensis, 500. Myzostomum, 20. Nandidx, 530. Nassa, 738. Nephelis, 50. Nereis, 52, 112. Netuma thalassina, 500. Noctiluca, 37, 65, 71. Nototheniidee, 548. OctocyNoDON, 535. ODONTOGLYPHIs, 527. Odontonectes erythrogaster, 528. Oliva reticularis, 397. Oonidus, 547. immaculatus, 547. reticularis, 547. Opalina, 69. Ophicephalidee, 530. Ophicephalus lucius, 531. pleurophthalmus, 531. polylepis, 531. SPIRITALIS, 530. urophthalmus, 531. Ophichthyide, 501. Ophiocara porocephala, 550.
|
Ophryotrocha, 9, 20. Opisthopterus macrognathus, 501.
| Osphronemide, 530.
Osphronemus goramy, 530. Osteochilus hasseltii, 500. kuhli, 500. Ostorhinchus, 518. Otolithus argenteus, 530. Pachynathus capistratum, 547. Paracanthistius maculatus, 521. Paralichthys polyspilus, triocellatus, 555. Parameeba, 69. Paramecium, 69. Parapegasus natans, 501. PARAPERCIS ATROMACULATA, 548. hexophtalma, 549. PArRKIA, 525. Pegaside, 501. Pentaprion longimanus, 517.
5950.
| Pentatoma, 12, 48, 104. | Perca rogaa, 522.
Pertica, 530. Petrometopon cyanostigma, 521. formosus, 521. pachycentron, 521. Pharopteryx corallicola, 530. Physa, 20, 21, 29, 43, 54, 104, 112. Platacide 544., Platax orbicularis, 544. vespertilio, 544 Platycephalidw, 550. Platycephalus indicus, 550. PLECTROPOMA PESSULIFERUM, 520. Plectorhinchus, 528. affinis, 527. radjabau, 528. sebae, 528. Pleuronectidw, 555. Pleurophylidia, 26, 29, 54, 104. Plotoside, 499. Plotosus anguillaris, 499. Polydactylus pfeitferi, 530,
Polycherus, 20, 51, 52, 98. Polynemidee, 530. Pomacentridee, 532. POMACENTRUS LEUCOSPHYRUS. 533. tripunctatus, 535. vanicolensis, 535. violescens, 535. Pomadasys commersonnil, 528. Premnas biaculeatus, 533. EPIGRAMMATA, 582. semicincta, 533. Priacanthidey, 524. Priacanthus tayenus, 524. Prostheceraeus, 13, 22. Psettodes erumei, 555. Pterois lunulata, 548. Pterotrachea, 29. Rachycentride, 513. Rachycentron pondicerianus, 513. Raizero, 524 Rana temporaria, 98 Rangia cuneata, 234, 281, 282, 482. Rasbora argyrotenia, 500. Rastrum, 509. Remora nieuhofii, 552. Rhinobatide, 499. Rhynchelmis, 17, 56, 59, 112. Rhynchobatus djiddensis, 499. Salamandra, 56, 70. Salarias oortu, 553. Salmo, 56. Sardinella brachyosoma, 501. Saurida tumbil, 501. ScAaRTICHTHYS BASILISOUS, 552. STIGMATOPTERUS, 553. Scaride, 541. Scarts CALLUS, 542. cantori, 544. quoyi, 544. PINGUIROSTRATUS, 541. rubroviolaceus, 544. Schismatorhynchus hetorhyn- chus, 500. Sciena macroptera, 530.
INDEX.
Scienide, 518, 530.
Scolopendra, 105.
Scolopsis bleekeri, 528.
Scomber kanagurta, 506. lactarius, 518.
Scomberoides tol, 506. toloo, 506.
Scomberomorus gutiatus, 506.
Scombride, 506. Scorpzenide, 548. Scorpeenopsis oxycephala, 548. Seyris indica, 513. Selar, 509, 510. Sebastopsis polylepis, 548. SERIOLA CRETATA, 506. purpurascens, 507. Serioloidei, 518. Serranide, 520. Serranus analis, 522. aurantius, 522. Sida, 56. SIGANITES,. 546. Sillaginide, 549. Sillago sihama, 549. Siluride, 500. Siredon, 35. Soleidae, 556. Sparide, 529. Sparus miniatus, 529. Sphaerechinus, 115. Sphyreena forsteri, 504. TOXEUMA, 502. toxeuma, 504. Sphyreenide, 502. Sphyrna zygena, 499. Sphyrnide, 499. SrrLoTicHtTiys, 528. Spirochona, 65.
Stethojulis phekodopleura, 535.
Stromateide, 513. Strombus gigas, 396. Sycotypus, 56, 78. eanaliculatus, 6.. Syngnathide, 501. Synodontide, 501. Teeniura lymma, 499,
Tetraodon palembangensis, 547.
leiurus, 547.
Tetraodontide, 547.
Teuthidide, 545.
Teuthis canaliculatus, 546.
corallinus, 546. javus, 545. vermiculatus, 546. virgatus, 546.
Toxopneustes, 35, 37, 48, 65. 54, 61, 66.
Thalassoma dorsale, 539.
hardwicke, 539. lunare, 537. MELANOCHIR, 537.
Theropon, 527.
jarbua, 527. theraps, 527.
Theraponide, 527.
Thysanozoon, 13, 15, 17, 19, 22, 56, 57. :
Toxopneustes, 35, 37, 48, 65.
Trachinotus ovatus, 513.
Triacanthide, 546.
Triacanthus brevirostris, 546.
nieuhofii, 546.
Trichiuridee, 506.
Trichiurus haumela, 506.
Trichopodus trichopterus, 530.
Triton, 20.
Turbellaria, 29.
Turriops turrio, 269.
Tylosurus crocodilus, 501.
3 leiurus, 501.
melanotus, 501.
Unio, 11, 14, 17, 19, 27, 29, 38, 54, 56, 58, 61, 69, 109, 112, 113.
Upeneoides moluccensis, 530.
sulfureus, 530. Upeneus malabaricus, 530. Urosalpinx, 56, 78.
cinerea, 6.
Variola louti, 521.
Venus cancellata, 433.
Zenodon, 546.
CmRULEOLORUM, 546.
niger,. 547.
Zeus insidiator, 517.
Zirpheea, 14.
KARYOKINESIS AND CYTOKINESIS
IN THE
Maturation, Fertilization and Cleavage
OF
CREPIDULA and other GASTEROPODA.
EDWIN G. CONKLIN, Pu.D.
PHILADELPHIA: 1902.
Ae ee ie
PART I. KARYOKINKESIS.
CONTENTS.
Introduction 5 : : eee ‘ : : ; : Methods : ‘ s 7 : E : F 2 : A
I. Marurarron. A. Predivision Stages. 5 : : : : : ‘ é 1. The Ovarian Ege : : : : : : Egg Laying. 5 ; : : é : . B. Maturation Divisions F : : : : : . 1. Nuclei : ; , : : : : : ; a. Chromatin 3 : : : Me ; j b. Nuclear Sap ‘ ; : , : : ; é . Nucleolus . ; : : : ‘ E ? . : a
‘. Chromosomes. : : : : : 0 $ . Centrosomes and Central Spindles ; . : : P : : 3. Polar Rays, Spindle Fibres and Spheres _. : : : : ;
4. Polar Bodies
II. Ferri.izarion. 1, Entrance of Spermatozoon : 2. The Germ Nuclei : : : 3. Egg and Sperm Asters and Spheres 4. Approach of Germ Nuclei and Spheres 5. Origin of Cleavage Centrosomes
III. Cieavace. 1. Nuclear Changes during Cleavage a. ait of Germ Nuclei. : . Chromatin “% te ution of Cinomazoms onl Boncarion of Date ntee Nuclei : Cet trosinaes and Central Spindles
bo
a. Centrosomes
b. Central Spindles ; BPN 3. Polar Rays and Spindle Fibres; Mid-Bodies 4. Spheres
TV. GENERAL CONSIDERATIONS AND COMPARISONS. 1. The Nucleus during the Cycle of Division : : é ; : a. Formation of Chromosomal Vesicles; Growth of Daughter Nuclei
b. Chromatie Differentiation ; Solution of Oxychromatin and Nuclear Membrane
c. Escape of Nuclear Substances; Aster and Spindle Formation d. Chromatic Elimination 2. Centrosomes and Central Spindles a. Structure and Metamorphoses b. Relation of Centrosome to Cell Body and Shien ec. Relation of Centrosome to Nucleus d. Comparisons e. The Centrosome as a P ersistent Cell Onan f. Homology of the Centrosome 3. Spheres, Idiozome, ete. 4. Archoplasm, Kinoplasm, ete.
wo 09 wD 0 OD OO
+ _
41
PART I. CYTOKINESIS.
CONTENTS. I. Srrucrure or CyrorpLAsM
Il. Movements oF Ceri Conrents A. During Maturation B. During Fertilization C. During Cleavage 1. First Cleavage 2. Second Cleavage . Third Cleavage 4. Fourth Cleavage : 5. Fifth and Sixth Cleavages 6. Subdivisions of First Quartette 7. Subdivisions of Second Quartette 8. Subdivision of Third Quartette
II]. AnAtysis oF MOVEMENTS DURING CELL DIviston. 1. Movements during Metakinesis : a. Movements in Spindle and Aster . b. Movements in the Cell Body Comparisons ; 3 . Movements during Telokinesis Comparisons : : : 3. Orientation of Centrosomes and Spindles a. Relative to Nucleus. 5 : : 3 b. Relative to Cell Body
bo
IV. Some Facrors oF DIFFERENTIATION. 1. Polarity ; a. Unsegmented Ege b. Blastomeres : : , c. Nucleus, Centrosome and Sphere 2. Differential Cell Division a. Rhythm of Division b. Direction of Division e. Size of Daughter Cells d. Quality of Daughter Cells
References. : : : P j ; ; 5 Explanation of Figures
SS SS RE SSE I a0 oO
Crnm C= Sco co
KARYOKINESIS AND CYTOKINESIS IN THE MATURATION, FERTILI- ZATION AND CLEAVAGE OF CREPIDULA AND OTHER GASTEROPODA.'
By Epwin G. Conxuin, Pu.D. 1;
INTRODUCTION.
Cell division, in a broad sense, includes not only nuclear division and the separa- tion of daughter cells, but also all the phenomena which lead up to these processes and which follow them; the terms Aavyokzneszs (Schleicher 78) and Cytokzneszs (Whitman °87) are used in this paper to include these nuclear and cytoplasmic activities of the entire cell-cycle from one division to the next. Flemming (’§2) has objected to the term Aaryokzneszs on the ground, among others, that nuclear move- ments are not characteristic of indirect division. But in view of the extensive movements of both nuclei and cytoplasm, which occur in the cell divisions deseribed in this paper, the terms Karyokzneszs and Cytokineszs have peculiar appropriateness.”
The phenomena of cell division still include an extraordinary number of dark problems, in spite of the fact that “all the search-lights of the biological sciences have been turned upon the cell.” Confusion and contradiction exist as to the nature and metamorphoses of the centrosome and central spindle, the origin and fate of the amphiaster, the characteristics and history of the attraction sphere, the existence or non-existence of a specific substance (Archoplasm, Kinoplasm, ete.) whose primary function is the division of the cell. Still less complete is our knowledge of the inter- relation of nucleus and cytoplasm during the various phases of division, of the phenomena and significance of the movements of cells and cell constituents and of the chemical, physical and physiological principles involved in the division of nucleus and cell body.
In the early development of the egg, cell divisions have a peculiar interest because of their bearings on problems of heredity and differentiation. Here are found phenomena of the most general occurrence and of the deepest significance, viz.: the maturation, fertilization and cleavage of the egg and the early differentiation of the embyro. The bearings of the phenomena of maturation and fertilization upon
‘ From the Zoological Laboratory of the University of Pennsylvania.
2 On the other hand Flemming’s term mitosis commends itself because of its brevity, and it is frequently employed throughout this paper.
1* JOURN. A. N. S. PHILA., VOL. XII.
6 KARYOKINESIS.
problems of heredity and differentiation are matters of common knowledge and need not be discussed here; the relation of cleavage to these larger problems is neither so generally recognized nor so freely admitted.
In the cleavage of the egg, differentiations occur to a remarkable degree in certain cell divisions, while they appear to be absent in others. Typically, cell divisions are rythmical, alternating, qualitatively and quantitatively equal, and con- sequently non-differential. The differentiations of cleavage cells are due to depart- ures from this typical condition in one or more particulars. In certain animals these departures are very notable, the cleavage being from the first differential. The differentiations of cleavage may have a far-reaching prospective significance, since in certain animals (annelids, mollusks, polyclades and nematodes) the principal axes and body regions of the future animal are marked out by the cleavage planes and entire organs are represented by a single cell or group of cells. In such cases the minutest details of unequal, bilateral or qualitatively dissimilar division of cells may be of great importance. The forms and peculiarities of such cleavage are inherited quite as certainly as are any adult features, and when the problem of inheritance may be reduced to a certain peculiarity of a certain cell division it is evident that we have this problem reduced to relatively simple terms.
The species which has formed the chief subject of this work and from which all the figures are drawn is Crepzdula plana Say. ‘The following species and genera have also been studied more or less completely :—
C. fornicata Lam., C. convexa Say, C. adunca Keep, Urosalpinx cinerea Stimp- son, Sycotypus canaliculatus Gill, Fulgur carica Conrad, Hamznea solitaria Say, Lolis papillosa Loven.
Many of the phenomena here recorded, particularly those relating to proto- plasmic movement and the history of the centrosomes during cleavage, have been observed in all of these seven species of Prosobranchs and two species of Opistho- branchs. These phenomena are, therefore, not isolated, and it is probable that they are of wide occurrence.
METHODS.
The eggs of Crepidula are in many respects peculiarly favorable for study. These mollusks are very abundant and the eggs are deposited in large numbers, a single female usually laying from eight to ten thousand eggs. No other egg with which. I am acquainted is so favorable for the study of cleavage. The eggs are inclosed in membranous capsules, which reagents readily penetrate; however, in order to insure rapid fixation I have usually punctured the capsules with a needle before putting them into the fixing fluid; they are then passed through the various reagents and finally imbedded and sectioned while still in the capsules. Each female of C. plana deposits an average of fifty capsules, with approxi- mately 175 eggs in each capsule. The eggs vary scarcely at all in size, each being about 0.136 mm. in diameter. The great advantage of being able, without further trouble, to handle in large numbers such small eggs will be at once apparent. On the other hand the eggs are unfavorable in having a relatively large quantity of yolk which is colored deeply by most stains.
Two general methods of observation have been followed: (1) the study of entire eggs; (2) the study of serial sections. For the former the best method of preparation is as follows: The living eggs are teased from the capsules into Kleinenberg’s picro-sulphuric fluid, or into Boveri’s picro acetic, where they are left from thirty minutes to two hours; they are then washed in alcohol until the eggs become nearly white, and are stained in the following solution for from five to ten minutes.
KARYOKINESIS. {i
Delafield’s hzematoxylin 5 é ; F : : 5 Allee. Distilled water : 3 5 3 : ; : : LOR CC: Kleinenberg’s picro-sulphurie fluid : é : : . 10 drops.
Eggs so stained are washed in alcohol, dehydrated and mounted in balsam. If the eggs of C. plana are mounted under a thin cover, supported at one side by a bit of glass .15 mm. thick (the eggs are about .136 mm. in diameter), they can be studied under an immersion lens and the relative posi- tions of nuclei, centrosomes, spheres and mid-bodies (‘“Zwischenkérpern”), can be determined as would be impossible in sections.
I have tried many modifications of this stain, but have found none so good as the one given. The yolk is left transparent yellow, while the protoplasm is of a rosy tint, and the chromatin, centro- somes and mid-bodies are blue or black. I have found by experience with the eggs of many mollusks, annelids, crustacea and insects, that this method is especially useful with eggs in which there is much yolk.
Eggs fixed in Kleinenberg’s fluid are always more transparent than those fixed in Boveri’s, the yolk stains less and the spheres, centrosomes, spindles and mid bodies stand out more clearly. On the other hand the structure of the cytoplasm is not so well preserved in Kleinenberg’s as in Boveri's fluid.
For sectioning the following fixing fluids have given good results: Flemming’s, Hermann’s, Boveri’s picro acetic, Graf’s picro-formol; while the following were less satisfactory: Kleinenberg’s fluid, Perenyi’s fluid, chromo-acetic, chromo-formic, sublimate-acetic, sublimate (sat. sol.). Many different stains were used, among them Heidenhain’s iron-alum and Delafield’s hematoxylin, either alone or in combination with orange G., Bordeau red, eosin or acid fuchsin; Hermann’s saffranin- gentian-iodine; Biondi-Heidenhain’s mixture; Auerbach’s acid fuchsin and methyl green, ete. The best results were obtained with material fixed in Boyeri’s picro-acetic and stained in Delafield’s hematoxylin and orange G.!
Many preparations which were so stained were afterwards decolorized and restained with iron hematoxylin and Bordeau red. In this way it was possible to observe certain details of structure which could not otherwise be determined with certainty. Since there are always devotees of certain fixing and staining media who refuse to accept results unless their favorite methods haye been employed, I may say here that every important result of my work has been confirmed by material fixed and stained in each and every one of the methods enumerated above. These results, therefore, cannot be attributed to the exclusive use of one or two media.
In general the eggs were imbedded in paraffine in the usual way, but owing to the fact that they contain a large quantity of yolk and are, therefore, more or less friable, some were doubly imbedded in celloidin and parattine, after the method suggested by Kultschitzky (’87) and by Ryder (88). This method is especially valuable for tracing the movements of the pronuclei and spheres through the yolk. The advantages of this method over ordinary celloidin imbedding are that thinner sections can be had, and that these can be cut in series or ribbons. Its most serious disadvantage is that the sections are usually much folded and cannot easily be flattened.
MATURATION. A. PRE-DIVISION STAGES.
1. Tue OvartAn Ecc.—In the stages just before its escape from the ovary, the egg of C. plana is irregular or polygonal in outline, being pressed into this shape by surrounding eggs, and it contains a very large germinal vesicle located near the center of the cell. The egg is filled with yolk spherules, which vary greatly in size; there is extremely little cytoplasm visible, and this lies close around the nucleus. The latter contains a single large nucleolus, within which is a
"The great value of Heidenhain’s iron hematoxylin method I have had abundant oceason to verify. Ihave used it extensively in this work and for certain purposes it is without an equal; but having said this, I wish to add that it is not by any means the best stain for all purposes and with all material, as some recent workers seem to suppose. It is not, as is now well known, a specific centrosome stain. In all mollusean eggs which I have examined the yolk stains as densely and retains the stain almost as tenaceously as the centrosomes. For this reason I have found it impossible by this method to recognize small centrosomes and asters in a dense mass of yolk, whereas with the picro-heematoxylin the yolk is left a clear yellow, while the aster is red and the centrosome black. The iron hzematoxlin is also a less delicate stain for spheres and cytoplasm than picro-hematoxylin while, of course, it cannot be used at all in the preparation of entire eggs.
8 KARYOKINESIS.
more densely staining body, usually eccentric in position, Plate I, fig. 1. The chro- matin is in the form of small granules, varying in size and irregular in shape, which are attached to linin threads stretching through the vesicle. Many of these gran- ules can be seen to be three- or four-parted, though others are rounded or irregular in shape. The four-parted granules are larger than the others but all are extremely small; their method of formation was not observed. At this stage no trace of cen- trosomes can be found anywhere in the egg.
2. Eaa Layine.—After the eggs leave the ovary they descend to the lower enlarged part of the oviduct, where they meet spermatozoa from the receptaculum semints and together with an albuminous fluid are surrounded by a glairy, mucous substance, which hardens into capsules. These capsules are attached together in a cluster and are fastened by a common stalk to the object upon which the female is seated. All the eggs laid by one individual begin development at nearly*the same time and proceed with remarkable uniformity, so that whenever examined they are all found to be in approximately the same stage.
The earliest stages of free eggs which I have seen were taken from the oviduct while the capsule was being formed. The outline of the egg at such a stage is usually elliptical or irregular, being in marked contrast with the spherical form which it attains after the entrance of the spermatozoon. The germinal vesicle is slightly eccentric in position and immediately around it there is a small amount of cytoplasm in the interstices between the yolk spheres; elsewhere in the egg the yolk spheres are closely crowded together. The nucleolus is now a single, homogeneous body and frequently exhibits an alveolar or recticular structure. The chromatin granules are rounder and a little larger than in the ovarian egg and many of them are arranged in rows or strands, fig. 2. In one egg of this stage I observed two minute granules in the cytoplasm, close to each other and in contact with the nuclear membrane; these are possibly the centrosomes, though no polar. radiations or central spindle was observed. At this stage the spermatozoon has not entered the ege (Plate I, fig. 2).
B. Maturation Divisions.
1. Nucreus.—The earliest trace of the first maturation division which I have seen appears about the time of the entrance of the spermatozoon. The centrosomes are now plainly visible, being surrounded by a few short radiations, and are connected by a central spindle. At the same time the nuclear membrane is indented opposite the poles of this spindle and fibres can he traced from the centrosomes to these indented areas, Plate I, figs. 3 and 4.
At this stage the germinal vesicle contains a great number of chromatin granules which are connected together by linin threads, also a single extremely large nucleo- lus, while the nuclear sap fills all the interstices within the nucleus and constitutes the greater part of its bulk.
(a). Chromatzn.—A few of the chromatin granules are larger than the others and their form is spherical, 2-lobed, 5-lobed or 4-lobed. They are probably identi- cal with similar granules found in the pre-division stages. They differ much in
KARYOKINESIS. 9
size and for this reason their number cannot readily be determined since they grade down to the smaller granules, which are innumerable (Plate I, figs. 5 and 4). These larger granules continually increase in size and become the chromosomes of the first maturation spindle; some of these granules stain less deeply at the center than at the periphery. As the chromosomes grow in size the remaining chromatin granules, which constitute the chief bulk of chromatin within the germinal vesicle, grow smaller and smaller and are gradually dissolved; on the disappearance of the nuclear membrane they escape with the nuclear sap into the cytoplasm, figs. 5, 6 and 7. At the same time the linin threads, which were plainly visible at an earlier stage, fig. 4, are no longer to be seen, but the arrangement of some of the granules in radiating lines, fig. 5, is probably to be taken as evidence that some of these threads still exist. In the early stages of the first maturation division, all the chromatin granules stain alike, in later stages the chromosomes stain more densely with nuclear stains, while the remaining granules show an increasing affinity for plasma stains. In such stages as are shown in figs. 3 and 4, there are no perceptible _ differences, save size only, between the granules which become chromosomes and those which dissolve; the fact, however, that the history of these two groups is so different shows there is some fundamental difference between them. It is highly probable that the faintly staining granules which are ultimately dissolved or trans- formed into linin are identical with the lanthanin or oxychromatin of Heidenhain (94).
(b). Nuclear Sap. Before the solution of the oxychromatin granules and nucleolus begins, the nuclear sap is a clear and almost non-colorable fluid. As the solution of these elements progresses the nuclear sap becomes granular and tingible, staining blue or purple with Delafield’s haematoxylin alone, though it stains deeply with plasma stains, such as eosin or orange G. when these are used after the hxema- toxylin. Even after the nuclear membrane has entirely disappeared the nuclear sap and oxychromatin can still be recognized as a granular mass, figs. 5, 6, 7.
Korschelt (95) has observed a similar increase in the staining properties of the nuclear sap of Ophryotrocha, where the “ Kernsaft”’ stains more and more deeply as division advances until it becomes so dark that the chromatin threads are invisi- ble. Then the sap loses some of its staining qualities and, at the same time, dis- dissolved nucleolar substance is probably added to the chromatin threads.
1 The term Achromatin, as used and defined by Flemming (’82, p. 375), is limited to “ that formed substratum of nuclear structures, as well as of the division figures, which is not colored by nuclear stains.” As thus defined, it is applicable only to the linin, and is not even applicable to it at all stages in the cell eycle, since the linin also is colored by nuclear stains at certain stages. Furthermore, it is extremely probable that oxychromatin is transformed into linin at certain stages, and that oxychromatin and perhaps linin are sometimes dissolved in the nuclear sap. This interrelation of these yarious parts of the nuclear substance makes it impossible to apply the terms “chromatin” and “achromatin” as used by Flemming. Nevertheless, it is convenient to have a term which will include all of the nuclear constit- uents which do not form chromosomes, as contrasted with that which does. Since the term achromatin has frequently been used in this sense, and since I am unwilling to further cumber cytological nomen- clature with new names, I shall employ the term “achromatin,” or “achromatic substance of the nucleus” to include all the contents of the nucleus except the chromatin, and even that portion of the chromatin
which does not form chromosomes. As thus used it includes linin, oxychromatin, nuclear membrane and nuclear sap.
2 JOURN. A. N. S. PHILA., VOL. XII.
10 KARYOKINESIS.
(c). Mucleolus —The chromosomes, which are at first widely scattered through the nucleus, figs. 3 and 4, gather together more closely and often lie immediately around and upon the nucleolus, figs. 5 and 6. In some cases it looks as if these chromosomes were being formed out of the substance of the nucleolus, and the fact that the nucleolus diminishes in size as the chromosomes increase lends color to this view. On the other hand, when the chromosomes first appear they are scattered through the entire nucleus and do not lie close to the nucleolus, and though it is possible that they may later receive substance from the dissolving nucleolus, it is impossible to suppose that they are fragments of the nucleolus. The latter is grad- ually dissolved without any fragmentation. Before the complete disappearance of the nuclear membrane the nucleolus has greatly diminished in size, and at the same time the nuclear sap stains more deeply, fig. 5. After the disappearance of the nuclear membrane the nucleolus comes to lie outside the spindle, while most of the chromosomes are found within it, though some of them may still be scattered among the polar radiations, fig. 7.
Within the cytoplasm the nucleolus continues to diminish rapidly in size and soon entirely disappears, figs. 8, 9, 12a. In this respect the history of the nucleolus in the first maturation of Crepzdula is the same as has been described by Hacker (93), Foot (94), Mead (95), Wheeler (95), Obst (’99), and others, in a considerable number of animals.
(d). Chromosomes.—The shapes of the chromosomes of the first maturation spindle are shown in figs. 7-15 and in text fig. I. In the early prophase the most common form is that of a 3- or 4-lobed body; in fact, such bodies are found in the nucleus of the ovarian ege. There can be little doubt that these are the “tetrads”’ of authors, though in Crepzdula they are not always 4-lobed. As these chromosomes increase in size a hole appears through the middle, between the lobes. There are also found circular or elliptical rings which may be completely closed or may be open on one side; also dumb-bell and cross-shaped bodies. All these forms are represented in text fig. I; all the 2-part chromosomes are shown in the first line (A), the 3-part ones in the second line (B), and the 4-part one in the third (C). These forms are grouped according to evident resemblances merely and it is not cer- tain that they always stand in the genetic relations indicated. For example, A, 4 and 5 may give rise to B, 6 and 7; B, 4 and 5 may be only variations of C, 3 and 4, ete. In all cases, however, the short chromosomes of the early prophase give rise to rod-shaped or elongated chromosomes in the metaphase. In some cases (e. ¢. line A) this is probably accomplished by these chromosomes becoming ring-shaped and by the opening of these rings on one side. If the ring shows no thickenings (A, 2 and 3) a rod-shaped chromosome is formed by its opening, which later becomes dumb-bell shaped (A, 4 and 5); if it shows three thickenings (B, 1-4) it gives rise when opened to a rod with a thickening at each end and one in the middle (B, 5). The 4-part chromosomes (C, 1) are frequently drawn out into cross-shaped ones; these crosses usually have a hole through the middle and each arm of the cross is split lengthwise from this hole nearly to the tips (C, 2.and 3). In later stages the
KARYOKINESIS. 11
arms which lie in the spindle axes lengthen, while the transverse arms shorten, the hole disappearing; in this way chromosomes are formed with three enlargements (C, 4 and 5) similar in all respects to those described above (B, 4 and 5). The resemblance of these two forms is so close that it is difficult to explain the differences in their mode of origin. It is possible that forms, such as those shown in B, 4 and 5, are really crosses with short transverse arms, the tips of all four arms being bent toward each other until they nearly or quite meet.
The striking differences in the shapes of the chromosomes of the prophase is continued into the metaphase where at least three distinct types may be recognized as shown in text fig. I, lines A, Band C. In the late anaphase, however, all come back to a cubical or tetrafoil condition; a hole is usually present through the middle of these as in the prophase.
. 3 4 5 Gieiana Clean e 10 11
Propnase Metaphase Anaphase
Fria. I.—Chromosomes of the First Maturation Division of Crepidula.
Various authors have called attention to the variety in the form of the chro- mosomes of the first maturation division (v. Klinckowstrém 96, Van der Stricht, 98, Foot 98, Lillie 98, Griffin 99). Foot and Lillie figure 3-part chromosomes in the metaphase of the first maturation of A//olobophora and Unio entirely similar to those in Crepzdula, and Lillie shows these chromosomes split longitudinally, as they must be, if formed from crosses as shown in text fig. I, line C.
It is difficult to say whether these differences in the shapes of chromosomes mean much or not. On the one hand it is possible that all the chromosomes of a
12 KARYOKINKESIS.
given mitosis cannot be reduced to a single type; that their differences in shape indicate differences in material substance, and that different chromosomes may therefore represent different heritable qualities. On the other hand these differ- ences in the shapes of chromosomes are generally limited to the first maturation division; they are rarely found in the second maturation and only to a limited extent in the cleavage. Furthermore, there are many evidences that the shapes of chromosomes are conditioned by their linin sheaths, and that the chromatic substance which fills the sheath is of a semi-fluid of viscid character. Thus in the metaki- nesis of the first maturation, it is always found that the chromosomes have enlarged ends toward the poles of the spindle, and that they are drawn out into thin connect- ing threads in the equator. In this region the chromosomes are frequently monili- form in shape (text fig. I, C, 8 and 9 and Plate I, fig. 15), and cross sections through the equatorial region of these elongated chromosomes shows many of the latter sur- rounded by a clear zone, which is bounded by a dark line (Plate I, fig. 12a). This clear zone is entirely lacking in sections through the enlarged ends of the chromo- somes. In fig. 12a one chromosome lies entirely outside of the spindle substance, and yet it is surrounded by this clear zone; this zone and its outer dark boundary is not therefore a mere expression of the absence of spindle substance around the chromosome, but is a structural peculiarity of the chromosome itself, and probably represents a linin sheath, which is separated from the contained chromatin in the equator, but is entirely filled by the chromatin at the poles. After the complete division of this thread of chromatin and its withdrawal into the enlarged ends of the daughter chromosomes, the linin sheath may still be seen for a long time con- necting the latter together, and constituting a connective fibre.
The chromosomes grow continually during the early stages of the first matura- tion division and reach their greatest size in the metaphase when each is from two to twenty times the size of the largest granule present in the germinal vesicle (ef. figs. 3 and 12). After this stage they do not appear to increase in volume. ‘The great differences in the size of chromosomes in the same spindle is almost as striking as their differences in shape ; for example, the volume of the largest 3-part chromo- some in Plate I, fig. 12 and text fig. I, B, 7 and 8, is about fifteen times that of the smallest; I am unable to say, however, whether this difference in the size of chro- mosomes is the same in all eggs or not. I do not remember that any one has recorded such enormous differences in the size of chromosomes as are here described. Montgomery (98) says that the chromosomes of one of the spermatocytic divisions in Pentatoma vary greatly in volume, the largest sometimes having six times or more the volume of the smallest. In no other mitosis in Crefzdu/a is there such variety in the size of chromosomes, and nowhere else are there such differences in shape.
The number of chromosomes in the first maturation division is thirty, as I have determined by a careful study of the entire mitotic figure, as well as by cross sections through the equatorial plate. Such a cross section is represented in Plate L, fig. 12a, and the whole number of chromosomes is there shown. This is undoubtedly the
KARYOKINESIS. 1:
Qo
reduced number of chromosomes; I have been unable to count with accuracy the number present after fertilization, but it is evidently about sixty.
In the early anaphase of the first maturation, the daughter chromosomes are either dumb-bell shaped, or cubical or spherical masses, frequently with a slender process of chromatin running from each daughter chromosome toward the other, figs. 15-15 and text fig. I. A little later each chromosome becomes cubical or quatrefoil in shape, and this form persists, and is universal until the metaphase of the second maturation division, when all become 4-parted, figs. 16-51 and text fig. II, and then cross shaped exactly as in the first maturation. In this ease, however, the arms of the cross are not split longitudinally, but the separation takes place between the arms, so that in the metakinesis two of the arms or spherules, in the form of a dumb-bell, go to one pole and two to the other, text fig. I], 5 and 6. In the anaphase of the second maturation, these dumb-bell or rod-shaped chromosomes again become cubical or spherical, as in the anaphase of the first maturation (Plate II, figs. 82, 33 and text fig. II, 8).
The cubical chromosomes at the close of the first maturation are about half
as large as the fully-formed chromosomes in
Wrest ce Sle ea eke atthe metaphase of that division, while those o 8 a in the anaphase of the second maturation
& & Y ap - are about one-half the size of those in the 8 ja prophase of this division, z. ¢., the volume
Prophase Metaphase | anaphase Of each chromosome at the end of the sec- ond maturation division is about one-quarter
Fic. I1.—Chromosomes of the Second Maturation p » : Division of Crepidula. (In the reproduction that of the fully-formed chromosome in the
pbeeeuie nas peoumred uced more Luan tel!) metaphase of the first maturation. There has been therefore no growth of the chromosomes after the metaphase of the first maturation. The number of chromosomes in the second maturation division is the same as in the first, viz. thirty, and the same number is left in the egg after the second polar body has been formed.
It is especially noteworthy that in the prophase and anaphase of both matura- tion divisions the chromosomes are cubical or quatrefoil in shape. In the metaphase of the second maturation, figs. 27-531 and text fig. II, the chromosomes look like typical “tetrads”’ and they would undoubtedly be called such if they occurred in the first maturation.
Similar 4-parted chromosomes in the second. maturation are figured by v. Kline- kowstrém (96) in Prostheceraeus, by Van der Stricht (98) in Zhysanozoon. and by Byrnes (99) in Lzmax.
In Crepzdula it is impossible to say whether the plane of division of the chro- mosomes is the same or is different in the two maturation divisions. At the begin- ning and at the end of both divisions the chromosomes are cubical or quatrefoil in shape, and one might as well speak of the ** longitudinal” or “transverse” division of a cube or sphere as of these chromosomes. — It is impossible, therefore, to deter- mine whether or not reduction in the sense of Weismann takes place in this case.
14 KARYOKINESIS.
Griffin points out that the division of the tetrad in 7halassema and Zirphea is unlike that in the Copepods, in that in the former each spherule of the tetrad becomes an arm of a cross and that these arms then split longitudinally, whereas in the Copepod type two entire spherules separate from the other two. The former he calls a “spurious tetrad” (cross form), the latter a “tetrad of the Copepod type.” In Crepidula, just as in Thalassema and Zirphea, the tetrads are of the “spurious” type in the first maturation, whereas in the second maturation we have chromo- somes which, in every respect, resemble tetrads of the Copepod type.
In the late anaphase of the second maturation the chromosomes which remain in the egg become vesicular and fuse together to form a few vesicles with large eranules of chromatin on their walls, Plate II, figs. 34-35. Finally all of these vesicles fuse into one, as is shown in fig. 36, e¢ seg.
2. CENTROSOMES AND CENTRAL SpINDLES.—The earliest stage at which centro- somes have been seen was in an egg from the oviduct, not yet fertilized, fig. 2. In this ege the centrosomes are already present as two minute bodies, in contact with the nucleus and without any apparent central spindle or polar radiations. In fact, because of the absence of these radiations it is impossible to be certain that the two granules shown in fig. 2 are really centrosomes. In other eggs from the oviduct, figs. 8-7, into which a spermatozoon has penetrated, the centrosomes are larger, a central spindle is present and polar fibres are abundant. I have been unable to determine whether this central spindle arises as a centrodesmus (Heidenhain) or whether its fibres grow out independently from the two centrosomes and afterwards unite to form the spindle. In general it may be said, that the formation of the mitotic figure usually begins with the entrance of a spermatozoon into the egg.
In the prophase of the first maturation the centrosomes are minute densely staining points; they grow larger as mitosis advances, and in the stages immediately preceding the metaphase, figs. 8, 8a, 11, and text fig. III, they are more or less irregular in shape, and when deeply stained and strongly destained with the iron- alum-haematoxylin of Heidenhain, they may be seen to contain a central clear area. Around this clear area the dense walls of the centrosome are thickened in places and ‘may, perhaps, represent large granules in contact with one another, as Lille has found to be the case in U/zz0. In the prophase of the second maturation, the cen- trosomes are so small that I have found it impossible to make out their structure with certainty, but they are in many cases slightly irregular in form (cf. figs. 27-31), from which I conclude that their structure is the same as in the prophase of the first maturation.
In the metaphase of the first maturation (fig. 12), the centrosomes are spherical bodies about 1 w in diameter. They contain a central area which stains faintly, around which is a thick, dense zone which corresponds to the irregular or granular zone of the prophase; in the metaphase, however, this zone is perfectly regular and gives no indication of being composed of granules as in the preceding stage.
In the anaphase of the first maturation the centrosomes become large hollow spheres, the peripheries of which stain deeply while their central areas remain
KARYOKINESIS. 15
clear. Within the central area a faintly staming body becomes visible, which, in its turn, becomes a hollow sphere (text fig. III, e and /); this is the “corpus- cle central” of Van Beneden (87), or the “centriole” of Boveri ('95).*
Up to this stage the centrosomes at the two poles of the maturation spindle are identical in size and struc- ture. When, however, the outer pole of the spindle comes into contact with the egg membrane, the sphere and centrosome at this pole become flattened, figs. 14-16, though the cen- trosome still shows its dark periph- ery, its central clear area and its cen- tral corpuscle.* In the late anaphase the outlines of the centrosome at the outer pole are marked by a layer of granules, while within the central clear area is the elongated central corpuscle, fig. 16a. Finally, after the complete separation of the first polar body, the granular outlines of this centrosome disappear, though the central corpuscle, or rather the cen- trosomes and central spindle to which it gives rise, are still found within it (Plate II, figs. 22, 28a, 29, D4, 36).
The centrosome at the inner pole of the spindle continues to en- large until it reaches a truly aston- ishing size, becoming fully 4 uw in diameter. Its peripheral layer is at first a solid zone of deeply staining material. In later stages this zone breaks up into plates, figs. 16-25 and text fig. III, and still later it ap- pears as a ring of close set granules.
Fig. I1I1.—Centrosomal Cycle in Fig. IV.- Centrosomal Cycle in the Maturation of Crepidula. the Cleavage of Crepidula.
‘ Boveri (’01) denies the general homology of his “centriole” with the ‘ corpuscle central” of Van Beneden. There can be no doubt that the structure in question in Crepidula is homologous with Boveri’s “centriole.”
2 Van der Stricht (’98) has observed a similar flattening of the sphere and centrosome against the egg membrane in the first maturation of Thysanozoon.
16 KARYOKINESIS.
The central corpuscle, which is shown in figs. 15 and 14 as a faintly staining, hollow sphere, soon becomes elliptical in shape, fig. 15. At the poles of this ellipse its walls grow thicker and stain deeply. These thickened points become the centro- somes of the second maturation spindle, while the remainder of the ellipse forms the central spindle? (cf. figs. 14-16a, and text fig. III, ¢, 2). The daughter centro- somes and central spindle lie within the mother centrosome; the outlines of the mother centrosome then disappear and the new amphiaster lies free in the granu- ular remains of the sphere. ;
During this metamorphosis the centrosome undergoes great changes in its stain- ing qualities; in the prophase and metaphase it stains deeply with haematoxylin; as it enlarges, however, the peripheral portion only takes haematoxylin, while the cen- tral part takes plasma stains; finally, in the late anaphase, even the peripheral por- tion takes plasma stains.
At no time during this metamorphosis do the astral radiations penetrate the centrosome. As long as they can be seen they remain attached to its surface, and even after the new amphiaster has arisen within the mother centrosome, the rays are still centered on the figure as a whole, figs. 25-28, and text fig. Ill, ¢ and 2. The new radiations which appear around the daughter centrosomes develop de novo, as MacFarland (97) and Griffin (99) maintain.
Up to the time when the second polar body is cut off, the history of the centro- somes during the second maturation is similar to that already described for the first ; at the beginning of division they are minute granules, as the division advances they become larger, and in the anaphase are large hollow spheres.
After the second polar body has been formed, however, the centrosome which remains in the egg becomes a very large sphere filled with many coarse granules and with a boundary layer of close-set granules, from which, in some cases, polar fibres proceed, figs. 34-36. I have never seen a peculiarly large granule which might be regarded as a centriole within this centrosome, nor have I seen the formation of a central spindle as at the close of the first maturation. On the other hand, the for- mation of a large number of granules within the centrosome is a phenomenon which occurs in the telophase of every cleavage (text fig. IV), and seems, there- fore, to be the more usual process. It seems probable, therefore, that the par- ticular manner in which the daughter centrosomes and central spindle arise within the mother centrosome at the close of the first maturation, is a modification of the more typical process shown in the cleavage, due, perhaps, to the entire omission of a resting stage between the two maturation divisions. In the maturation, there-
‘In view of the remarkable resemblance of this stage to a corresponding stage in the division of the “centrosome” in Diaulula (see MacFarland’s figs. 36, 37, et seq.) it may be supposed that the vesicular body which I have called the ‘central corpuscle’’ is really the centrosome and that the surrounding body is only the inner zone of the sphere. Fortunately, however, the outlines of the centrosome are so perfectly distinct and its history, as shown in my preparations, so continuous, that there can be no question as to its identity in this case. The outlines of the old centrosome remain until after the central corpuscle has given rise to a perfect spindle within it; so that in Crepidula, and several other gastero-
pods which I have studied, the new centers and central spindle arise from the central corpuscle and not from the entire centrosome as in Diaulula.
KARYOKINESIS. 17
fore, the centrosomes, ike the chromosomes, undergo an unusual type of division. The centrosome which goes into the second polar body completely disappears.
The egg centrosome is surrounded by a sphere with grows with the growth of the centrosome until it becomes very large and is filled with faintly staining granules which are held in a delicate reticulum, fig. 56. In some cases the outlines of this sphere are very distinct, in others more faint, but im all cases there are strongly marked polar fibres, which run from the periphery of the sphere for a considerable distance through the egg; some of these fibres may also be traced into the spheres where they appear to break up into rows of granules. In stages later than fig. 56 I am unable to recognize the egg centrosome; its granules merge with those of the surrounding sphere and its outlines are no longer visible.
In structure and metamorphosis the centrosomes of the maturation divisions of Crepidula are very similar to those described by MacFarland (97) in Dzau/lu/a, Lillie (98) in Unzo, Vejdovsky and Mrazek (98) in Rhynchelmzs and Van der Stricht (98) in Zhysanozoon. The further consideration of these centrosomes will be deferred until after the description of the cleavage.
3. Poutar Rays, SprnpLte Fipres AND SpHERES.—At the time of the entrance of the spermatozoon into the egg, figs. 5 and 4, the centrosomes are surrounded by polar fibres and the nuclear membrane is indented in the region of each centrosome. A large number of fibres, forming a cone or half spindle, can be traced from each centrosome to the indented portion of the nuclear membrane. Within the nucleus the linin threads, with their attached chromatin granules, are arranged in the same radiating lines as these fibres and form an intranuclear continuation of them, figs. 5 and 4.
In early stages of this division both the extra- and intra-nuclear portions of the spindle consist of branching and anastomosing threads, along which are ranged oxy- chromatin granules. These two groups of fibres, z.¢., those imside and those out- side of the nucleus, are so essentially alike that I cannot doubt that both are derived from the same substance, viz., the linin and oxychromatin of the nucleus, in which case the extra-nuclear portion of the spindle must be formed from nuclear constitu- ents which have escaped from the nucleus at the indented areas mentioned above.
The polar fibres also consist of threads along which granules are attached, figs. 4,5 and 6. In the first maturation they closely resemble the spindle fibres. and, like them, may be derived from the achromatic nuclear substance. These granules are rarely arranged in concentric spheres as Heidenhain, Driimer and Lillie have found them. As mitosis advances, both polar and spindle fibres become smooth. There can scarcely be any doubt that this is accomplished by the transformation of these granules into the substance of the fibre (cf. Boveri, °88, p. 80, Wilson °95, Griffin *99). Again, in the dissolution of the spindle one frequently observes that the fibres become varicose, as in the early stages of mitosis, while the portion of a fibre between granules becomes less and less prominent, figs. l6a@ and 53. I have been unable to observe these same varicosities on the fibres of the central spindle in the maturation divisions, but they can be seen in the central spindles which are found during the cleavage of the egg (Plate LV, figs, 74, 75 and 76, and text fig. IV).
3 JOURN. A. N. S. PHILA., VOL. XII.
18 KARYOKINESIS.
The infolding of the nuclear wall mentioned above must, of course, be accom- panied by the escape of some substance from the nucleus. Coincidently with this infolding of the nuclear membrane the polar fibres and extra-nuclear spindle fibres become longer, stouter and granular. At the same time the spindle and the sphere surrounding the centrosome stain more deeply, owing to the presence of an interfilar substance, which stains like nuclear sap... The nuclear membrane then completely disappears, but the nuclear contents preserve for some time the outline of the nucleus and can easily be distinguished from the surrounding cytoplasm because of greater affinity for stains, figs. 6 and 7. The entire mitotic figure, with the excep- tion of the polar systems, lies within this granular area and the enormous growth of the spindle is at the expense of this intra-nuclear substance. Not all of the achromatic substance of the germinal vesicle is confined to the spindle and the polar systems, a large part of it passes directly into the cytoplasm, which is increased in quantity after the nuclear membrane dissolves.
Both the aster and the spindle are plainly composed, during the early stages of mitosis, of two constituents, viz., fibres with their attached granules and an inter- filar substance. Between and around the spindle fibres, both in the first matur- ation and in all subsequent divisions, there is a homogeneous interfilar substance which colors deeply with plasma stains. This substance is sharply delimited from the surrounding cytoplasm, as is shown in figs. 12 and 12@ and also in later stages of both the maturation and cleavage.
A cross section through the equator of the spindle in the metaphase, fig. 12a, shows the interfilar substance of the spindle as a homogeneous mass, staining deeply with orange or eosin, and with stellate radiations running out into the cytoplasm in all directions. These radiations around the equator are shorter, blunter and more irregular than those at the poles. They are probably caused as follows: in the forma- tion of the definitive spindle there is a general elongation of the linin reticulum in the direction of the spindle axis and a contraction of the reticulum at right angles to that axis; at the same time there is a condensation of the interfilar substance, the more fluid karyolymph being squeezed out of the spindle. In this equatorial shrinkage some of the linin threads probably remain attached peripherally and thus cause the stellate radiations.” The chromosomes le within this interfilar substance, though occasionally one is found just outside of it, fig. 12a, and they are scattered through the entire thickness of the spindle, and not merely in a wreath around the periphery. There is, therefore, strictly speaking, uo ‘‘central spindle” in Crepzdu/a as contrasted with a ‘peripheral spindle,’ but the fibres of the two must be intimately com- mingled.
* Driiner calls attention to the fact that the nuclear membrane is dissolved at points opposite cen- trosomes and that coincidently the rays grow stronger; and R. Hertwig (99) has observed in Actino- spherium that nuclear material probably escapes into the plasma cones, since the latter stain more deeply about the same time that the nucleus shrinks in size.
* Similar radiations around the equator of the spindle have been figured by Korschelt (795, figs. 131, 139), Mead (’98, fig. 18), and Gardiner (98, figs. 28, 34).
KARYOKINESIS. 19
An interfilar substance, which is to all appearances similar to that of the spindle, surrounds the centrosomes and radiates along the polar fibres, so that in all middle stages of mitosis it is difficult to recognize the polar fibres and spindle fibres when once they are surrounded by it.
In the later stages of mitosis this interfilar substance moves to the poles of the spindle, again allowing the spindle fibres to be seen distinctly ; it also moves inward toward the centrosome, leaving the polar fibres sharply marked, fig. 22, and thus ageregated, from the spindle and polar rays, forms a sphere with rather indefinite out- lines. This sphere differs notably in character from that which is found in many other animals, e. g.. the outer sphere of Unzo (Lillie 98) and the couche corticale of Thysanozoon (Van der Stricht’98). The latter are clear zones of definite outline, with faintly staining astral rays running through them; in Crepzdu/a, on the other hand. this zone is indefinite in outline until the late anaphase or telophase, and is even then not so sharply bounded as Lillie and Van der Stricht have shown it; further, instead of being a zone which is clearer than surrounding parts, it is denser and more deeply staining. The spindle fibres and polar rays can be traced through this sphere to the centrosome in early stages of mitosis, but in middle stages fewer radiating fibres can be seen in it (cf. figs. 5-8a@ with figs. 11-16). In later stages again polar rays can be traced through it to the centrosome (cf. figs. 22-24 and 54-56).
The origin of this interfilar substance is difficult to determine. In the aster it seems to be principally derived from hyaloplasm (interalveolar substance) of the cell body, which is aggregated toward the centrosomes, the larger alveoles of the cyto- plasm and the yolk spherules being crowded out from the centrosome as the inter- alveolar substance moves in toward it. In the spindle, on the other hand, the interfilar substance seems to be formed in large part from achromatic material of the nucleus; such interfilar substance exists before the nuclear membrane is broken, though it is at this stage much less dense than in the fully formed spindle. When the nuclear membrane dissolves at the poles this substance escapes into the extra nuclear spindle and spheres; it is quite possible that at the same time there may be an invasion of the spheres and spindle by hyaloplasm from the cell, this double movement being in the nature of a diffusion in both directions. The fact that the interfilar substance is denser than either the nuclear sap or hyaloplasm may perhaps indicate that it is a new substance formed by a combination of the two. While this suggestion as to the origin of the interfilar substance accords well with all my obser- vations as to its character and movements, it cannot be considered as more than a suggestion.
The form and size of the first maturation spindle varies greatly in different phases. From the prophase to the metaphase it increases in length and diameter, becoming most stout in the metaphase ; in the early anaphase it continues to increase in length, becoming about as long as the radius of the egg, figs. 11-14, and at the same time it grows very slender; finally, in the late anaphase it again shortens, be- coming stouter, until it is not more than one-half as long as in the metaphase or early anaphase, figs. 15 and 16, and at the same time the chromosomes are pushed
20 KARYOKINESIS.
right into and through the sphere until they come into contact with the cell wall. In no other mitosis is there such a shortening of the spindle; in fact, in all other divisions, with the possible exception of the second maturation, the spimdle continues to grow longer throughout the whole of the mitosis. A similar shortenmg of the first maturation spindle has been observed in Ascarzs (Boveri 87), Branchipus (Brauer °92), Opkryotrocha (Korschelt °95), MZyzostomum (Wheeler °95), Cerebra- tulus (Coe 99), Polycherus (Gardiner ’98), Axolot/ and Trzton (Carnoy and Lebrun 99, see their figs. 110 and 112). The principal cause of this shortening is to be found in the peripheral movement of the mitotic figure, as will be described in Part IL; its chief result is the production of a much smaller polar body than would be possible if the spindle maintained its maximum length throughout the later stages of division. At the time of its greatest length the first maturation spindle is about one-half as long as the diameter of the egg, and since the division of the cell body always takes place through the middle of the spindle, the first polar body would have a diameter one-quarter that of the egg were it not for this shortening of the spindle. ;
I agree with G. Niessing (799), that the shape of the spindle, z.e., whether it is stout or slender, is due to the quantity and location of the interfilar substance, but this depends upon the degree of contraction of the linin reticulum. Both reticulum and interfilar substance are widely distributed through the nuclear cavity in the early prophase, and at this stage the spindle is very stout; in later stages, as the reticulum contracts and the interfilar substance passes to the poles, the spindle grows slenderer. In the late anaphase, when the spindle becomes shorter, it again crows stouter, figs. 15 and 16.
‘The second maturation spindle arises within the centrosome left in the egg at the close of the first maturation. At first it occupies but a small part of the cavity of this centrosome, but it grows rapidly until it fills the whole of it. The outlines of the mother centrosome then disappear and the spindle lies free in the sphere sub- stance. Here it grows rapidly in size, but never becomes more than half as long as the first maturation spindle, though it is relatively stouter. Its mantle fibres are not formed directly from a linin reticulum, since there is no vesicular nucleus, though they may possibly be formed from nuclear material which escaped from the germinal vesicle at the previous division.
During the prophase, the direction of the first maturation spin dle bears no con- stant relation to the egg axis. It may lie obliquely or even at right angles to that axis, figs. 9 and 10, but ultimately it moves into a radial position, fig. 12, e¢ seq.
The direction of the second maturation spindle, like that of the first, varies greatly, though in all cases it ultimately becomes approximately radial. As in Physa (Kostanecki and Wierzejski °96), the outer pole of the second maturation spindle les at the very point where the mid-body (Zwzschenkorper) of the first maturation spindle was formed. The second polar body is given off immediately under the first, so that the latter becomes separated from the surface of the ege and remains mounted upon the former. This happens irrespective of the initial direc-
KARYOKINESIS. 21
tion of the spindle, which always ultimately turns so that one pole lies immediately under the first polar body. If one may judge by the figures of many authors, this must be a phenomenon which occurs among a large number of animals.
4. Porar Bopirs.—Finally, the outer half of the mitotie figure, with a small amount of surrounding cytoplasm, protrudes from the general surface of the ege. The furrow separating the first polar body begins to form at the periphery and pro- ceeds toward the middle of the stalk connecting the polar body with the egg. In some cases the spindle seems to retain its full diameter, even when the cytoplasm has been completely constricted by the dividing furrow, fig. 22, as has also been observed by Kostanecki and Wierzejsky (96) in Physa. Afterward, the spindle itself becomes constricted in the middle, fig. 25; and the constricting ring of darkly staining substance finally cuts the spindle in two and itself becomes a spherical mid-body. Durig and after the separation of the first polar body, one first becomes aware of the fact that there is an egg membrane, which takes no part in the con- striction, but is lifted from the egg by the polar body, Plate II, figs. 22, 23, 28, 30 and 351.
The second polar body is smaller than the first and is separated from the egg in the same way as the first, a mid-body being formed, as shown in figs. 34 to 41. This mid-body is larger and persists longer than that of the first maturation, as Mark and Kostanecki and Wierzejski have also observed. When fully developed, it consists of a central granule and two surrounding spheres, the inner one small, dense and sharp in outline, the outer one large, less dense and indistinct in outline. The remains of the spindle can be seen running through this outer sphere as two cones, their apices being in contact with the inner sphere and their bases with the two nuclei.
The first polar body divides by mitosis into two, figs. 27, 28a, 32, 34 and 36, and each: of these may subdivide amitotically into a large number of cells, some of which are unequal in size and recall the macromeres and micromeres of developing ova, figs. 41, 45, 61, 69, 73 and 81. I have never seen the second polar body dividing.
II. Ferrinization.
1. EnrrANCE OF SPERMATOZOON.—Copulation occurs only at long intervals, perhaps once in the life time of a female, and the spermatozoa are stored after copulation in a tubular outgrowth of the uterus. Ova and spermatozoa meet in the uterus, and here the entrance of the spermatozoon occurs, though the later stages in the approach of the egg and sperm nuclei do not occur until after the capsules have been formed and deposited. In the examination of thousands of eges taken from the egg capsules I have never found one which was unfertilized and very few into which more than one spermatozoon had entered.
A mature spermatozoon is shown in fig. 17; there is a relatively large head with pointed apex, separated, in preserved material, by a clear space from the tail. I am inclined to regard a minute, darkly staining cap which covers the posterior end of the head as the middle piece. It is extremely small and appears to contain no
22 KARYOKINKESIS.
archoplasm.' A spermatozoon enters the ovum almost immediately after it reaches the uterus and while the germinal vesicle is still intact, fig. 5, e¢ seg. The sperm may enter at any point on the surface of the egg, except within a small area im- mediately surrounding the animal pole; usually, however, it enters near the vegetal pole. Polyspermy is exceedingly rare; one sometimes finds several spermatozoa. attached to an egg, and in a few cases two spermatozoa may be found penetrating the egg membrane or lying just within it, fig. 10, but only on one or two occasions have I seen two well-developed sperm nuclei within one egg. The pointed head of the spermatozoon bores through the egg membrane, figs. 18, 19, 20, though the tail does not enter. After the sperm head is well through the egg membrane several granules are found just behind the head; these are probably derived from the middle piece. Their number and arrangement is variable, but there are always more than two, so far as I have observed, and they are never grouped at the poles of a spindle. After its entrance, the head occupies such positions as to justify the belief
that it turns around, as is known to be the case in many other animals (ef. figs 18, 1955205521):
Foot (94 and ’97) has described in Allolobophora a number of dark round bodies which stain as intensely as the sperm head itself, and which lie on each side of the head or at its posterior end. These she calls the sperm granules and suggests that they may be formed from metamorphosed archoplasm. They are not constant in appearance and may be the result of degeneration.
Byrnes (99) has also observed in Lewax a number of darkly staining granules which accompany the sperm head. She suggests that they are derived from particles of chromatin constricted off from the sperm nucleus. Later they disappear and become scattered through the cytoplasm of the egg.
In the main the resemblance of these “sperm granules,” both of Ad/olobophora and Lzmax, to those which I have observed in Crepzdu/a, is striking enough. I can- not believe, however, that they are degeneration products in Crepzdula and for that reason, among others, have not adopted Foot’s name for them.
2. Tur Germ Nucret.—Immediately after the sperm head has entered the ege it is seen to be a pointed rod with three constrictions and four enlargements, having much the same size and shape as one of the 4-part chromosomes found in the meta- phase of the first maturation division, fig. 18. It soon grows shorter and thicker and becomes dumb-bell shaped, fig. 20, then nearly spherical, figs. 10, 21, and then irregular or amoeboid, figs. 39, 40. Up to this stage it has remained chromatic throughout, but from this time forward spaces filled with achromatic substance appear within it and it begins to grow vesicular. V. Klinckowstrém (96) and Van der Stricht (98) have observed a similar transformation of the sperm nucleus in Prostheceraeus and in Thysanozoon, the sperm head being first moniliform, then spherical, then vesicular.
‘Byrnes (’99) finds no middle piece in the spermatozoon of Lima and suggests that it may possibly be surrounded or overgrown by the sperm head.
KARYOKINESIS. 23
The egg nucleus is formed by the fusion of the chromosomal vesicles left in the egg at the close of the second maturation, as described on p. 14.
The further changes of the germ nuclei may now be briefly followed as far as the prophase of the first cleavage. The developments of both germ nuclei are entirely parallel, so that a single description will serve for both. As soon as the vesicular stage of each nucleus is reached the chromatin is found to be stretched through the nucleus in the form of a reticulum. figs. 56-41. As the nuclei enlarge the chromatin takes more and more the form of rounded masses, Plate III, figs. 44, 45 and 46, while the reticulum connecting the masses becomes extremely tenuous and does not stain. In short, there is at first a chromaten reticulum, which in later stages becomes a /27z” reticulum with the chromatin aggregated at nodal points. The chromatin masses differ considerably in size, fig. 45, and are at first quite solid. In later stages, figs. 49-53, these masses become hollow spherules. Those spherules which develop into chromosomes become connected together into a linear series, and either remain solid or at least have thicker walls than those spherules which take no part in the formation of the chromosomes. The further history of the chromatin will be taken up under the head of the first cleavage. As soon as the vesicular stage of each germ nucleus is reached there appears within it a single large nucleolus.’ This persists until a stage when the two nuclei come into contact, fig. 44, when it is usually dissolved in the nuclear sap, though sometimes traces of the nucleoli may be seen in later stages, e.¢., fig. 49.
3. EeG ANp Sperm Asters anp Spueres.—The history of the egg centrosome and sphere in the second maturation division has already been considered, pp. 16 and 17. At the same time that the egg aster is being transformed into the enormous ege sphere, figs. 32-36, a sperm aster has appeared and is undergoing a parallel trans- formation. The various stages in this process occur at approximately the same time in the two, though the sperm sphere and nucleus remain slightly smaller than those of the egg until the nuclei lie near each other. We may now follow in detail the origin of the sperm sphere. ;
After its entrance the sperm head lies among the yolk spheres in a small quantity of cytoplasm, while the granules derived from the middle piece lie just behind the head. There is at this stage no trace of astral radiations anywhere in the ege, except in connection with the first maturation spindle. The sperm nucleus lies in this position, near the periphery of the egg, without any trace of astral radiations near it, until the anaphase of the second maturation division. At this time the nucleus has become irregular or amceboid in shape and some distance from the nucleus, toward the center of the egg, the sperm aster appears. It is a noteworthy fact that no sperm aster appears until the sperm nucleus begins to absorb achromatic material, and this suggests that the two processes stand in some causal relation to each other. Furthermore, the fact that the two spheres are pro- portional in size to their nuclei, and that the sperm sphere remains smaller than
"Mark (’81) observed in an undetermined species of Limax that each of the germ nuclei contained a single nucleolus.
24 KARYOKINESIS.
the egg sphere as long as the sperm nucleus is smaller than that of the egg, lends further weight to this suggestion.
The earliest stage in the formation of the sperm aster which I have seen is shown in figs. 59 and 40. I have examined thousands of eggs of earlier stages, but have failed to find a sperm aster in any of them. The aster when first seen is a radiating figure in the cytoplasm, with several dark granules at its center. The number, position and size of these granules is not constant, and in later stages they greatly increase in number and stain less darkly than at first; there can be little doubt that they are identical with the granules derived from the middle piece. The sperm aster with the granules at its center ultimately becomes more rounded in outline and forms a large sphere from which radiating fibers proceed in all direc- tions. This sphere exactly resembles the sphere in contact with the egg nucleus, fig. 41.
From the time of their first appearance each of these spheres les close to its own nucleus, and they do not wander from these relative positions so that there is no possibility of confusing or mistaking them. During the approach of the sperm nucleus and aster to those of the egg, one or two small accessory asters appear in the egg, usually at some distance from the sperm and egg nuclei (figs. 42 and 45); these resemble the minute asters described by Mead (98), and Lillie (98) as “accessory asters.” They contain no centrosomes or large granules, and their origin at.a distance from the egg and sperm asters shows that they are independent of either of these. These accessory asters are present for a brief period only and then completely disappear.
At no stage in their development do the egg and sperm spheres show the com- pact and densely staining qualities which the spheres show throughout the cleavage stages; this added to the fact that there is a less perfect separation of cytoplasm and yolk during the fertilization than in the cleavage makes the study of these structures difficult, and this is especially true in the stages just before and after the fusion of the spheres. While designating these structures “spheres,” both because of their form and also because of the derivation of the ege sphere from the sphere left in the egg at the close of the second maturation, I would not be understood as positively homologising them with the “outer sphere” or ‘“ cortical zone” of authors.
4. Approach or GERM NucLEI AND SpHERES.—The egg nucleus and sphere remain at the upper pole, immediately beneath the polar bodies, and do not move from this position. The sperm nucleus and sphere move toward those of the egg in a path which is at first directed toward the center of the egg (“ Entrance path,” Roux), and then toward the egg nucleus (“ Copulation path”’). If the sperm enters near the lower pole, the course of the sperm elements is nearly straight through the egg from the lower to the upper pole; if it enters at any other point than the vegetal pole, the path is a curved one, the “entrance path” curving more or less sharply into the “copulation path,’ depending upon the distance of the point of entrance from the vegetal pole. In all cases the sperm elements approach those of
KARYOKINESIS. 25
~
the egg from the lower side, and during the prophase of the first cleavage the germ nuclei usually, though not invariably, occupy the same relative positions, the egg nucleus being above and the sperm nucleus below, figs. 42-55. The positions of the spheres relative to the germ nuclei is not perfectly constant, though the sperm sphere usually precedes the sperm nucleus and the egg sphere lies on the central side of the egg nucleus. The spheres remain distinct during the approach of the germ nuclei, one being quite as evident as the other, and neither showing any trace of degeneration. A number of yolk spherules are carried before the sperm into the protoplasmic area surrounding the egg nucleus and sphere, and thus it happens that several yolk spherules are usually found between the two germ nuclei and spheres, and more or less isolated from the remainder of the yolk. The germ nuclei first come into contact, as shown in figs. 44 and 45, and afterwards the spheres meet. inclosing still some of the yolk between them; the spheres then completely fuse. figs. 45, 46, 47, 49, 50, 51, #S.
Before fusion the spheres consist of masses of faintly staining granules, and a more or less distinct boundary line separates them from the remaining cytoplasm ; from this boundary a few fibres or rows of microsomes radiate. This boundary line is sharper in some cases than in others, but is always faintly marked. Immediately before and after the fusion of the spheres it can be seen that the coarse eranules in the spheres are nodal points in a very delicate reticulum, figs. 45-47 and 49-51. As soon as the spheres have fused, their substance surrounds the nuclei and spreads in a faintly staining mass into the cytoplasm above the nuclei and immediately below the polar bodies. A similar area of darkly stained protoplasm has been observed by Coe (99) in Cerebratulus (see his figs. 25-28), and is said by him to be derived from the germinal vesicle. In Crepzdu/a there can be no doubt that this area is derived from the egg and sperm spheres, though these in turn may be derived from material escaped from the germinal vesicle. All this time very faint radiations proceed from the periphery of the fused spheres, figs. 46, 49, 50. In Arenicola, according to Child (98), the germ nuclei, when they meet, are sur- rounded by an area of reticular cytoplasm from which radiations run into the sur- rounding substance of the egg. Child regards these radiations as possibly the result of the absorption of liquid by the germ nuclei, while the reticulum, he thinks, may indicate an accumulation of liquid around the nuclei.
In Crepzdula the spheres are present during the period when the germ nuclei are growing most actively ; they lie in close contact with these nuclei and appear to be associated with their rapid growth. I am inclined to regard them as the expres- sion of certain chemical and physical processes, taking place between the nuclei and the cytoplasm, rather than as structures of high morphological significance.
5. ORIGIN OF CLEAVAGE CEeNTROSOMES.—In several cases I have observed two large granules among the microsomes at the periphery of the spheres, from which stronger radiations proceed into the cytoplasm, but not into the spheres, figs. 47, 40, 51. These granules are but little larger than others in the peripheral layer of the spheres, and the radiations proceeding from them are but a trifle stronger and more
4 JOURN. A. N. S. PHILA., VOL. XII.
26 KARYOKINESIS.
perfectly centered. Nevertheless they are the only structures in the egg at this stage which at all resemble centrosomes, and I believe, though I cannot positively affirm it, that they become the centrosomes of the first cleavage spindle. In a slightly more advanced stage, figs. 48, 52, 55, unmistakable centrosomes are present ; they are no larger than the granules of the preceding stage, but the radiations are larger and more numerous, and they proceed im all directions from them. Those radiating fibres which are directed toward the germ nuclei come into contact with the nuclear membrane, which becomes infolded at this point, and at the same time a darkly staining, homogeneous fluid escapes from the nucleus thus forming a cone or half spindle, the base of which is applied to the nucleus, while the apex reaches to and surrounds the centrosome.
As soon as the undoubted centrosomes appear the fused egg and sperm spheres lose their boundaries, and their granules are either dissolved, or are scattered through the cytoplasm, figs. 48, 52, 55. The cleavage centrosomes are from the first independent of each other, and not until a later stage (figs. 54 and 55), is there any trace of a “central spindle” between them; these fibres grow out from each centrosome until they meet and fuse, just as MacFarland (97) has observed in the first cleavage of Pleurophylidza.
In view of the controversy as to the origin of the cleavage centrosomes in different animals, it is important to know what relation these centrosomes bear to the egg and sperm spheres of Crepzdu/a. Unfortunately no conclusive answer can be given to this question since the centrosomes do not appear until after the spheres have fused.t| There are certain evidences, however, which point to the conclusion that each sphere gives rise to one of these centrosomes. The evidences are the following :—(1) in fig. 45 a number of yolk spherules lie between the eee and sperm spheres which are here entirely separate; in figs. 46 and 47, the principal mass of yolk within the fused spheres probably marks the line of fusion between the two spheres; in fig. 47 a centrosome lies on each side of the principal aggregation of these yolk spherules, and therefore it is probable that one centrosome has arisen from that part of the fused sphere which was the sperm sphere, and the other from the half which was the egg sphere; (2) until the time of fusion each sphere is closely connected with, in fact partially surrounds, its own nucleus. Even after the fusion it can be seen, fig. 46, that a denser portion of the fused sphere is connected with each of the germ nuclei. Now, if the centrosomes arose, one from the ege
1 Since this was written more recent work on this subject has shown conclusively that centrosomes and spindles may arise separately in connection with each germ nucleus. If the recently fertilized eggs of C. plana are put into a 1 per cent. solution of sodium chloride in normal sea water for 4 hours, a perfect karyokinetic spindle, though about one-half the size of the usual cleavage spindle, appears in connection with the egg nucleus, although the latter may be separated from the sperm nucleus by almost the whole diameter of the egg. If the sperm nucleus is small and densely chromatic no spindle is formed in connection with it; if, however, the sperm nucleus has grown until it contains a consider- able quantity of achromatic material a perfect spindle may be formed in connection with it also; in such cases the two spindles usually lie close to each other and may form a tetraster. This experiment suggests that the contradictory observations of different investigators on different animals may find an
explanation in the varying rates of growth of the germ nuclei within the egg or in slight differences of the environment.
KARYOKINKSIS. 27 sphere, the other from the sperm sphere, we should expect to find a centrosome in connection with each germ nucleus and with no connecting central spindle between them. This is just what occurs. In figs. 50-51 the two centrosomes are so placed as to suggest that one is related to the egg nucleus and the other to the sperm nucleus, and in figs. 48-53 there can be no doubt about this fact. In no egg ex- amined is there a trace of a central spindle connecting the two centrosomes until after the centrosomes are in their definitive positions and the nuclear membrane is broken down at the poles of the spindle, figs. 54-55. Even though the centrosomes may lie in their definitive positions at an early stage, a thing which sometimes occurs (fig. 52), they are still quite independent, there being no central spindle fibres between them. This evidence, therefore, although not entirely conclusive, is favor- able to the view that one of the centrosomes of the first cleavage spindle comes from the egg sphere and the other from the sperm sphere.
Such a conclusion as to the origin of the cleavage centrosomes is at variance with all observations which have been made heretofore,’ and it is with much hesi- tation that I bring it forward without being able to demonstrate its truth in the clearest and most satisfactory manner. I have finally determined to publish these observations only after having spent several years in trying to get indisputable evidence upon this point, so far without success. However, the evidence, as far as it goes, points to the conclusion that both egg and sperm spheres contribute to the formation of the cleavage centrosomes.
In view of the fact that, in Crepzdu/a, egg and sperm centrosomes and spheres undergo parallel metamorphoses and that both spheres persist until their union, the commonly accepted view that the spermatozoon alone contributes to the cleavage centrosomes seems in this case highly improbable. Further, there is no particle of direct evidence in favor of this view; there is no sperm amphiaster as in many other cases; when the cleavage centrosomes first appear there is no central spindle between them, as would be the case if both were derived from a single sperm cen- trosome ; a centrosome usually appears in connection with each germ nucleus, which is also inexplicable on the supposition that both have come from the spermatozoon. These same facts are equally strong against the supposition that both cleavage cen- trosomes are derived from the egg centrosome.
On the other hand it is quite possible that both cleavage centrosomes are new formations, z. e., are not directly derived from the egg and sperm centrosomes, but have arisen independently of these and of each other, in the remains of the fused spheres. Apart from the evidence that one centrosome comes from each of the
1Tt most closely resembles the results of Carnoy and Lebrun (’97) on Ascaris, though it differs fundamentally from these in that these authors claim that the cleavage centrosomes arise from nucleoli, one of which comes from each of the germ nuclei.
Since the above was written Lillie’s (1901) complete paper on the maturation, fertilization and cleavage of Unio has appeared, and the account which he gives of the origin of the cleavage centro- somes in that animal is strikingly like my observations as to the origin of these centrosomes in Crepi- dula. In brief he finds that one cleavage centrosome arises in connection with each germ nucleus, that there is no central spindle between them and that they arise near or in the margin of the sphere sub-
stance. He does not consider that they are descendants of the egg centrosomes or sperm amphiaster, but that they are ege products of new origin.
28 KARYOKINESIS.
spheres, there is no reason to be alleged why both may not be new formations with- out genetic relationship to egg or sperm centrosomes, except the analogy of the cleavage stages, where a persistence of centrosomes in all stages can be clearly established.
In a former account of the fertilization of Crepzdu/a (Conklin °94) I described a form of “ Quadrille of the Centers,” in some respects similar to that observed by Fol (91) and Guignard ('91). In this account I expressly stated that I had not seen the centrosomes during the fertilization, but only the egg and sperm “asters.” My account of the persistence and approach of both the asters until they come into contact, I am now able to confirm. However, my account of their subsequent divi- sion into halves and the union of these halves by pairs to form the cleavage asters was incorrect. Judging by what I have since seen I am convinced that in my former paper I mistook lobulations of the egg and sperm spheres such as are shown in fig. 44, for division of those spheres, and other similar lobulations of the fused spheres, figs. 46-49, for the union of half-spheres to form the cleavage asters.
The present stand of the question of the centrosomes in fertilization is so well known that it demands no extensive treatment here. Following the publications of Fol, Guignard, Blane and myself, papers on this subject came “ fast and furious.” Boveri (95), Wilson and Mathews (795), Hill (95) and Reinke (95), showed that no quadrille occurred in the Echinoderms; Kostanecki and Wierzejski (96), Mac- Farland (97), and more recently Griffin (99), and Linville (1900), held that it did not occur among mollusks; Guignard’s work has failed of confirmation by other writers; Van der Stricht (95), who held that a quadrille occurred in Amphioxus, has been followed by Sobotta (97), who maintains that there is no quadrille in that animal, and Blane (95), who described a form of quadrille in the trout, has been followed by Behrens (98), who finds that both the cleavage centrosomes in that animal comes from the sperm; and so the quadrille went to its death.
On the other hand Boveri's (87-92) view that the cleavage centrosomes were introduced by the spermatozoon, and that “7z¢ zs the centrosome alone which incttes the division of the egg, and ts, therefore, the fertilizing element proper” (Wilson 96, p. 140), was eagerly championed by more than a score of writers; in fact this doctrine was much more cautiously held by Boveri than by many of his followers. However, there has been accumulating a body of evidence to show that the cleavage centrosomes do not, in all cases at least, come from the spermatozoon. Apart from the long known fact that cleavage centrosomes are present in parthenogenetic eggs, many observations have been made on fertilized eges which tend to show that these centrosomes may come from the egg centrosome or may possibly arise independently of either the egg or sperm centrosomes. I refer particularly to the work of Wheeler, Foot, Mead, Lillie, Child and myself. In almost every case so far observed there is a period, more or less prolonged, during which no centrosomes are visible (ef. Coe, 98, p. 455). In only a few cases is it affirmed that the sperm centrosomes can be traced without a break into the cleavage centrosomes.
So far as the mollusks are concerned there does not seem to be a single case in
KARYOKINESIS. 29
which the cleavage centrosomes are undoubtedly derived from the sperm centrosome. As to Physa, in which this origin is strenuously maintained, Kostanecki’s figures are capable of another interpretation than that which he puts upon them. All of his figures which show the two germ nuclei and the two centrosomes up to the time when the latter have taken their final position at the poles of the nuclei (his fig. 33a) show one centrosome in connection with each nucleus and nowhere in these stages is a central spindle shown, except in fig. 30, which shows a single fibre con- tinuous from pole to pole; even in the later stage, fig. 33a, there is no central spindle. Further, it is a significant fact that when the egg centrosome disappears the sperm centrosome also disappears (fig. 25-28), while the next stage figured (fig. 30) shows two large and well marked centrosomes, and in fig. 31 one of these lies in close connection with each of the germ nuclei.
In Pleurophylidia, according to MacFarland, the sperm asters and centrosomes disappear completely during the formation of the second polar body and for a rela- tively long period no centrosomes are present. After the germ nuclei are in con- tact the cleavage centrosomes appear, and since they frequently occupy positions similar to the sperm centers, the author thinks they are derived from these.
In Unzo Lillie finds that both egg and sperm centrosomes and asters completely disappear and that accessory centrosomes and asters also arise and disappear. Finally the two cleavage centrosomes arise independantly of each other and of any of their predecessors. ;
In Lemnea Linville finds that both egg and sperm centrosomes disappear for a time, but since the cleavage spindle first involves the sperm nucleus, he concludes that the cleavage centrosomes are of spermatic origin. His figures, however, do not bear out this interpretation ; fig. 6 shows the incipient cleavage spindle in connection with what is surely the egg nucleus, though he calls it the fused germ nuclei, (so far as | am aware the germ nuclei do not fuse in any mollusk.) Fig. 18, which is one of the earliest of his figures showing the cleavage centrosomes, shows one in con- nection with each germ nucleus and with no central spindle between them.
Boveri's figure of Pterotrachea (90. fig. 10), which is so widely copied in the text-books, shows one centrosome in connection with each germ nucleus and no central spindle between the two.
In other groups of animals the evidence in favor of Boveri's hypothesis is by no means conclusive, while much positive evidence has been brought against it. Among Zurbellarza I know of no single case clearly favorable to this view; (ef. Klinckowstrém °96), Van der Stricht (98), Gardiner (99), Van Name (99). Coe’s (99) work on Cerebratulus affords very good evidence that the sperm centers become the cleavage centers in that animal, and the same is true of Cheloplerus (Mead °97), and of 7halassema Griffin (99). On the other hand Foot ('97) has shown in a convincing manner that the cleavage centrosomes are new formations in Allolobophora and Child (98) holds the same position with regard to Arenzcola.
If all these accounts are to be believed, therefore, the cleavage centrosomes may come from the sperm, from the egg, or from both, and it is at once apparant that
30 KARYOKINESIS.
processes which vary so much can have no fundamental or general significance.t And even if all these accounts are not accepted, they show that the problem is a peculiarly difficult and complicated one and that it is still too early to formulate gen- eralizations with regard to it.
The evidence is certainly very convincing that in some cases both of the cleavage centrosomes come from the sperm, but in other cases the evidence that they do not have this origin amounts to a demonstration. This is shown in the clearest possible manner in phenomena of normal and artificial parthenogenesis in which of necessity the cleavage centrosomes must have their origin in the egg. Boveri recognizes par- thenogenesis as an exception to his generalization and indicates that in such cases: the egg centrosome may not degenerate. This view presupposes a fundamental dif- ference between amphigony and parthenogenesis such as does not actually exist. It is well known that in certain animals the determining causes of amphigonic or par- thenogenetic development are slight differences in extrinsic conditions ; for example there is no fundamental difference between the ova of the honey bee which develop parthenogenetically and those which are fertilized, and how purely accidental in this case are the causes which determine whether there shall be parthenogenesis or amphigony. There is no world wide distinction between these two methods of development and the differences as to the manner of origin of the cleavage centro- somes cannot be fundamental. If in some species the ege centrosome is capable of being preserved or reorganized, it 1s certainly quite possible that in others it may not degenerate at all. There is therefore no a frzore reason for supposing {that Boveri’s hypothesis is of general application and, as I have already attempted to show, it is not in accord with all the facts.
Certainly when one looks at the problem of fertilization from a general point of view, when one considers the universality of sexual reproduction, when one reflects upon the multitudes of exquisite adaptations which exist for securing the union of ege and sperm he will be loath to believe that the essential feature of fertilization is the addition of a centrosome to the egg cell or the supplying of a stimulus to its development which is not needed in all cases and can as well be supplied by changes in density, salinity, temperature, etc., as by the entrance of a spermatozoon.
III. CLeavacs.
I have already described the cell lineage or what may be called the external phenomena of cleavage in Crepzdula (Conklin ’97) and must refer to that paper for any detailed account of that process. I may be permitted here, however, to recall a few of the more important features in the early cleavage. In this gasteropod, as indeed also in all mollusks, the cleavage is of a peculiarly determinate, z. @., con- stant and differential, character.
The first cleavage is equal and divides the egg into two blastomeres which are approximately anterior and posterior in position; the second cleavage is also equal, dividing the egg into right and left portions, Plate V, figs. 80-88. Not only are the
" See foot-note.
KARYOKINESIS. 31
four cells thus formed (A, B, C and D,) equal in size, but they each contain about the same quantity of yolk. Two of the cells (B and D) meet at the vegetal pole in a polar furrow, whereas all four cells usually meet in a point at the animal pole.
From these four cells thus formed three groups or “quartettes” of small cells, without yolk, are cut off (Plate VI, figs. 89-96). These three quartettes (la—ld, 2a—2d, 3a-3d) form the whole of the ectoblast of the embryo. The fourth quartette (4a—-4d) consists of large cells containing yolk and one of the cells of this quartette (4d) is the mesentoblast and gives rise to most of the mesoblast and also to the posterior part of the intestine. The other three cells of this quartette (4a, 4h, 4c) are purely entoblastic. The first division of the first quartette (la—id) is very unequal, giving rise to four large ‘‘cephaloblasts’” and four small “trochoblasts” (figs. 93-96). The latter are peculiar in structure and history, being clear and non- granular as compared with the cephaloblasts; they divide but once and grow to a great size, giving rise to parts of the velum and head vesicle. The first subdivision of the cephaloblasts is also unequal (figs. 97, 98), giving rise to four small “apical cells” and four large peripheral ones which become the “basal cells” in the arms of a cross of ectoblastic cells, which lies with its center at the apical pole and one arm in each quadrant (figs. 99,100). The first division of the second quartette is nearly equal (figs. 96, 97), while at the second division four small cells arise which forms the “tip cells” im the arms of the cross (figs. 96-100).
Now as contrasted with these external phenomena of cleavage, which are chiefly concerned with cell boundaries, the internal phenomena consist of certain cyclical changes in the nucleus, centrosome and cytoplasm, each cycle being in the main like every other, though often differing in details. It is in these internal phe- nomena that the causes of determinate cleavage must be sought and to a study of _ these phenomena we now turn.
No sharp line of demarkation can be drawn between the fertilization and the first cleavage, since the two overlap, to a certain extent, in point of time. For con- venience, however, we may consider the fertilization ended and the first cleavage begun when the centrosomes have taken their definitive positions at the poles of the incipient mitotic spindle. Such a stage is shown in figs. 54-45.
1. THe NucLear Cuances Durinc CLeavace.—a. /udependence of Germ Nuclez.'—Until the metaphase of the first cleavage the chromosomes derived from the two germ nuclei are plainly separated into two groups, one derived from the egg nucleus, the other from the sperm, figs. 55-56, text fig. V. During the metaki- nesis no such separation is recognizable, but in the late anaphase the chromosomal vesicles fuse together into two groups and as the daughter nuclei become vesicular a partition wall is left between these groups, fig. 60, and text fig. VI. In the telo- phase this partition wall gradually disappears, persisting longest on the side of the nucleus next the centrosome, where a grooye marks its position, fig. $1; this groove usually disappears at the height of the nuclear “rest” or * pause,” but it appears again in the early prophase of the next division and in almost exactly
1 An abstract of this section appeared in the Biological Bulletin, Vol. II, 1901.
32 KARYOKINESIS.
ST GsenVie
Fie. VIII.
Fie. IX. Fie. X.
Fias. V-X.—INDEPENDENCE OF THE GERM NUCLEI OF CREPIDULA.—The duality of each nucleus is shown in the metaphase and telophase of the first cleavage (figs. V, VI), in the prophase and telophase of the second cleavage (figs. VII, VIII, X) and in an abnormal egg (fig. IX).
KARYOKINESIS.
Fic. XII. Fic. XI.
Fig. XIV. Fic. XIII.
Fie. XV. Fic. XVI.
Fics. XI-XVI.—INDEPENDENCE OF THE GERM NUCLEI OF CREPIDULA.—The dual character of each nucleus is shown especially well in the telophase of each division,
5 JOURN. A. N.S. PHILA., VOL. XI.
3 KARYOKINESIS.
the same position in which it was last seen, fig. 82, text figs. VIL and VIII. In this groove the new central spindle for the next cleavage lies and in the following division each half of the double nucleus is divided equally, frequently showing a double chromosome plate in the metaphase. In the succeeding anaphase and telo- phase each nucleus is again plainly double, being separated by a partition wall into two parts. The dual character of the cleavage nuclei has been observed in the telophase of every cleavage cell up to the 29-cell stage and in many cells up to the 60-cell stage, text figs. V-XVI.
It is very probable that the halves of these double nuclei descend in unbroken continuity from each of the germ nuclei and for the following reasons:
1. In the first and second cleavages the nuclear halves are distinct at all stages except during the metakinesis, and the relative positions of these halves cor- respond to those of the germ nuclei. Even in later cleavages the relative positions of the nuclear halves indicate that the one lying nearest the animal pole is probably from the egg nucleus and the other from the sperm, text figs. V-XVI.
2. In the first, second, third and fourth cleavages, and probably in all, the central spindle when first formed in the early prophase, lies in a groove between the nuclear halves, and hence in the only plane in which it could he if the nuclear halves are to be equally divided. Since successive cell divisions in Crepzdu/a alter- nate in direction, it follows, if the plane of nuclear division is always at right angles to the plane of contact between the two halves, that the nuclei or nuclear spindles must rotate at every cycle of division. This actually occurs, as a glance at the text figures will show; the rotation usually occurs in each nuclear cycle before the pro- phase but sometimes as late as the metaphase.
3. In certain abnormal cases blastomeres are found with two entirely separate nuclei in the resting stage; in other cases two entirely separate mitotic figures lie side by side in the same cell and in one such case, text fig. IX, there are thirty chromosomes in each of these spindlés, the same number which is found in each of the germ nuclei.
4. Finally there is always a single nucleolus in each of the gerin nuclei before their union, and in all of the cleavages, so far as I have observed, there are two and only two nucleoli present in the telophase, but during the resting period, particularly if it be prolonged, they may fuse into a single one. In view of current teaching with regard to the significance of the nucleolus this persistence of a definite number of nucleoli in each telophase is a somewhat surprising fact and may possibly indi- cate that there is a persistence of some structure which may act as a center for the formation of the nucleolus in each cell generation.’ Since the nucleolus itself is dissolved at the beginning of each mitosis, may not some achromatic structure, in which or around which the new nucleolus is formed, persist and be transmitted by
* Montgomey (’99) has compiled tables showing the number of nucleoli in the egg cells of 170 or more genera representing almost every phylum in the animal kingdom. As a result of this work he concludes that the number is not constant for a species, that it does not depend upon the amount of
yolk, mode of cleavage nor upon the manner of deposition of the egg, and that the facts do not warrant an attempt to explain the factors limiting the number of nucleoli.
KARYOKINESIS.
ro) Or
)
division to the two daughter cells? However this phenomenon may be explained, the fact that there is a single nucleolus in each germ nucleus before their union, and that there is a single nucleolus in each half of the dual nuclei during the cleavage, is additional evidence that the halves of these dual nuclei actually represent the germ nuclei.'
Such a case as that of Crefzdula indicates that the apparently single vesicular nucleus of the resting stage may really be double in character, and the fact that out of such a nucleus there may arise in the anaphase and telophase dual daughter nuclei shows that the germ nuclei may still preserve their individuality, though no trace of such separateness may be apparent at other periods. Further, it is possible, even in advanced stages of the cleavage to determine with considerable probability which part of a double nucleus is derived from the egg and which from the sperm, the egg half always lying nearer the animal pole than the sperm half (see text figures V-XVI).
This independence of the germ nuclei during the cleavage of Crepzdu/a is funda- mentally like the observations of Hacker (92, °95) and Riickert (95) on Cyclops, here also the germ nuclei do not completely fuse throughout the early cleavage, their independence being most clearly shown in the telophase. Riickert also finds evi- dence of a similar independence of the germ nuclei in the figures of Fol (79) on Toxopneustes and of Bellonci (84) and Kolliker (89) on Szvedon. Some of these figures referred to furnish very doubtful evidence. For example only one of Fol’s figures (pl. VII, fig. 7) shows a dual nucleus, while the figure in Kélliker’s textbook (fig. 56) is most probably a case of the indentation of the nuclear mem- brane opposite the centrosomes, a thing which frequently happens im the early pro- phase. Bellonci’s figs. 1 and 20 show an indentation on one side of the nucleus which may correspond to a division between the germ halves, but none of his fig- ures, with the possible exception of fig. 20, show a fusion of the chromosomal vesi- cles into two separate groups. Coe (’99) figures an indented nucleus in the telophase of the first cleavage of Cerebratulus (see his fig. 40) which probably represents the incomplete fusion of the germ nuclei in this animal. With the exception of Hacker and Riickert, none of the authors named call attention to these indented nuclei or suggest their possible significance, and I think it may fairly be said that Crepzduda affords the most satisfactory and convincing evidence of the independence of the germ nuclei which has yet been discovered.
These observations are intimately related to the important discovery of Herla (95) and Zoja (95) that the egg and sperm chromosomes of Ascarzs remain inde- pendent at least as late as the twelve cell stage, and this discovery was anticipated
1 More recent experimental work on Crepidula egg has shown that when the chromosomal vesicles are prevented from fusing a single nucleolus usually appears within each; in general, one nucleolus is found within each nuclear vesicle, and the fact that two are so generally found in the telophase is probably due to the fact that at this stage the nucleus consists of two vesicles, whereas the more com- plete fusion of these vesicles in later stages may lead to the formation of a single nucleolus. Such a
view would bring the size and number of nucleoli into relation with the size and number of nuclear vesicles present at any stage.
3 KARYOKINESIS.
by the hypothesis of the individuality of the chromosomes, first advanced by Rabl (85) and afterward ably defended by Boveri (87, °88, °92).
(6). Chromatin. At the beginning of the prophase of the first and second cleavages the nucleus contains a large number of rounded chromatin granules, which are connected together by a faintly staining linin network, figs. 45-55 and 62-64. These granules are at first solid bodies, but later become hollow spherules,’ figs. 45—52, and in these stages they all stain alike. Some of these spheres then become united in a linear series, to form the chromosomes, while the others (a large proportion of the whole number) take no part in the formation of the chromosomes and are finally dis- solved in the nuclear sap, or are transformed into linin threads. Those spherules which enter into the formation of the chromosomes again become solid and stain more deeply than the others (basichromatin) figs. 50, 51, 62, 63, while those which do not form chromosomes stain less deeply with nuclear stains and gradually come to take plasma stains, (oxychromatin.)
In the prophase of the third, fourth and fifth cleavages the chromatin exists in the form of a reticulum, figs. 70, 71, 74, 75, and not in the form of separate sphe- rules. In the rest preceding the prophase, however, this reticulum is formed of chromatin spherules as in the first and second cleavages, though these spherules are never so evident in later cleavages as in the first two. Some of the threads of this chromatin reticulum become chromosomes; others which show that they are com- posed of granules, fig. 70, stain much less deeply with nuclear stains, finally taking plasma stains only, and have no part in the formation of chromosomes, but are dissolved in the nuclear sap, or are transformed into linin.
This differentiation into two kinds of chromatin, one of which (basichromatin) forms chromosomes and the other (oxychromatin) does not, occurs in the early pro- phase ; in the preceding rest stages all the chromatin, both reticulum and spherules, stains alike, figs. 45-55 and 61-64 and 69-70. In the first and second cleavages the oxychromatin granules are scattered through the whole of the nucleus and most of them dissolve zz sztu, figs. 53,54, though some of them become attached to the mantle fibres of the spindle, fig. 55, and text figs. XVII and XVIII, where they are either transformed into spindle fibres or are dissolved, exactly as in the prophase of the first maturation. These dissolving granules sometimes remain hollow and in this case their morphology sufficiently identifies them with the chromatin spher- ules of preceding stages, figs. 49-52; in other eggs the dissolving granules become solid and gradually grow smaller and smaller until they disappear in an almost homogeneous nuclear sap, figs. 55 and 63. In some of the cleavages, particularly the second and third, I have observed that the basichromatin, in the form of a densely staining reticulum occupies that portion of the nucleus lying nearest the centrosome (‘* Pol” of Rabl °85), while the oxychromatin, also in the form of a reticulum, occupies the opposite half of the nucleus (‘Gegenpol” of Rabl ), figs. 62
* These hollow spherules with clear center and dark periphery directly reverse that common staining phenomenon, such as is characteristic of yolk spheres, where the periphery becomes clear, on
destaining, and the center remains dark. They have also been figured by Korschelt (95) in Ophryo- trocha and by Coe (799) in Cerebratulus in the prophase of the first cleavage.
Ns)
KARYOKINKESIS. ‘
= U
oO
and 70. These nuclei are very similar to those figured and described by Calkins (98) in Mocteluca and by R. Hertwig (99) in Actenosphe@rium, where the basi- chromatin is aggregated’ at one pole (** Hauptpol’’) and the oxychromatin at the other (‘*Gegenpol”’).
In all cases the oxychromatin granules or reticulum completely disappear as such, though this may not happen until after the spindle is well formed, e. ¢., fig. 55. Wilson (95) maintains that a portion of the chromatin (oxychromatin) is transformed into linin in 7oxopneustes, and Griffin (?99) holds the same view as to Thalassema; see also Lillie (1901, p. 250). I have no doubt that this is the case also in Crepzdula, where many of the oxychromatin granules are arranged on the linin fibres and are here dissolved and apparently transformed into the substance of the fibres (see text figs. XVII, XVIII). The further history of the achromatic sub- stances will be followed under the head of the mitotic spmdle and spheres.
The basichromatin is transformed into chromosomes in the manner already indicated (p. 56). In no nucleus in Crepzdu/a have I ever been able to find a single continuous spireme thread. The chromosomes are formed by the union into a linear series of the chromatin spherules or from portions of the chromatin reticulum, but from the first there is a large number of these segments, though I cannot determine whether the number is the same as the final number of chromosomes. Perhaps this method of formation of chromosomes without a preceding spireme is to be looked upon as a modification due to a precocious segmentation of the spireme.
In the early prophase of several cleavages, particularly the first division of the first quartette, the chromatin is aggregated into a dense mass at the center of the nucleus, leaving a peripheral zone inside the nuclear membrane which contains no chromatin, text fig. XXIX. Such nuclei resemble in appearance the “synapsis” stages (Moore, Montgomery) of spermatogenesis. This condition is the result of the aggregation of the chromosomes, a phenomenon which occurs in every prophase, while the resemblance to the synapsis is due merely to the persistence of the nuc- lear membrane for an unusually long time.
c. Separation of Chromosomes and Formation of Daughter Nucler. The chromosomes, which are at first widely scattered through the nuclear cavity, text figs. XVITand XVIII, are first drawn into the equatorial plate and then transported to the poles of the spindle in the usual manner.
The splitting of the chromosomes in the first cleavage, however, greatly resem- bles a heterotypic mitosis. In this division many of the chromosomes are shaped like rings, ellipses or triangles, and the parts of these figures lying in the equator grow thinner and thinner, the chromatic substance aggregating in the portions of the chromosomes turned toward the poles, until only a faint linin thread is left com- pleting the otherwise open rings or triangles, fig. 56. I am not sure that this type of division of the chromosomes in the first cleavage occurs in all eggs, sce I have found it in only a few cases and have been unable to find it in others of apparently the same stage (cf. figs. 56 and 57).
1 Montgomery (1900) has rendered these names into the convenient English terms “central pole” and ‘‘ distal pole,” which terms I shall adopt in this paper.
38 KARYOKINESIS.
The separation of the chromosomes coincides in point of time with the flow of the interfilar substance of the spindle to the spheres. The chromosomes move to- ward the poles until they come into contact with the spheres and even spread around them to a certain extent, figs. 59, 66, 67. Such a fact is irreconcilable with the theory that the chromosomes are moved solely by the contraction of the spindle fibres, as Wilson (95) and Griffin (99) have pointed out, and suggests that the movements of both interfilar substance and of chromosomes may be due to the chemotropic attraction of spheres and centrosomes, as Strasburger (93) maintains.
When the chromosomes have reached the borders of the sphere at the end of the spindle they do not enter into the sphere but spread somewhat over its surface figs. 59, 66, 67. In this position the chromosomes are rapidly transformed into vesicles, which grow larger and larger. These vesicles then fuse together and the nucleus becomes an apparently single vesicle, though divided by a partition wall as described above (p. 34). A reticulum of chromatin is then formed within the daughter nuclei, which probably arises from the walls of the chromosomal vesicles, and on each side of the partition wall there appears a single nucleolus, fig. 60. While these chromosomal vesicles are in contact with the sphere, the latter frequently becomes pear-shaped with the pointed extremity toward the chromosomes, fig. 67. In all cases the daughter nuclei have processes which extend partially around and even into the spheres, figs. 60, 81. Gradually, however, the processes disappear as the daughter nuclei increase in size and the latter finally become rounded on the side next the spheres, figs. Gl and 68. The significance of these processes of the nucleus which project into the spheres is not far to seek. The daughter nuclei are at this stage increasing their achromatic substance at a great rate, and the form of these nuclei at once suggests that this substance is absorbed in large part from the ¢ cells of Dytescus, as described by Korschelt (89), are similar im form, and perhaps in function, to this stage of these cleavage nucle,
spheres. The nuclei of the growing eg
Lillie (99) has observed that just before the “inner sphere” begins to expand, after the second maturation division in U/nzo, it is three-quarters surrounded by the chromosomes, and he suggests that there may be at this time a diffusion of chromatin into the sphere, the interior of which stains more darkly than before. According to my interpretation of the similar phenomenon in Crepzdu/a, the chromosomes are at this time absorbing substances from the spheres; not until much later does the “inner sphere” or centrosome again stain more deeply.
During this rapid growth of the daughter nuclei the spheres decrease some- what in size (ef. figs. 60 and 61, also 67 and 68), in spite of the fact that at this time sphere substance is collecting into the spheres from the astral radiations so that the decrease in the size of the spheres is not so great as it would otherwise be.
The chromatin reticulum which is formed in the daughter nuclei gives place in the next prophase to chromatic granules connected together by linin threads, figs. 61, 62, 70.
In early stages of the prophase, when the centrosomes are just moving into
KARYOKINKSIS. 39
position at the poles of the nuclei, the latter frequently put out one or more short blunt processes, text figs. VII, VIII, XII. These processes contain chromatin and sometimes dark masses which look like nucleoli. Unlike the nuclear processes of the anaphase, described above, these are usually found on the side of the nucleus away from the centrosome and nearest the mid-body. It is probable that they are withdrawn into the nucleus before the nuclear membrane is dissolved. Their sig- nificance is unknown.
This completes the account of the cycle of changes which the nucleus undergoes from one prophase of the cleavage to the next. With the exception of certain minor details, as has been pointed out, each cleavage is like every other in the matter of these nuclear changes. Apart from the equal division and distribution of the chromosomes in each mitosis, the most obvious and striking fact in this nuclear cycle is the escape of so large a part of the nuclear constituents into the cell body during mitosis and the reabsorption of a part of these by the daughter nuclei.
2. CENTROSOMES AND CENTRAL SPINDLES.—a. Centrosomes.—The origin of the centrosomes for the first cleavage has already been described in detail (pp. 25-30). These centrosomes are at first minute granules, quite independent of each other. A few fibres are inserted in them and radiate for a short distance into the cytoplasm, Some of these fibres grow toward the nucleus and form a cone or half spindle (figs. 48, 55), while others grow between the two centrosomes and unite them, thus forming a * central’ spindle” in the manner observed by Hermann (91), Drier (94) and MacFarland (’97). From the time the central spindle appears, the history of the centrosomes of the first cleavage is almost identically like that in the other cleavages so that the following description, unless otherwise specified, applies to any and all of the cleavages.
The minute centrosomes of the prophase (figs. 52, 53, 54, 63, 70, text fie. IV, a and 6) become much larger in the metaphase (figs. 57, 65, 72, 76, text fig. IV, c and @) and stain less deeply at the center. In the anaphase (figs. 58, 59, 66) the centrosomes continue to enlarge, the periphery alone staining with haematoxylin while the central area takes the plasma stain. Finally in the late anaphase and in the telophase the centrosomes become relatively enormous spheres (figs. 60, 67, 73, text fig. IV, e and /), frequently 6 to 8 win diameter. The peripheral layer or centrosomal membrane grows thinner and thinner until it reaches such a degree of tenuity as to be scarcely visible, ultimately breaking up into granules (figs. 68, 69, 73, 74). In all these respects the metamorphoses of the centrosomes throughout the cleavage are the same as in the maturation divisions.
The central area of the enlarged centrosome is at first apparently homogeneous (figs. 58, 66), but gradually minute granules begin to appear within it and then extremely delicate threads connecting them into a reticulum (figs. 59, 60, 66, 67, 73, text fig. IV). Sometimes one sees, as in figs. 60, 61, 62, one or two granules within the centrosome which are slightly larger than the others; but during the telophase all of these granules are extremely minute and stain very faintly with plasma stains. Gradually they grow larger until they fill the entire centrosome and
40 KARYOKINESIS.
their affinity for nuclear stains increases until in the resting stage of certain cleavage cells, these centrosomes look like small nuclei filled with amass of minute chromatin eranules, figs. 61, 69, 76, text fig. IV, £ In other cleavage cells these centrosomes with their contained granules remain much less conspicuous. Of all the early cleavage they are most plainly visible in the macromeres just before the formation of the first and second quartettes, figs. 69, 74,75. The cause of this difference in the appearance of the centrosomes in different cells depends largely upon their size and affinity for stains. The size of the centrosome is always proportional to that of the cell in which it hes; its affinity for stains, during the resting period, increases as it approaches the free surface of the cell, so that although it may stain faintly when a short distance from the surface (figs. 61, 62, 68) it stains deeply when in contact with it (figs. 69, 74). Upon these two factors then depends the relative conspicu- ousness of the centrosome during the resting period.
In all the cleavages which I have studied, with the exception of the first, the new centrosomes, and probably also the central spindles, arise within the mother cen- trosome, as in the case of the second maturation spindle. The origin of centrosomes for the second cleavage is shown in figs. 62 and 63, though the origin of the central spindle could not be clearly made out in this case. The origin of centrosomes and spindles for the third cleavage is shown in fig. 70, while those for later cleavages are shown in figs. 74,75 and 76. In all these cases the centrosomes appear as shghtly enlarged granules within the old centrosome. These granules stand at some distance from each other, and in no case in the cleavage have I seen the division of a single granule to form these two; they are, however, connected by the reticulum of threads and granules which fills the mother centrosome, and when the time arrives for the formation of a new mitotic figure the mother centrosome elongates, becoming shghtly elliptical m outline, the daughter centrosomes, as two enlarged granules, lie at the extremities of this ellipse, and the reticulum which fills the mother centrosome is drawn out into an irregular spindle shaped body composed of threads and granules, figs. 70, 74, 75, text fig. IV, 2, 2, 2. This elongation continues and the spindle shaped body becomes the central spindle (fig. 76), which in this case consists, not of straight fibres running from pole to pole, but of irregular and anastomosing fibres with granules at their nodes. The daughter centrosomes soon become surrounded by a little area free from granules, which is due to a halo of radiating fibres, so fine that few of them can be seen at this stage. This is the first appearance of the sphere (‘‘couche corticale”’) and it also arises, at least in certain cleavages, within the mother centrosome, figs. 70 and 76, the membrane of which may still persist at this stage.
At a slightly later stage these radiating fibres become very evident, and with the formation of the cones or half spindles, as described at the beginning of this account of the centrosome, we have the completion of the cycle of changes under- gone by the centrosome from one prophase to the next. In a word, the most im- portant features of this cycle are (1) the great increase in size of the centrosome and its transformation into a sphere filled with a reticulum of fibres and granules, and
KARYOKINKESIS. 4]
(2) the origin of the new centrosomes and central spindle from this reticulum; in some cases at least the cortical zone also arises within the old centrosome, so that the entire initial spindle of one cell generation arises within the centrosome of the preceding generation.
b. Central Spindles. In the case of the first cleavage the central spindle is formed after the centrosomes have taken their definitive positions at the poles of the germ nuclei, figs. 55 and 56. It first appears as a few fibres running from centrosome to centrosome in the line of contact between the egg and sperm nuclei. These fibres run independently from pole to pole and do not branch and show cross anastomoses with one another, so far as I have been able to observe; there are no varicosities or granules on them, as is the case in the later cleavages. The central spindles for the second cleavage are shown in surface view in fig. 82 and text fig. VIII, running from centrosome to centrosome over the surface of the nuclei and in the groove between the nuclear halves. In a section of a somewhat earlier stage (fig. 65), I have been unable to detect the fibres of the central spindle, though there is a clear area free from granules lying immediately over the nucleus and between the centrosomes in the position of the central spindle.
In the later cleavages the origin of the central spindle within the mother centro- somes can be plainly observed figs. 70, 74, 75, 76. The central spindle is in these cases a long drawn out reticulum with granules at its nodes. These granules gradu- ally disappear as the spindle elongates and their substance is evidently transformed into the central spindle fibres.
In Crepzdula, then, there appear to be two methods of origin of the central spindle : in the first cleavage the spindle arises in the cytoplasm between two inde- pendent centrosomes ; in all the other cleavages the centrosomes and central spindle arises as a unit structure within the mother centrosome; in the former case the fibres arise de novo between the centrosomes, in the later they arise as a centro- desmus (second maturation) or from the centrosomal reticulum (later cleavages).
3. Potar Rays Anp SprnpLe Fisres.—When first visible the polar rays are extremely short and delicate fibres and their presence is to be recognized rather by the clear area (“cortical zone’’) surrounding the centrosome than by the recognition of individual fibres, figs. 70, 76. Soon these fibres become larger and longer and are plainly visible, figs. 52, 55, 63. Those directed toward the nucleus become stouter and more numerous than the others, and the nuclear membrane is fre- quently indented where they come into contact with it, figs. 55, 54, 71. In some cases, however, the nuclear membrane is not indented, but is drawn out into a cone, the apex of which lies near the centrosome. Whether the membrane is invaginated or evaginated, there is in both cases an escape of achromatic nuclear substance at the poles, and it is due to this substance that the extra-nuclear fibres grow stouter and become covered with oxychromatin granules, text figs. XVII-XIX. In the first maturation division, not only the fibres of the extra- nuclear spindle, but also all the polar fibres are studded with these granules; in the cleavage, however, I have not observed them on the polar fibres. In early
6 JOURN. A. N. S. PHILA., VOL. XII.
42 KARYOKINESIS.
prophases the fibres of the extra-nuclear spindles are directly continuous with the linin threads of the nucleus, which they closely resemble in every respect, text figs. XVII and XVIII. Like the linin they branch and anastomose and are studded with oxychromatin granules. This resemblance is so striking that I cannot doubt that the fibres of the extra-nuclear spindles are really derived from the achromatic substance of the nucleus.
As in the maturation, so also in the cleavage there is an interfilar substance which fills the spaces between the fibres and which constitutes the greater part of the bulk of the amphiaster. This interfilar substance is probably derived in part from the hyaloplasm of the cell body and in part from nuclear sap containing dissolved oxychromatin.
Fic. X VII.—Prophase of first cleavage of Crepidulu. Fie. X VIII.—Prophase of second cleavage of Crepidula.
Throughout the metaphase the spindle-fibres are to a great extent concealed by this interfilar substance which fills in the whole space between them. In strongly destained specimens, however, the fibres can always be seen in the spindle. After the metaphase, however, no fibres can be seen crossing the dark zone which now surrounds the centrosome; both polar fibres and spindle-fibres appear to stop at the boundary of this cortical zone, or rather sphere. In the anaphase the structure of the sphere is such that one may be quite sure that neither polar nor spindle-fibres run through it, figs. 58, 59, 60, 66, 67, 68. In both metaphase and anaphase the polar fibres are not always centered on the centrosome, and if they were continued in a straight line through the sphere some of them would not touch the centrosome at all, figs. 57, 58, 60, 65, 67.
Just before the chromosomes reach the boundary of the spheres the mitotic eure is cylindrical in shape and consists almost entirely of interzonal filaments, figs. 58, 66. As soon as the chromosomes have reached the spheres and are transformed into vesicles, figs. 59, 67, the spindle again becomes wider in the middle than at the ends and contains many fibres which do not reach from pole to pole.
KARYOKINESIS. 43
The spindle greatly increases in length from the prophase to the telophase. R. Hertwig (99) has observed that in Actznospherium the spindle more than doubles in length during this period, and in Crepzduda the lengthening is nearly as ereat. The shape of the spindle varies greatly from prophase to telophase, being largest at the equator in the prophase and smallest in the telophase.
Mip-sopy.—When the new cell-wall is formed the spindle is constricted in the middle and a very remarkable mid-body (Zweschenkorper) is formed. This mid- body is elliptical in outline, and is surrounded by a dark area from which radiations proceed in all directions; into this dark area the cell-membrane and the two halves of the spindle enter, fig. 60. This mid-body is for all the world like a centrosome with its surrounding sphere and aster, and recalls Watase’s (95) comparison of the mid-body to an intercellular centrosome. This apparent resemblance is still further supported by the fact that the mid-body in this case becomes a hollow sphere before it finally disappears, fig. 61, just as the centrosome does.
The mid-body is surrounded by a darkly staining substance which resembles the sphere substance. This recalls Moore’s (95) observations on the larval Salamander, where he-finds a mass of archoplasm on each side of the mid-body, also Kostanecki's (92) statement that the mid-body is formed from granules of the sphere (archo- plasm). Kostanecki (97) has observed a mid-body in Physa consisting of a ring around the central spindle-fibres, from which radiations proceed. In some cases this ring divides through the middle into two rings. A similar ring is called by Heiden- hain “ Zel/nabel.’ Moore and Meves have seen mid-bodies connected with the centrosomes around the nucleus. as is plainly the case in the third cleavage of Crepidula (see fig. 73).
The cell-membrane adjoining the mid-body is thicker and more protoplasmic than at the periphery, and is in process of formation at this place. The mid-body persists through the whole of the resting period and until the prophase of the next succeeding division when it gradually disappears. As long as it is present there ean be no doubt as to protoplasmic continuity between the daughter cells.
4. SpuEres.—Before the nuclear membrane is indented, the centrosomes are surrounded by a clear area consisting of a halo of radiating fibres, figs. 63, 70, 76. This condition may exist even within the mother centrosome (see fig. 70). This clear area is the first appearance of what I shall call the sphere (‘‘ outer sphere” of Lillie, “couche corticale” of Van der Stricht). When the nuclear membrane is dissolved at the poles substances escape from the nucleus into this area surround- ing the centrosomes. At the same time hyaloplasm from the cell body is probably drawn in through the astral rays into the same area. There is thus a commingling of hyaloplasm and chromatic nuclear sap which constitutes the interfilar substance of the aster. There is at this stage no clearly marked sphere, since the central area of the aster is in no way delimited from the surrounding radiations.
In middle stages of mitosis it is difficult. even in thoroughly destained speci- mens, to trace the polar rays and spindle-fibres through the interfilar substance to the centrosome. In the anaphase the interfilar substance of spindle and aster
44 KARYOKINESIS.
collects into the central area surrounding the centrosomes, and this area, thus delim- ited from the surrounding plasm, is the sphere; at the same time the spindle-fibres again become plainly visible while a reticular or alveolar structure appears within the spheres, fig. 58.
In the late anaphase the spheres become much larger and are bounded by a layer of microsomes from which fibers radiate. The interior of the spheres is composed of a fine reticulum with nodal thickenings, and the whole sphere stains much less densely than in earlier stages. Finally in the telophase the spheres reach their greatest size and become filled with granules, the reticulum being scarcely visible, or disappearing altogether (figs. 61, 68, 73).
During the whole of the resting period the spheres persist, usually pressed close to the cell-membrane, and as long as the centrosomes remain in them they preserve a regular form (figs. 68, 69, 75, 74, 76). They are composed of coarse granules, which stain deeply with plasma stains, and they are sharply bounded by one or more layers of microsomes. As soon as the daughter centrosomes and central spindle arise from the mother centrosome, they migrate out of the sphere and the latter at once begins to lose its regular form. It becomes ragged in outline and is finally flattened out to a thin layer of densely staining granules immediately under the cell-membrane (figs. 63, 65, 70, 71, 72, 73, 74, 76).
These granules, the remains of former spheres, can frequently be recognized through -two generations of cleavage cells; ¢. g., the spheres which appear in the second cleavage (figs. 68-72) can still be recognized after the completion of the third cleavage, located in the first quartette of micromeres (figs. 73 and 74, see also figures of entire eggs in Plates V and VI). From the time when the daughter centrosomes issue from the spheres the latter are degenerating structures, and although their remains may persist for a surprisingly long, time they ultimately disintegrate and are apparently dissolved in the cytoplasm.
To sum up the history of the spheres: we find that they arise around the cen- trosomes at a very early period in the mitosis, in some cases within the mother centrosome. With the disappearance of the nuclear membrane at the poles of the spindle they are invaded by an interfilar substance; they have no clearly marked boundary. In the anaphase and telophase the spheres greatly enlarge, but their growth is always proportional to the size of the cell in which they are found. They are largest in the anaphase just before the chromosomal vesicles begin to form and they probably contribute to the growth of the daughter nuclei. At first they have a delicate radiating structure, this gives place to a homogeneous condition, and this to an alveolar or reticular one; finally, in the rest stage they are granular. Their fragments persist long after the daughter centrosomes have moved out of them, and they ultimately dissolve and disappear in the cytoplasm.
‘ In surface views of entire eggs the sphere may seem larger in the resting stage or early prophase than in the telophase, e. g., figs. 81 and 82, 86 and 87, etc.; this is due, as sections show, to a flattening
2 c ¢ 5 2; ° oe Z 2 7 5 of the sphere against the cell-membrane and a spreading of the sphere substance through the influence
of the astral rays, and not. to an actual increase in its volume (figs. 71, 72, 76).
KARYOKINESIS. eg
IV. GENERAL CONSIDERATIONS AND COMPARISONS.
I propose to give in this section a brief synopsis of the changes which the nucleus, centrosome and sphere undergo during the whole cycle of division in the mollusks which I have studied; to compare these observations with closely related ones in other animals and to indicate the general conclusions to which these obser- vations lead.
1. Tae NucLevus puRING THE CycLe or Divistoy.—The history of the nuclear changes during the cycle of division may be summarized as follows: (1) The chromosomes, consisting of chromatin inclosed in a linin sheath, divide and move to the poles of the spindle where they partially surround the spheres. (2) Here they become vesicular, the interior of the vesicle becoming achromatic, though frequently containing a nucleolus-like body, while the wall remains chromatic. (3) These vesicles continue to enlarge and then unite into the “resting nucleus” ; the nuclear membrane is composed of the outermost walls of the vesicles, while the inner walls stretch through the nucleus as chromatic partitions; the chro- mosomal vesicles from the egg and sperm nuclei remain distinct longer than those from the same nucleus. (4) The chromatin of these inner alveolar walls then aggregates into threads, giving rise to a ‘chromatic reticulum,” though the linin still preserves, for a time at least, the alveolar structure. (5) The chromatin of these threads then aggregates into spherules, which are connected together by linin threads; these spherules vary in size, and at first all are solid and stain alike. (6) They then become hollow and are differentiated into oxy- and basi-chromatin. (7) In the first maturation each of the basichromatin spherules, or bodies, grows into an individual chromosome; in the cleavage the basichromatin spherules unite into several linear series, thus forming a segmented spireme. (8) The oxychromatin spherules grow smaller and some are dissolved in the nuclear sap while others are arranged in series on the linin threads into which they are transformed; these threads with attached spherules form the spindle fibres. (9) During the differentia- tion of the chromatin the nucleus swells in size and the membrane becomes less chromatic, while the nuclear sap becomes more so; the nuclear membrane then dissolves at points opposite the centrosomes and linin, oxychromatin and nuclear sap here escape. (10) The spindle, which at first fills the entire nuclear cavity, then grows longer and slenderer and contains an interfilar substance; the nuclear membrane entirely disappears; the equatorial plate stage is then reached and the eycle is complete. In a word, the daughter chromosomes absorb achromatic sub- stances, and unite to form the nucleus, within which the chromosomes and spindle of the next division arise, while nuclear sap and dissolved chromatin escape into the aster and cell body.
Taking up now in more detail some of the individual steps in this cycle:
(a) Formation of Chromosomal Vestcles.—Growth of Daughter Nuclet.— When the chromosomes have reached the ends of the spindle, and in some abnormal cases even before this (see text fig. IX), they begin to absorb achromatic material and to swell into spherical vesicles. Such vesicles are found generally, if not uni-
46 .KARYOKINESIS.
versally, in the early divisions of ova, though they are not usually found in other mitoses. What is the cause of this difference? It occurs to me that it may be due to differences in the size and in the rapidity of division of blastomeres as compared with tissue cells.’ The following observations favor this view : — The chromosomal vesicles are proportional in size to the size of the cell (quantity of cytoplasm) in which they lie. The daughter chromosomes which go to the two poles of the spindle are always equal in size however unequal the cell division may be, until the time when the daughter cells are separated by the new cell wall. Immediately after this separation a difference appears in the size of the vesicles in the two cells, if the division was unequal, the larger cell containing large chromosomal vesicles while in the smaller cell they remain small or do not show the vesicular structure at all.
The chromosomes which go into the polar bodies do not appear vesicular at any stage, though after the division of the first polar body they fuse into a single nucleus in each cell which contains very little achromatic material. The smallest cells in the early stages of cleavage are the ‘“ trochoblasts” (fig. 97, la’-ld°); these cells do not again divide for a very long period, and in them the chromosomal vesicles are at first very small. Chromosomal vesicles appear in the anaphase of all the other cleavages, but as the cleavage advances and the blastomeres grow smaller these vesicles become less and less apparent.
From these observations I conclude that in large cells where divisions succeed one another at short intervals the chromosomes begin the growth characteristic of the daughter nuclei, z.¢., the absorption of substances from the cell body, before they have fused together, whereas in small cells or cells which divide only at. long intervals the chromosomes fuse before the absorption of achromatic material begins.
After the fusion of the chromosomal vesicles to form the daughter nuclei, the latter continue to absorb achromatic material, growing larger and larger, until the prophase of the next division. A part at least of the achromatic material absorbed is derived from the sphere which in turn contains interfilar substance of the spindle and aster. This recalls the conclusions of O. Hertwig (75), in which he points out that in the formation of the daughter nucleus the chromosomes absorb “ Kernsaft” and become vesicular, the process being the reverse of what occurs in the beginning of division, when “Avernsaft” is set free into the cell body. A similar view was held by Butschli (’76).
In the growth of the nucleus the nuclear membrane has the properties of a semi-permiable membrane, z.¢., substances pass readily through the membrane in one direction, but not in the other. Reinke (1900) has suggested that the nuclear ground substance is a diosmotic material, which, by taking up substances from the cell, produces a substance of higher osmotic pressure. When, on the other hand, the nuclear membrane dissolves and the ground substance of the nucleus mingles with the fluid substance of the cell, the peripheral layer of the latter assumes the
‘ Flemming (92) formerly held that all chromosomal vesicles were artifacts. Now that they have
been found, however, in so large a number of ova, prepared by the best modern methods, such an idea cannot be maintained.
KARYOKINKSIS. 47
role of a semi-permeable membrane and thereby the swelling of the dividing cell is produced which Reinke calls ‘‘ mitotic pressure.’ The manner of growth of the nucleus, its turgescence and the infolding of its membrane in the prophtse preclude the idea that the nuclear membrane is full of pores as held by Carnoy, Watase, and formerly by Reinke, and indicate that the growth of the nucleus is a phe- nomenon of diosmosis.
However unequal the division of the cell body may be the daughter nuclei are at first entirely equal, but the subsequent growth of the nucleus is proportional to the quantity of cytoplasm in which it lies; this is shown not only in the cleavage of the egg, but also in the formation of the polar bodies. The nuclei of the polar bodies rarely become vesicular but remain chromatic throughout. The fact that the size of the nucleus is proportional to the quantity of the cytoplasm in which it hes indicates that the achromatic substance absorbed by the nucleus is also propor- tional in quantity to the volume of the cytoplasm.
It sometimes happens, especially in eggs in which more than the normal num- ber of centrosomes and asters are present, that some or all of the chromosomal vesi- cles do not fuse, but remain distinct through the whole of the resting period. In such cases each of the vesicles behaves like a miniature nucleus, absorbing achro- matic material and forming a network of chromatin either within the vesicle or on its walls, In this growth and differentiation the vesicles keep pace, step by step, with the normal nucleus, so that one must regard the resting nucleus as virtually composed of vesicles, though their union may be so intimate as to hide this structure. The resting nucleus is not, therefore, a single structure any more than is the equato- rial plate. It is composed of units, each of which, so far as known, has the properties of the entire nucleus, and the union of these vesicles into a single one may be con- sidered as a secondary character. It is altogether probable that the chromosomes, and hence the chromosomal vesicles, preserve their identity throughout the resting period, and I venture the suggestion that the daughter chromosomes will be found to arise within the chromosomal vesicles, as the daughter centrosomes, or centrioles, arise within the mother structures.
(6) Chromatic Differentiation ; Solution of Oxychromatin and Nuclear Membrane.—In the early prophase of each division in the mollusks which I have studied, the chromatin becomes sharply differentiated into oxy- and basi-chromatin (Heidenhain). This differentiation occurs before the solution of the nuclear mem- brane, but at a time when the nucleus is growing rapidly in size and is therefore actively absorbing substances from without. This suggests that the rapid absorption of cell substance and the differentiation of the chromatin are associated, but whether this absorption is the cause or the result of the chromatic differentiation, I am un- able to determine.
The solution of the nucleoli usually precedes that of the oxychromatin spherules and the nuclear membrane, but in the case of the first maturation the enormous nucleolus is thrown out into the cytoplasm before it is completely dissolved. Many oxychromatin granules are not dissolved in the nuclear sap,
48 KARYOKINESIS.
but contribute directly to the formation of the linin threads and spindle fibres, as Wilson (°95) and Griffin (’99) have found to be the case in Zovopneustes and in Thalassema. This is especially the case in all small nuclei, whereas the larger the nucleus the greater the quantity of oxychromatin which dissolves and passes into the cytoplasm.
The solution of the nuclear membrane goes on coincidently with the solution of the oxychromatin, so that it seems probable that the causes of the two are the same. Before its solution the membrane changes its staining qualities, becoming more and more plasmatic in reaction. The membrane is in all cases first dissolved at points opposite the centrosomes. In this process the membrane undergoes one of two modifications: either (1) it isdrawn out toward the centrosomes into a cone-like figure, or (2) it is indented opposite the centrosomes. Both of these methods may coexist in the same animal, though one or the other is usually predominant. Among mollusks of all classes, the membrane is usually indented. The difference between these two methods is not great, depending upon the time at which the membrane is dissolved and upon the rate of outflow of nuclear substance; if the membrane is thin and dissolves early a cone is formed; if it dissolves slowly and only after a considerable quantity of nuclear substance has escaped, it becomes indented. The strength of the nuclear membrane in Cyefzdu/a is shown not only by the degree of indentation which it suffers before it is completely dissolved at the poles, but also by its long persistence in the equator of the nucleus (see figs. 84 and 88). Even when the membrane persists for a long time and becomes deeply indented at the poles it need not be supposed that pressure is brought to bear by the polar fibres or by other means upon the membrane; on the contrary, the indentation is chiefly or entirely due to the escape of nuclear sap and the consequent collapse of the nuclear wall.
The infolding or outfolding of the nuclear membrane at points opposite the centrosomes is a very common phenomenon among all classes of animals. It would be useless to attempt to summarize all the observations on this point, and I shall refer to only two recent works which touch upon this subject :—
Montgomery (98) has observed a cone-shaped protrusion of the nuclear mem- brane opposite each centrosome in Pexdatoma. These cones contain a dark sub- stance which he believes to be of nucleolar origin.
Fischer (99) interprets the openings at the poles of the nucleus as due to a greater growth of the nucleus at these pots. The fact, however, that it first occurs opposite the centrosomes and in connection with the formation of the half spindles, indicates that the opening in the nuclear membrane is due rather to the solvent action of some substance which diffuses to or from the centrosomes.
(c) Escape of Nuclear Substances; Aster and Spindle Formation.—\t may be considered certain that the infolding (or outfolding) of the nuclear membrane at points opposite the centrosomes is due to an outflow of nuclear substance at these points. This is conclusively shown by the fact that the linin reticulum, with its attached chromatin granules, here extends outside the nuclear wall nearly to the
KARYOKINESIS. 49
centrosome and forms the extra-nuclear portion of the spindle. The aster also stains more deeply after this outflow than before and very like the chromatic nuclear sap.
The shape of the spindle depends in part upon the degree and stage of this nuclear outflow. The spindle is at first as wide at the equator as the entire mother nucleus, but as the flow of nuclear substance toward the poles continues it grows longer and slenderer, the centrosomes at the same time moving farther and farther apart, until in the late anaphase almost the whole of the interfilar substance has moved out of the spindles into the spheres.
Whether or not the spindle fibres and linin threads exist as such in the living cell or are artefacts must still be left an open question. It can scarcely be doubted, however, that they do represent substances which are different from the surround- ing materials of the cell, and this is after all the important thing. That the spindle- fibres and especially the connective fibres sometimes show considerable elasticity and rigidity has been pointed out repeatedly by those who hold that the centrosomes are pushed apart by their activity. Nowhere is this better shown than in the first maturation of Crepedu/a, where in the shortening of the spindle the fibres are bent and kinked and the chromosomes at the outer pole are pushed clear through the polar body into contact with the opposite cell wall. In spite of this, however, it seems to me very questionable whether the spindle fibres are anything other than a fluid more viscid than the surrounding cell substance.
I agree with those authors (Butschli, Fischer, Rhumbler, Wilson), who hold that the astral rays represent diffusion streams in the cytoplasm, rather than a stable system of fibres. There are certain evidences that the astral rays are composed in the main of cytoplasmic material, principally hyaloplasm or interalveolar substance ; chief among these is the fact that the aster is always proportional in size to the extent of the cytoplasmic area which comes within its influence, a fact which Wilson (96) emphasized and which I also (94) pointed out and have since had abundant opportunity to verify. But while the aster and astral rays are in the main composed of hyaloplasm it is probable that in normal mitoses certain nuclear substances enter into their formation. In the mollusks which I have studied there can be no doubt that certain achromatic substances from the nucleus and spindle flow into the aster and at the same time the central area of the aster as well as its rays stain more deeply than the hyaloplasm throughout the cell body. Whether there may not be a centrifugal movement of escaped nuclear substance along the astral rays as well as a centripetal movement of the hyaloplasm must be left an open question.
In this connection I must refer to one of the first observations ever made on indirect nuclear division,—that of Auerbach (’74) on the living eges of certain nematodes. He observed the double suns (asters) with their connecting stalk (spindle) and supposed that they were formed by the collapse of the nucleus and the passing out of nuclear sap into the cytoplasm, “the astral radiations being merely the expression of the paths along which fine streams of nuclear sap pass out into the protoplasm.” Biitschli (76) also observed the passage of nuclear sap
7 JOURN. A. N.S. PHILA., VOL. XII.
50 KARYOKINESIS.
into the cytoplasm during nuclear division; in Caucudlanus and Nephelzs this loss of nuclear fluid may amount to as much as two-thirds of the volume of the unal- tered nucleus. Biitschli held that this fluid escaped at the two poles of the nucleus and accumulated in the central areas (asters), from which it radiated into the cell body. Further Biitschli observed that the more a daughter nucleus grows, the more the central area of the neighboring radial system diminishes, whence he inferred that the latter furnishes material for the growth of the former. (See Mark, ’81, p. 321.)
In 1892 Biitschli reversed his former view as to aster formation, holding that it is due to a flowing of plasma into the spheres or centrosomes and not from them. He supposed that the centrosomes attracted substances dissolved in the enchylemma as a hygroscopic substance attracts water, and that the diffusion movements thus produced cause the astral radiations. Although Biitschli in his 1892 work, and since 1n 1898 and 1900, maintains that the astral radiations are due to an attraction exerted by the centrosome, he expressly stated in the first mentioned work that it is unimportant whether the diffusion streams move in one direction or the other (z. e., centrifugally or centripetally).
I fully agree with Biitschli that the astral rays are the expression of diffusion streams. In the process of diffusion the commingling substances may move in opposite directions at the same time, and it is quite possible that the balance of flow between the centripetal and the centrifugal diffusion movements may lead to a cen- trifugal flow at one time and to a centripetal flow at another.
Rhumbler (96, 99) has also developed an elaborate theory of aster and spindle formation which is based in the main upon this view of Biitschli’s. He holds that the astral rays are reducible to tension on the alveolar radii; this tension being due to the fact that the centrosomes, and later the nucleus, take up fluids from the sur- rounding plasm. He also holds that the two spheres exert a pull on the nucleus which leads to the formation of the spindle and to the escape of nuclear sap into the equatorial plane of the cell, where the division wall will form.
I accept Rhumbler’s views as to the flow of cell and nuclear substances toward the centrosome, but cannot agree with him that the nuclear sap escapes largely or entirely in the equatorial plane. Much of the nuclear sap as well as the oxychro- matin and linin escapes at the poles of the nucleus and, although nuclear contents escape into the cytoplasm in all directions when the nuclear membrane is com- pletely dissolved, there is no evidence in the cases which I have studied that this has to do with the formation of the division wall.
Fischer (99) holds that the spheres of animal eggs are to be explained as substances escaped from the nucleus, and he suggests that the astral rays may be normally formed by the diffusion of substances from the nucleus into the cytoplasm and the production there of non-soluble substances. He considers that such radi- ations are not persistent structures, but that they may appear and disappear re- peatedly during the course of a single division. I agree with Fischer that there is an escape of nuclear substance at the poles of the nuclei, and that the astral rays
KARYOKINKSIS. 51
represent diffusion streams through the cell body, but Iam not sure that I under- stand him when he says that the rays are composed of non-soluble substance, since they certainly disappear (as he also maintains) either by dissolving in the cyto- plasm or by being absorbed into the sphere.
Meves (’99) criticises Fischer's views on aster formation by saying that such rays as Fischer describes could not grow interstitially, as normal rays and spindle fibres are known to do; sometimes also extensive rays appear in the anaphase, long after the mingling of nuclear and cytoplasmic substances. If, however, these rays be considered as diffusion streams to or from the centrosome, in the sense of Biitschli, these criticisms lose most, if not all, of their force. Finally, that the astral rays are not fixed structures stretching between the centrosome and the cell membrane, as Heidenhain and Kostanecki hold, is shown by the fact that in many mitoses the spindle is free to turn and move through the cell, and yet the astral rays show neither twisting, bending, nor distortion. This is shown especially well in the first maturation, and in the first three cleavages of Crepzdula, where there are considerable movements of the amphiaster even after the metaphase ; but in no case is there a corresponding bending of the rays, as there would be if these were fixed structures (see observations of Ziegler, Lillie, e¢ a/., Part II, Sec. II).
(d). Chromatic Elimination.—\n the maturation and early cleavages of the eggs which I have studied, the total amount of chromatin which is transformed into lmin or dissolves and escapes into the cell body is greater than that which goes to form the chromosomes; the amount of cytoplasm in the cell is also noticibly greater after the nuclear membrane is dissolved than before.
In this connection other observations of a somewhat similar character may be recalled. Almost all persons who have studied the maturation of the egg, have commented upon the large quantity of nuclear material which is set free into the cell body during the first maturation division. In the starfish, according to Wilson (95, p. 458), at least nine-tenths of the chromatin is thus set free. Gardiner (798, p- 97) estimates that in the egg of Polycherus not more than one five-hundreth part of the chromatin which is present in the germinal vesicle goes into the chro- mosomes, all the rest being thrown out into the cell. Most observers agree in identifying as chromatin this nuclear material which escapes into the cell body, though in most cases it stains less deeply than the chromosomes and its subsequent dissolving shows that it must be different from the chromosomes, which never dis- solve. Gardiner (98, p. 98) argues that there must be two kinds of chromatin, the one soluble, the other not, and Griffin (99) believes that the soluble chromatin arises as a nuclear reticulum which at first takes plasma stains and later nuclear ones.
Boveri's (92 and °99) observations on the diminution of the chromosomes in the somatic cells of Ascarzs may be recalled in this connection. In this case the ends of the chromosomes pass into the cytoplasm during the mitosis and there gradually undergo solution or disintegration. This case, however, differs greatly from that of Crepzdu/a since it occurs only in differentiation of somatic cells, whereas in Crepzdula the outtlow of nuclear material occurs at each and every
Or
52 KARYOKINESIS.
mitosis. Hacker (97) has described an elimination of nuclear constituents in the Keimbahn of Cyclops; in the first cleavage a large number of granules (“ekto- somes”), which Hacker considers escaped nucleoli, collect around one attraction sphere but not around the other. This process is repeated in subsequent cleavages, the cells in which the ektosomes appear marking out the Aezmbahn. Finally, in the division of the genital cells the ektosomes are found around the entire spindle figure. The elimination of the ektosomes in Cyclops, like the diminution of the chromosomes in Ascarzs, differs fundamentally from the chromatic elimination in Crepidula, in that the latter occurs in all the cleavages irrespective of whether the blastomeres are progenitors of the germ cells or not.
In the ovarian eggs of many animals an elimination of nuclear constituents has been observed (for a list of these cases see Meves, 94, p. 149); all these cases deal with elimination during the resting period of the nucleus. On the other hand my observations mentioned above, as well as those of Wilson (96, p. 141) on Werezs, Mathews (95) on Asterzas, Gardiner (98) on Polycherus, Griffin (99) on Thalas- sema, and many others, show that there is an escape of chromatic substance from the nucleus into the cytoplasm during the period of mitosis. In the cases just mentioned this elimination occurs during the first maturation division, and Griffin at least, affirms that it does not oceur in the cleavage mitosis. In Crepzdula, on the other hand, it occurs in every mitosis (except that of the second maturation), though it is, of course, most evident where the nucleus is large and the amount of chromatin great.
In this connection the theoretical conclusions of De Vries, Weismann and Roux, concerning the nuclear control of the cell should be recalled. De Vries holds that there is an actual migration of pangenes from the nucleus into the. cell body, these pangenes giving character and direction to all cytoplasmic processes, in fact, both De Vries and Weismann assume that the entire cytoplasm is the product of the pangenes. Roux holds that the nuclei become progressively specialized during development, and that these specialized nuclei determine the character of the cyto- plasm, but he does not suggest how this determination occurs. Weismann accepts and unites both the views of De Vries and those of Roux.
Judging these theories by the facts of chromatic elimination in Crepzdula and other gasteropods, [ am compelled to conclude that in all nuclei the chromatin appears the same in character, differing only in quantity; in all nuclei the chro- matin is differentiated into oxychromatin and basichromatin, the latter alone form- ing the chromosomes, while the former is eliminated; there is no evidence of progressive differentiation of the nuclei. That these facts, however, are not con- clusive against the theory of Roux is shown by the fact that in Ascarzs there is a specialization of the somatic cells as distinguished from the germ cells; if such a specialization occurs in Crepzdu/a it must begin at a much later period than in Ascaris. On the other hand, the fact that the eliminated chromatin is differentially distributed to the cleavage cells (see Part II, Sec. II) may be held to afford evidence of the fact that it plays some part in the differentiation of blastomeres.
KARYOKINESIS. 53
But whatever the theoretical bearings of this elimination may be, there can be no doubt of the fact that in Crefzdu/a, and perhaps in a large number of animals, there is a very extensive exchange of material between the nucleus and cytoplasm, and, further, that a large part of that most characteristic nuclear substance, the chromatin, passes into the cytoplasm in the form of oxychromatin during every cell cycle, while a relatively small portion is preserved for the purpose of reproducing the daughter nuclei.
There is thus in karyokinesis a rhythmical growth and collapse of the nucleus as a whole, the new nuclei arising endogenously, z. e., from chromosomes, within the old. In fact, one might speak of these changes in the nucleus as a systole and diastole (Ryder, *94), by means of which an exchange of nuclear and cytoplasmic material is brought about.
The nuclear membrane appears to permit the passage of materials inward but not outward during the resting period, whereas the escape of nuclear material into the cell is brought about by the disappearance of the nuclear membrane during karyokinesis. Such a process is not universal, for in cells where karyokinesis occurs very rarely, or not at all, the interchange between cytoplasm and nucleus has been observed to take place, but the phenomena described are characteristic of mitosis in general. 2
2. CENTROSOMES AND CENTRAL SPINDLES.—a. Structure and Metamorphoses.— It is evident from the history of the centrosomes of Crepzdu/a that throughout the maturation and cleavages up to at least the 60-cell stage, the centrosomes are abso- lutely continuous from one cell generation to the next, with the possible exception of the fertilization stages. Of this fact there can be no particle of doubt. With the exception of the earliest origin of the centrosomes of the first maturation, and with some reserve as to the origin of the centrosomes of the first cleavage, I be- lieve that I have seen every step in the origin and metamorphoses of the centro- somes up to the 12-cell stage; while in all the cleavages up to the 60-cell stage, or even later, I have observed and drawn the centrosomes at almost every stage in the cell cycle. Fortunately, this is not a very difficult thing to do, since the centro- somes are so large and distinct that even during the height of the resting period they can be seen in entire eggs, and their elongation to form the central spindles plainly observed (see Plates V, VI).
The principal features of the entire centrosomal cycle from one cell division to the next may be summarized in a single sentence: Zhe minute granules at the poles of the central spindle enlarge until they become hollow spheres within whitch new centrosomes and central spindles appear. The individual variations character- istic of the maturation and the different cleavage stages have been mentioned in detail, and need not be reviewed here; it is sufficient to say that the history of every centrosome conforms to the above statement.
From this it is evident that the centrosomes and central spindles form a unit structure, as Heidenhain (94), MacFarland (’96) and Boveri (01) maintain. Only in the fertilization is this not the case, and there are few, if any, well authenticated
5A KARYOKINESIS.
cases on record in which the centrosomes and central spindle of the first cleavage form a unit structure. Even in many of those cases in which there is a division of the sperm centrosome and a well-marked central spindle between the halves (e.2., Physa, Pleurophylidia, Unio, Cerebratulus, Thalassema, Arentcola), this central spindle completely disappears and the definitive spindle is formed de novo between independent centrosomes. In view of the unit structure of centrosomes and central spindles in other divisions, this is certainly a striking phenomenon, and indicates that the centrosomes of the first cleavage are in their first appearance more independent of each other than in any subsequent cleavage. It also suggests a possible way of unifying the conflicting accounts as to the origin of the cleavage centrosomes.
(6). Relation of Centrosome to Cell Body and Sphere.—In the mollusks which I have studied, the centrosome is at all stages in its cycle sharply delimited from the surrounding cell-body and sphere. The outer zone of the mother centrosome does not disintegrate and lose its outlines until after the daughter centrosomes and spindle have appeared within it, so that in all stages of the cell cycle there is a clearly marked centrosome. In the cleavage the outlines of the centrosome are most difficult to distinguish in the anaphase (figs. 59 and 67), but even at this stage there can be no doubt of its sharp separation from the surrounding sphere. Only in the egg and sperm asters during the approach of the germ nuclei is this separa- tion completely lost. No clearly marked sperm centrosome can be recognized at any stage, and the egg centrosome which is very evident during the anaphase of the second maturation, and which during this period undergoes a typical transfor- mation into a hollow sphere (figs. 52-56), loses its outlines and completely disappears in the surrounding sphere before the union of the germ nuclei. This again marks a peculiarity in the centrosomes during fertilization not found in any other cell cyele.
The centrosomes and spheres grow simultaneously reaching their greatest size in the telophase or resting period when the astral radiations are smallest; the astral radiations again become prominent when the new centrosomes have moved out of the old centrosomes and sphere and are growing rapidly in size. The growth of the centrosomes and spheres is not coincident with that of the nuclei; on the contrary they are smallest when the nuclei are largest, viz., in the early prophase, and they have nearly reached their largest dimensions when the nuclei are smallest, z. ¢., in the late anaphase before the formation of the chromosomal vesicles (cf figs. 3 and 16, 27 and 34, 53 and 59, 65 and 67). This, as well as other morphological phe- nomena involved in the escape of achromatin at the poles of the spindle and the coincident growth of the spheres and centrosomes, together with the changes in the staining reactions of the latter, indicates that the spheres and centrosomes grow in part at the expence of substance escaped from the nucleus. That this is not a complete statement of the facts, however, is shown by all cases of unequal cleavage, in which the centrosomes and spheres at the two poles of the spindle are always equal until the constriction of the cell body begins (figs. 15, 55, 72, ete.), but immediately after this they become unequal in size, and in the end are proportional
KARYOKINESIS. 55
in size to the quantity of cytoplasm in which they he (figs. 16, 34, 73, ete.). We have already seen that the nucleus is always proportional in size to the cytoplasm in which it les, and we are also compelled to conclude that the size of the centrosome and sphere depends ultimately upon the quantity of cytoplasm. This must be taken to indicate that both centrosome and sphere receive substance from the cytoplasm during their period of growth, and on the other hand it can plainly be seen that the remnants of the old centrosomes and spheres are slowly transformed into cyto- plasm or cell membrane after the new centers have moved out of them. There is, therefore, an interchange of substance between cytoplasm and centrosome, wholly similar to that between cytoplasm and nucleus (see p. 53),
(c). Relation of Centrosome to Nucleus.—In certain cleavages the centrosomes, especially during the resting period, are very large and conspicuous, ¢. g., in the pause preceeding the third and fourth cleavages (figs. 69, 74), they are fully six «w in diameter. They contain a reticulum of material which stains blue or black with hematoxylin, and on the whole they present an appearance remarkably like nuclei. In no case, however, have I seen any evidence that the centrosomes are directly derived from the nuclei, though this may possibly be the case in the origin of the centrosomes of the first maturation division; on the other hand they may, as indicated above, receive substance which escapes in a dissolved condition from the nucleus during every mitosis.
Whatever the ultimate origin or phylogenetic relationships of the centrosome may be, there is a remarkable parallelism between it and the nucleus, as the follow- ing statements will show :
1. Both begin their developmental cycle as small granules, the central cor- puscle in the case of the centrosome, the chromosomes in the case of the nucleus.
2. Both grow enormously by the absorption of surrounding substances and become vesicular; in the cleavage of the egg the vesicular condition is followed by a reticular condition in both.
3. Both undergo radical changes in their staining qualities during this enlargement, passing from a condition in which they are uniformly chromatic to one in which they are almost entirely plasmatic in reaction; finally, they again become largely chromatic, so that the centrosomes in the resting stages of certain cells look like small nuclei filled with a chromatic reticulum (cf figs. 69, 75, 74, 75).
4. When they have reached their greatest volume both are proportional in size to the size of the cell-body in which they are found ; this probably indicates that the substances absorbed by both in their growth are derived from the cytoplasm.
5. In both, the daughter structures (centrosomes or nuclei) are but a fraction of the mother organ, the remainder of the latter passing sooner or later into the cytoplasm. |
The centrosome thus repeats the history of the nucleus; at one period it takes up substances from the cytoplasm; when it has reached its greatest size the new centrosomes and central spindle arise within the mother centrosome from a part of
56 KARYOKINESIS.
its substance, and the remainder of the latter passes back into the sphere and ulti- mately into the cytoplasm. It is evident from this description that, as in the case of the nucleus, so also in the centrosome there is a sort of diastole and systole, the phases of the one alternating with those of the other.
(@.) Comparisons.—I have found centrosomes, similar in all respects to these just described, in three species of Crepzdula, and in Urosalpinx, lllyonassa, Fulgur, Sycotypus,and Aeolzs. In structure and history the centrosomes in all these gastero- pods are similar to those which have been observed in Dezau/ula (MacFarland, ’97), Unto (Millie, 98), Thysanozoon (Van der Stricht, °98), Rhynchelmzs (Vejdovsky, "88, and Vejdovsky and Mrazek, °98), Actznosphaertum (R. Hertwig, 99), and Flaminea (Smallwood, 01), while they bear many resemblances to those which have been observed in Echznus (Boveri, 01), Ascarzs (van Beneden, ’87, Boveri, ’88, 01, Brauer, 795, First, 98), Szda (Hacker, ’93), Salamandra (Rawitz, 96, Niessing, 99), Salmo (His.,’98), Faleur (MeMurrich, 96), Zzmax (Byrnes, 96 and 799, Lin- ville, 1900).
Van der Stricht’s interpretation of the relations of the centrosome to the attraction sphere in 7hysanozoon seems to me most satisfactory, not only because he has for it the approval of Van Beneden and Boveri, but also because by it the various forms of centrosomes present in the animals named above and particularly the remarkable centrosomes of mollusca can be satisfactorily related to one another and to other forms. We need not here concern ourselves with the origin of the centrosomes of the first maturation; Van der Stricht believes that they arise from the nucleus, and this view is supported and extended by the recent observations of Schockaert (1900). Soon after its appearance the centrosome of Zhysanozoon is differentiated into a central corpuscle and a medullary zone (couche medullaire). These two together constitute the centrosome of Boveri, the central corpuscle being his centriole. The medullary zone is homogeneous in structure, and no astral fibres penetrate it except at the time of origin of the new spindle figure; it is usually bounded peripherally by a dark line (in reality a sphere).'| Around this is a clear area traversed at all stages by delicate astral fibres; this is the cortical zone (couche corticale), and it is possible that it is derived, in part, from the centrosome. Around this is the zone of astral rays, which sometimes may be subdivided into an inner dark anda peripheral clear zone. The centrosome and cortical zone continually enlarge, as the division advances, until they reach a great size. In the division of this centre, the central corpuscle first divides, usually into two, and the central spindle appears between the halves; the medullary zone then becomes bounded by granules, and this boundary gradually fades from view, though Van der Stricht believes that the entire attraction sphere persists and divides, thus giving rise to
1'The clear zone which is so generally found around the central corpuscle is believed by many observers to be the result of destaining. Such a zone is produced by destaining yolk-spheres; nucleoli, etc. Fischer (99) calls this “Spiegelfirbung.” It will be observed that the manner in which these centrosomes stain, which have a dense periphery and clear central area, is the exact reverse of ‘“Spie- gelfarbung.”
KARYOKINESIS. 57
the daughter attraction spheres. The relations of the parts of the centrosome, the attraction sphere and the aster of Zysanozoon may be indicated as follows :— Central Corpuscle Medullary Zone Cortical Zone Inner Zone t oa oes Peripheral Zone J
The most important points in which my observations differ from those of Van der Stricht are the following :—
1. The peripheral boundary of the centrosome (medullary zone) is much denser and more deeply staining than in 7hysanozoon.
2. During the rest stages in the cleavage the central corpuscle is represented by an enormous number of granules, only two of which form the new centrosomes.
3. Neither the medullary nor the cortical zones of the attraction sphere ever divide as a whole, but after the origin of the new amphiaster they are slowly dis- solved in the cytoplasm.
4. The central corpuscle of one generation gives rise to the entire centrosome, z. e., central corpuscle and medullary zone, of the next.’ This is most plainly seen in the anaphase of the first maturation.
5. At no time do the astral rays traverse the medullary zone,’ though new
Centrosome
lof Attraction Sphere.
rays which are not part of the old system may arise within that zone around the new centers.
In Deaulula, according to MacFarland, the centrosome increases greatly in size from the prophase to the anaphase and a single granule appears within it. This granule soon divides into two which move apart and become the new centrosomes. The whole of the old centrosome is transformed into the new centrosomes and cen- tral spindle. The rays are inserted on the centrosome, not on the central granule. Even after the new spindle figure has reached a considerable size the rays continue to be centered on the figure as a whole.
My observations differ from MacFarland’s only in one respect: The whole of the old centrosome is not transformed into the new spindle figure, but the latter arises zwzthzu the old centrosome. This is plainly true of the maturation stages, corresponding to those which MacFarland has studied, and in the first three cleav- ages, but the case is not so clear in the later cleavages, as a glance at my figures 74-76 will show.
According to Lillie (98) each centrosome in the prophase of the second matur-
‘Tam not quite certain whether this may not be involved in Van der Stricht’s statement that the centrosome becomes differentiated into a central corpuscle and medullary zone. The fact, however, that he maintains a persistence of the attraction sphere leads me to suppose that he regards each me dullary zone as derived from a preexisting one.
2 In this important respect my observations agree entirely with those of Boveri and his pupils, MacFarland and Fiirst, and differ from those of Van Beneden. Indeed, it may be doubted whether the term “medullary zone” should be applied to a structure which shows no radiations. Other con- siderations, however, render it extremely probable that the peripheral layer of the centrosome in gastero- pods and the “medullary zone” of Thysanozoon are homologous.
8 JOURN. A. N.S. PHILA., VOL. XII.
58 KARYOKINESIS.
ation division of (/zzo is composed of several large granules into which rays are inserted. In the metaphase these granules subdivide, and some of the fragments are distributed in the form of a sphere, the ‘‘inner sphere;” ‘‘one of the granules remains behind as the centrosome of the new inner sphere,” but “a large part of the centrosome granules is changed into the red-staining substance of the sphere.” In the anaphase the granules of the inner sphere, together with the peripheral accumulation of its ground substance, fuse into a continuous membrane. ‘The centrosomes are united to the membrane of the inner sphere by a few irregular threads which are not part of the system of radiations.’ Within the sphere the daughter centrosomes and central spindle arise.
Lilhe emphasizes the fact that the inner sphere is not the centrosome, and he says that Boveri's “ centrosome” is really the inner sphere, while his “centriole” is the real centrosome. He also holds that MacFarland’s “centrosome” is really the inner sphere.
Lillie’s conclusions seem at first sight to be very different from any of those mentioned above, and yet on consideration it will be seen to be rather a difference of terms than of facts. His ‘‘inner sphere” is undoubtedly homologous with Boveri's centrosome, his “centrosome” with the central corpuscle or “centriole” of Boveri. As to the genesis of these parts, I have never been able to observe the formation of the “inner sphere membrane” from granules derived from the central corpuscle as Lillie has done, nor have I observed the fragmentation of the central corpuscle and the transformation of these granules into the ground substance of the centrosome (substance of medullary zone), though in early phases the centrosome of Crepzdula is irregular in outline, as if composed of closely connected granules. In all cases which I have observed the central corpuscle enlarges but does not fragment; its substance accumulates peripherally and forms a continuous membrane which szdbse- quently is transformed into a layer of granules. If the central corpuscle of Uxzo were to remain a single structure, and were to continually expand, the result would not be unlike my observations. The one critical point in the comparison of Lillie’s observations with those of other investigators, is to determine whether the whole of his “inner sphere” is derived from the central corpuscle ; if it is, the differences are only matters of detail. It should be remembered that, according to Lillie, the inner sphere is itself a structure which sooner or later disintegrates and passes into the outer sphere or cytoplasm, and that it should disintegrate at different stages in different eggs is quite possible.!
In a later paper on the subject, Lillie (99) says that the inner sphere enlarges very rapidly after the formation of the second polar body, and its substance gradually merges with the general cytoplasm. Its interior is occupied by the vesicular sphere substance at the nodes of which are deeply staining granules. In this respect there is considerable difference between Unzo and Creprdu/a, for in the latter the inner
‘T ought to add that I have had the pleasure of seeing Lillie’s beautiful preparations, and they leave no ground for doubting the accuracy of his observations. Professor Lillie also personally assures
me that he is quite convinced that the whole of the “inner sphere’’ is derived from the central cor- puscle. (See also his recent work, 1900, p. 242.)
KARYOKINESIS. 59
sphere remains much smaller and persists for a considerable period, while the outer sphere undergoes a metamorphosis similar to that which Lillie describes. The inner sphere is very faint and difficult to detect in Crepzduda, and it may be that Lillie has overlooked it, or it may disintegrate sooner in U/nzo than in Crepzdula (cf. my fig. 36 and Lillie’s (99) fig. 14).
Vejdovsky ('88) found in the egg of Rkychelmzs immediately preceding the first cleavage, a hyaline sphere, the “ Periplast.” within which in the course of cell-division a new element, the *Tochterperiplast” appears; the latter divides into two spheres which represent the poles of the new spindle. During the first and second cleavages a new element, the “ Enkelperiplast” arises within the ‘* Tochterperiplast,’ while the ‘*‘ Mutterperiplast’ degenerates or fuses with the cytoplasm. This was the first obser- vation tending to show that the new centers arise endogenously within the old.
More recently Vejdovsky and Mrazek ('98) have confirmed and extended this account. They find at the poles of the first cleavage spindle, a large sphere at the center of which is a deeply staining granule, the ‘centrosome, (central corpuscle) ; this is surrounded by a hyaline sphere the ‘“ Tochterperiplast’’ (medullary zone), at the periphery of which the rays are attached; surrounding this is the *‘ Mutterperi- plast” (cortical zone). For the sake of uniformity we shall use the terms in paren- thesis in place of Vejdovsky’s terminology. As mitosis advances the medullary zone (Tochterperiplast) grows rapidly, becomes reticular or alveolar in structure and is bounded by a dense peripheral zone ; the central corpuscle decreases in size and stains less densely, while radiations appear around it within the medullary zone. Around the central corpuscle a new medullary zone appears within the old one. The central spindle is formed, after which the new medullary zone divides. Vejdovsky and Mrazek consider that centrosome and periplast are persistent organs of the egg and that they represent a single whole. The entire periplast, however, does not persist, but the inner zones give rise to the outer ones which gradually disintegrate into the cytoplasm.
The resemblances between these observations and those which I have described in the preceding pages are very striking. The only important difference between my own observations and those of Vejdovsky and Mrazek is the following: A new medullary zone does not form around the central corpuscle before the latter divides, but only afterwards, z. e., in the mollusks which I have studied the two daughter centrosomes are present before a new medullary zone is formed.
R. Hertwig has observed in Actznospherium a large * spongy centrosome ” within which a “‘ reduced centrosome” (central corpuscle) appears and divides into two; the latter then enlarge to form new ‘‘ spongy centrosomes,” while the former “spongy centrosome’ does not divide, but disappears in the cytoplasm. The resem- blances in this case to those discussed above are too obvious to need comment.
In Lzmax, Byrnes ('99) has found centrosomes which in many respects resemble those of Unio, Creprdula, Aeolis and other mollusks. In the metaphase of the first maturation there is at each pole a group of central granules, within a clear area, which is surrounded by a broad, densely staining zone. In the anaphase the
60 KARYOKINESIS.
central granules divide into two groups and a spindle appears between them and within the ‘‘centrosphere.” From my own observations I am convinced that both the central clear area, with its contained granules and the denser zone which sur- rounds it, belong to the centrosome, which, when fully formed in the anaphase, consists of central corpuscles and medullary zone, the latter bounded by a narrow line or layer of granules. If this be correct the spindle in Zzmavx arises wzthin the medullary zone, as in many other cases.
Linville (1900) has observed a similar centrosome in Zzyzax and other Pulmo- nates, around which is a cortical zone of radiating structure. He has not followed the metamorphosis of the centrosomes in detail, but his figures give evidence that the history of the centrosome in these Pulmonates is not different from what is known in other mollusks.
Finally, it seems quite possible to interpret most of the multitudinous forms of centers which have been described in the eggs of various animals in accordance with the Van Beneden-Boveri idea, as extended and defined by Van der Stricht. and particularly by Boveri (1901), provided that the remarkable changes in the structure of the centrosome from prophase to telophase be kept in mind. In mollusks the centrosomes are characterized (1) by the great breadth and density of the peripheral portion of the centrosome which, about the middle stage of mitosis, forms a dense ring or sphere surrounding a clear area, and which in all stages sharply separates the centrosome from the surrounding sphere, (2) by the fact that the centrosomes grow to an unusual size during mitosis, and (3) by the origin of the entire amphiaster of one generation within the centrosome of a pre- ceding one. Possibly the second and third of these characteristics are the results of the first, since the sharp boundary of the centrosome at all stages make it unusually easy to recognize the great growth of the centrosome and also the place of origin of the new centrosomes and central spindles.
Boveri's (01) masterful contribution on the nature of the centrosome reached me some time after my paper had been completed, and I have therefore been unable to make the extended use of it which I could have desired. In broad outlines my conclusions as to the centrosome are fundamentally like those of Boveri. The one most important point of difference between us is that Boveri considers the centriole as a differentiation of the centrosome, perhaps a continuous and persistent strue- ture, around which a portion of the centroplasm always remains to form the new centrosome. On the other hand, I hold with R. Hertwig (99) that the centriole gives rise by growth to a centrosome, within which a daughter centriole differenti- ates, z.¢., the centriole undergoes in its cycle of development a metamorphosis into centrosome and daughter centriole. In each generation the outer zone of the cen- trosome is thrown off, while the new centrosomes and central spindle come from the center of the old. There is thus a kind of endogenous formation of centro- somes, as Vejdovsky and Mrazek maintain.
Since receiving Boveri's paper I have carefully re-examined the critical stages in my preparations to see whether I could have overlooked an outer zone of centre-
KARYOKINESIS. GL
plasm around the deeply staining body at the center of the aster. There is a faintly staining zone surrounding the central body in the prophases both of the maturation and cleavage divisions (see figs. 4-7, 26-50, 52-56, 70-72, 76), but this zone accord- ing to Boveri's definition, does not belong to the centrosome, since even at its first appearance (cf. fig. 70) it is traversed by radiations; furthermore, a study of con- secutive stages shows that it develops step by step into the inner portion of the sphere. On the other hand, I believe that I have followed the central, deeply staining body through every stage of its growth and metamorphosis, having seen it not merely in the stages represented in the plates, but in thousands of others, many of which were carefully drawn. The result of this study convinces me that the small, deeply staining granule of the early prophase becomes the dense, spherical body of the metaphase and the large, hollow sphere of the anaphase, and that this body is the centrosome. The fact that in the mollusks generally the peripheral layer of the centrosome stains more densely than the central portion, makes it unusually easy in these animals to distinguish between the centrosome and the sur- rounding sphere. The result therefore of the re-examination of my preparations in the light of Boveri's work does not in any respect lessen my confidence in the accu- racy of my observations and interpretations, at least as far as Cyepzdu/a is con- cerned.
The type of centrosome represented by Crepzdula, Unto, Haminea and Aeolzs, viz., one within which the new centers and central spindles arise from the centriole while a considerable part of the mother centrosome fades away into the sphere, agrees much more closely with the types of centrosomes found in Ascarzs, Thalas- sema and Echznus, than does that of Dzau/ula. Boveri represents these four types in text figures (pp. 102-1035), and it can be seen at a glance that in the first three types the daughter centers and spindles occupy but a small part of the old centro- some, whereas in the fourth type (Dzau/ula) they occupy the entire centrosome. 1 have found that the relative size of the central spindle and daughter centrosomes (Netrum of Boveri), as compared with the inclosing centrosome, differs considerably in different cleavages of the egg. Thus in the first, second and third cleavages the netrum is much smaller than the mother centrosome, whereas in the fourth, fifth and later cleavages the netrum almost entirely fills the mother centrosome. In view of these facts I venture to suggest that a re-examination of Dzaulu/a with regard to this point might show that the outlines of the netrum are not coincident with those of the mother centrosome, but that the former hes wz/hzz the latter as is the case in the other mollusks named above, as well as in other types of centro- somes described by Boveri.
(e) The Centrosome as a Persistent Cell Organ.—There is no more perplexing problem in connection with the cell than that of the significance of the centrosome. On the one hand there are the well established facts as to (1) its persistence from cell cycle to cell cycle (my own observations showing that in the cleavage of Crep- zdula it persists without interruption to a stage with more than sixty cells and probably throughout the entire development); (2) its independent growth and
62 KARYOKINESIS.
division (shown not only by observation in many animals, but particularly by Boveri's (97) experiments on echinoderm eges); (3) its characteristic structure and metamorphoses, which in a large number of animals (perhaps in all) can be reduced to a common type.
These features are of such character and importance that they justly entitle the centrosome to the rank and title of a permanent cell organ (Van Beneden, Boveri). One who has followed the history of the cleavage centrosomes through several cell cycles, who has observed their unfailing persistence, the regular cycle of changes in form and staining reactions which they undergo, their complex struc- ture, their form of division, their parallelism in these and in other respects to the nuclei (see p. 55), can no more doubt that these centrosomes are persistent cell- organs than that nuclei or plastids are.
On the other hand there are the well known facts (1) that, according to the best testimony, there are no centrosomes whatever in the higher plants (Strasburger, Osterhout, Mottier, e¢ a/.); (2) that the persistence of centrosomes has been denied in the tissue cells of some animals, and even in certain stages of the ege, particu- larly during fertilization (Foot, ’97, Lillie, 98, Child, 99); (3) that various stages intermediate between centrosomes and microsomes or other cytoplasmic constituents have been described (Burger, Reinke, Watase, Mead, Eismond, Erlanger) which indicate that the centrosome is only a temporary differentiation of the cytoplasm ; (4) that artificial asters and centrosomes may be formed in egg cells by the action of various solutions, and that these may function as normal asters and centrosomes (R. Hertwig, Morgan, Loeb, Wilson).
The contradiction between these two classes of evidence is so complete, and the phenomena in both classes are apparently so well attested, that one would be inclined to seek refuge in the conclusion that in some cases the centrosomes are persistent cell organs and in others temporary structures, were it not for the fact that this contradicton may concern one and the same object (e. g., the eggs of Echi- noderms and of © hetopterus).
There is certainly no ground to doubt that in the cleavage of the eggs of many animals the centrosomes are, under normal conditions, absolutely continuous from cell generation to cell generation. Nor is there any possibility of doubting that in certain animals the centrosomes show independent growth and division, and that they pass through certain characteristic metamorphoses in this cycle. The only possible interpretation of these undoubted facts is that, in some cases at least, the centrosome is a cell organ of morphological as well as of physiological significance.
Is the contrary evidence irreconcilable with these well established facts, and must we, therefore, conclude that the persistence of centrosomes, their growth, meta- morphoses and division have no general morphological significance? I think not.
(1) If it be granted that the centrosomes are not present at any stage in the cell cycle in the higher plants, this does not necessarily contradict the centrosome theory of Van Beneden and Boyeri, since the fact that they are present in the lower plants indicates that their absence in the higher plants must be the result of degene-
KARYOKINESIS. 63
rative changes. If, however, centrosomes may degenerate in whole classes of the plant kingdom, the centrosome is surely neither so ubiquitous nor so necessary a cell organ as the nucleus.
(2) So far as animals are concerned the centrosome has been found in almost all kinds of metazoan cells, and at nearly every stages of the cell cycle. The history of biology shows that the failure to find structures, even by many observers, is no proof that they do not exist, and this is particularly the case with structures so difficult to observe and undergoing so great metamorphoses as the centrosomes. As to the alleged disappearance of the centrosome in the fertilization of the ege (Foot, Lillie, Child,), it must be said that this like negative evidence in general is not wholly conclusive. Certainly, so far as my own work goes, I cannot. affirm that both egg and sperm centrosomes entirely degenerate, although they do dis- appear, nor can I affirm that the cleavage centrosomes are new formations, although I am unable to trace any connection between them and the centrosomes of the ege and sperm. Even if these centrosomes disintegrate, it may be that the new centro- somes arise from some of their fragments; in fact, such would seem to be the case in Crepidula (see p. 27). The history of the centrosomes in the fertilization is at best a complicated one, and is by no means as clear as in the cleavage of the ege or in the division of tissue cells, and until we have more exact knowledge of the origin of the centrosomes in the fertilization, this doubtful evidence against the continuity of the centrosomes should not be permitted to outweigh the’ positive evidence in favor of their continuity afforded by ordinary mitoses.
(5) The view that the centrosome is only the meeting point of astral rays or that it represents merely a condensation of the cytoplasm, or that it is an enlarged microsome, entirely neglects to take account of the complex structure and metamor- phoses of the centrosome, as well as of its division and persistence. These are by all odds the most characteristic features of a centrosome, and until it has been shown that the cytoplasmic structures mentioned above are capable of reproducing these characteristic features, it may well be doubted whether they are really centrosomes. The mere formation of cytoplasmic radiations is in itself no positive indication of the presence of a centrosome, since such radiations are found in the higher plants where centrosomes are wholly lacking (Osterhout, Mottier), in non-living substances such as carbolic acid and chloroform, gelatin and albumen (Roux, Biitschli, Fischer), where there is certainly no centrosome with the characteristics described above ; around mid-bodies (see figs. 60, 61), and in many of the multiple and accessory asters found in cells under normal and artificial conditions, which show no body at the center of the rays (Mead, Lille, Morgan, e¢ a@/.).
(4) The fourth class of facts which speak against the theory of the persistency and morphological importance of the centrosome forms by all odds the most serious objection to that theory which has yet been raised ; I refer to the experimental pro- duction of centrosomes both in fertilized and in unfertilized egg cells by the action of various solutions (R. Hertwig, Mead, Morgan, Loeb, Wilson). It may well be doubted whether all of these structures are centrosomes, but that some of them
64 KARYOKINESIS.
are such is beyond dispute. Certainly, structures which function as centrosomes through a long period leading up to the production of larva (Loeb, Wilson), are enough like centrosomes to pass under that name. And even in cases where larvee are not produced (experiments of Hertwig, Mead and Morgan), there can be no reasonable doubt that centrosomes are found in the larger asters, even if the smaller ones do not contain them.
In the case of Morgan’s experiments on fertilized eges it might be maintained that the numerous asters and centrosomes observed are derived by division or frag- mentation from those already present in the cell-body, where it not for the fact that similar asters and centrosomes have been observed in the case of unfertilized eggs (Hertwig, Morgan, Wilson) where no centrosomes are present in the cytoplasm. The phenomena in these two cases are so similar that one cannot believe that they are due to wholly different causes; we may, therefore, safely class them together.
Hertwig maintains that in his experiments the centrosomes were formed from the achromatic constituents of the nucleus. He says: ‘‘Ich deute somit die Cen- trosomen as selbstandige gewordene geformte achromatische Kernsubstanz, eine Deutung fiir die ich wiederholt eingetreten bin.’ In part Morgan agrees with this position, though he also holds that centrosomes may arise at a distance from the nucleus and therefore from the cytoplasm. ‘I agree,’ he says, “ with Hertwig that the centrosomes may develop out of the achromatic substance of the nucleus, but I see no ground to extend this statement to include all centrosomes. . . . There is good evidence to show, I think, that similar bodies may arise in the cytoplasm also, as shown by Reinke, Mead, Watase and myself.’ It is a notable fact, how- ever, that in all these cases cited by Morgan the nuclear membrane has disappeared or is much shrunken and collapsed, showing that nuclear substance has escaped from it. This is true at least of the figures of Reinke, Mead and Morgan; Watase gives no figures of the egg of MZacroédella which he describes as containing “a series of thirteen asters, ranging from the miniature aster, with the microsome in its center, to the normal aster with a veritable centrosome.’ In the figures of Reinke, Mead and Morgan one is much struck by the fact that at the time when the asters appear in the cell the nuclear membrane is either entirely lacking or is much shrunken, showing that achromatic material has escaped into the cell. Of his own experiments Morgan says (p. 464): “ The first effect (of the salt solution on the egg) is to cause a shrinkage of the nucleus; then after the return of the eggs to sea water the division of the nucleus and subsequently that of the egg takes place: . . . central bodies are present in the artificial astrosphzres in almost all the stages.” Again (p. 517) he says: “ At the time when the nuclear wall dis- appears the astrospheres throughout the egg, whether in contact with the chromo- somes or not, become conspicuous and then fade away again as the chromosomes pass into the resting nuclei. There is some connection between the setting free of the chromosomes and the development of the astrosphwres ;” or rather, as it seems to me, between the escape of some nuclear constituent and the development of the astrosphaeres. The fact that achromatic nuclear substance may be distributed widely through the cell in normal mitoses was pointed out in the section on aster
KARYOKINKESIS. 65
formation, and I see no evidence in the cases brought forward by Morgan to indi- cate that the centrosomes or asters in all these cases may not be derived from escaped nuclear material."
There is certainly a close relationship between the nuclei and the centrosomes. The achromatic substance of the nucleus contributes to the growth of the cen- trosome in every normal cell cycle (see p. 54), and it is probable that the daughter nuclei in their growth resorb from the spheres a portion of this same achromatic substance. The peripheral spindle fibres are formed out of this substance (vzz., linin and oxychromatin), and bear a striking resemblance to the central spindle fibres at an early stage (cf. figs. 55 and 75). In the first maturation of the egg the centrosomes or asters do not appear until substances have escaped from the nucleus, as is shown by the breaking or indentation of the nuclear membrane (cf Coe 99, Carnoy and Lebrun ’99, Gardiner 98, Griffin 1900, Mead °98, Van der Stricht “98, Schockaert 1900), and, finally, the granular or reticular centrosome undergoes the same changes in reaction to stains as does the oxychromatin and linin, being at one time uniformly chromatic, and later uniformly plasmatic in reaction. In all these respects the centrosome behaves like an isolated portion of the oxychromatin and linin.
A large number of investigators have observed the formation of centrosomes and spheres from some of the nuclear constituents, particularly among the Protozoa, (Brauer 793 in Ascarzs, Riickert 94 in Cyclops, Ishikawa °94 and Calkins ’98 in Noctiluca, Balbiani 95 in Spzrochona, Schaudinn 96 in Acanthocystes, Hertwig (99 in Actinospherium, et al.).
' Wilson’s (01) recent work on Toxopneustes shows that asters and centrosomes may arise in eggs treated with Mg Cl, not only far from the nucleus, but even in enucleated fragments. Wilson says (p. 542) ‘ There is absolutely no evidence for, and the clearest evidence against, the view that the original cytasters form at or near the nucleus, to migrate thence toward the periphery, or that they arise by the multiplication of a single primary center.” He holds, therefore, that centrosomes and asters may arise de novo in the cytoplasm. Such a view, if generally true, would be fatal to the one which is set forth in this paper, and it deserves more extended treatment than can be accorded to it in a foot-note. In brief, the critical questions as to Wilson’s experiments are these: (1) Are the bodies in question real centrosomes; (2) do they arise de novo in the cytoplasm? Iam not disposed to question the fact that these bodies are really centrosomes, but I am inclined to doubt the statement that they arise de novo, if by that it is meant that they arise without genetic relation to other centrosomes or to the nuclei. The fact that these “artificial”? centrosomes may appear far from a nucleus, or even in enucleated frag- ments, does not necessarily imply that they are wholly independent of them. The achromatic sub- stance of the nucleus may be widely distributed throughout the cell during mitosis, and I have observed in the eggs of Crepidula, which have been placed in 2%-3% Na Cl for several hours, that the achro- matic portion of the nucleus may exist as one or more vesicles, with definite walls, quite distinct from the chromatic portion. In some cases these achromatic vesicles are in contact with the chromatic one ; in others they are widely scattered throughout the entire cell. Furthermore, many of these vesicles apparently give rise to centrosomes. If the achromatin of the nucleus is genetically related to the cen- trosome as I have maintained in this paper, and if achromatin, diffused throughout the cell, may, under certain stimuli, be aggregated into vesicles, which then give rise to centrosomes, Wilson’s observations need not necessarily mean that centrosomes arise de novo.
In all cases in which “artificial” asters and centrosomes have been produced, a large amount of nuclear substance has been present in the cytoplasm. No one, so far as I can recall, has observed asters in egg cells while the germinal vesicle is still intact; with the escape of achromatin from the germinal vesicle, however, numerous asters and possibly centrosomes may appear in the egg. I have tried hy various means to produce asters in egg cells before maturation, but always without success as long as the germinal vesicle remains intact. I believe that it may be laid down as a general principle that escaped nuclear material is essential to the formation of an aster, and that an aggregation of such material is necessary to the formation of « centrosome.
9 JOURN. A. N. S. PHILA., VOL. XII.
@P) (or)
KARYOKINESIS.
Others have observed that centrosomes disappear within the growing daughter nuclei in certain cases. For example, Mead says that the centrosome left in the egg at the close of the second maturation of Chesopterus is last seen “in the midst of the fusing (chromosomal) vesicles, its position being indicated by the point of con- vergence of the rays of its waning aster.’ Exactly the same thing is true of Cere- bratulus (Coe), Thalassema (Griffin), Asterzas (Wilson and Mathews) and probably of other animals. In all these cases the centrosome is probably taken into the egg nucleus.
All of these facts seem to me to indicate that the centrosome is intimately related to the “formed achromatic” substance of the nucleus and that, in some manner, artificially produced centrosomes are formed out of this material as R. Hertwig maintains.
Loeb has found, by a series of remarkable experiments, that artificial partheno- genesis may be caused in the eggs of echinoderms and of Chetopterus by the action of a variety of substances upon the eggs, and he concludes that in general this par- thenogenesis is the result of diosmotic action of these substances and the with- drawal of water from the egg, though other factors also enter into the problem in the case of Chetopterus. In the light of the many observations and experiments which go to show that asters and centrosomes are produced from escaped nuclear material, the thought suggests itself that artificial parthenogenesis may be caused by any method which will bring certain nuclear constituents into the cell body and yet not seriously injure either nucleus or cytoplasm.
One of the most interesting chapters of Boveri's recent work on the centrosome is that in which he undertakes to account for the production of centrosomes by arti- ficial means. Boveri recognizes, as have many others, the intimate relation between the achromatic material of the nucleus and the centrosome. In cases where real centrosomes can be produced from unfertilized eggs he holds that they are formed by a kind of regeneration (veparatron, Driesch ‘97) from the achromatic substance of the nucleus. Not all nuclei, however, have this power, and, accordingly, Boveri distinguishes between (1) szzclez which are purely nuclear in character, and (2) centro-nuclet which contain a cytocenter. An example of the former is found in Ascaris, and of the latter in many Protozoa, and in some Metazoa, particularly in the echinoderms and in the ovocytes of many animals—perhaps of all.
The question at once arises: What reason is there for supposing that among Metazoa nuclei are divided into these two classes?’ Boveri himself has asked this question, and he concludes that in Ascarzs at least the nucleus cannot be a centro- nucleus, since in certain pathological eggs, in which the spermatozoon remained at the periphery or did not enter at all, the egg went through the maturation divisions and the egg nucleus came to the period of the solution of the nuclear membrane without a trace of fibre differentiation, of centrosomes, or of spindles or spheres. He therefore concludes that the nucleus of Ascarzs is a pure nucleus which has lost the capacity of forming centrosomes.
The evidence upon which such an important generalization is based seems to
KARYOKINESIS. 67
me to be insufficient. No small amount of evidence is required to prove that nuclei, which are so similar in all other respects, are so different in this one. I agree with Boveri that research only can determine where (and I should add whether) this is true. :
Returning now to Boveri's idea that in the artificial production of centrosomes they are formed by a kind of regeneration :—Morgan’s experiments make it ex- tremely probable that numerous centrosomes may be formed independently of each other in the cytoplasm. If these are formed of achromatic nuclear material it is easy to understand that they appear wherever a sufficient accumulation of this material is found in the cell body. But achromatic material separated from the nucleus is not necessarily a centrosome with all of the morphological and physi- ological features which that body exhibits, as is shown by the fact that such mate- rial is distributed through the cell at every mitosis. Either there must be an escape of some centrosomal substance or structure, or the condition of the cell must be such as to bring about centrosome formation from ordinary achromatic material ; the latter is | believe Hertwig’s view (see quotation on p. 64); the former is held by Boveri who considers that the centrosome may be regenerated (repaired) from the achromatic substance of centro-nuclei only.
This very suggestive hypothesis of Boveri's makes it possible to harmonize the well established fact of the persistence of the centrosome as a cell organ with that other apparently contradictory fact that centrosomes may be produced experimen- tally in the cell; and this it does by practically adding another phase to the series of changes through which the centrosome may pass, vzz.: