m -JT

\ ■.'. t •» » i f /• ■•• I •

mm,

■ii^^m^^^^'

H

8

H|

w^^mmmm

ny

\

T C

> ■c,v '

^

^

SK

^^:

^^t/k.

'♦^Hl'

■<•»-:•

■4r<^

'■•\

%

^ift

i

THE LIFE HISTORY OF GRIFFITHSIA BORIIETIAITA

Dissertation submitted to the Board of
University Studies of the Johns Hopkins Uni-
versity in conformity with the requirements
for the degree of Doctor of Philosophy,

by

I . F. Lewis

1903

•00--

■p o

THI'] LI?.'^ HISTORY OF GRIFFITHSIA BORTOTIAiTA.

The red alga, Griff 3 thsia Bornetiana , was first de-
scrilDOd by W. G. Farlov/ ('77). It has 'been reported as oc-
curring coinrnonly from northern l/Iassachiisetts (COLLIITS, '0 ,

p, 50) south to Long Ir;lr.nd Sound, rnd h^.r. Iieen i-or;r,,->ded

(1)
f ro:ii Nev7 Jersey (3RITT0H, '81). It forms rosy tufts, 2.5

to 15 centimeters high, on rocks, "v.li&rves, sponges, shells

and occasionally on Zostera" (PARLOV/, »79, p. 131), South

of Cape Cod it is found grov/ing from one to four feet "be-

lov; low water mark, in protected sitiiations such as the

Lit.:.lo Hr.;:"oor at Wood's Hole, Ifess,, and on rocks in more

exposed localities.

The present investigation /as "begun i i 1905 on ma-
terial collected at Cold Spring Ha^rhor, ITev; York, "by D. S.
Johnson in 1902, and has heen cor'tinued during 1906 lUid
1907 at .Vood's Hole and at the Jolins Hopkins University,

In all plants examined, v/ith two exceptions noted be-
lov/, the anther idia, cystocarps, and tetraspores are borne
on separate individuals, which ma:' ^e readily distinguish-
ed v;ith the aid of a hand lens. The male plant is smaller
and more compact than either the feiiiale or tetraaporic

(1) The form reported from as fcLr south as the Barbadoes
by laie. VICKP^S ('05) is "b^lir^ved by Dr. FARLOW not to
be identical with G. Bornetiana-

plant, and may "be identified b;- the a"brupt terminations of
the filaiaents (fig. 1). It rarely becomes more than 4
centimeters high. The female plant is more loosely tufted
than the male, and reaches a much larger size, becoming 12
to 15 centimeters high. The cystocarps form deep red dots
at the sides of the nodes (fig). 2), The filaments of the
female plant do not end abruptly, but become gradua.lly
smaller tov/ard their tips (fig. 3). The tetrasporic plant
is more slender the^n the female and to the eye more nearly
like it than the itiale. It may be distinguished by the
v/horls of tetraspores, which form complete rings at the
nodes (fig. 4). The tetrasporic plant, lilce the sexual
individuals, sometimes produces reproductive organs v/hen
consisting of but 10-20 cells and ha,ving a height of less
than half a centimeter.

One plant was seen m which a few antheridial branches
occurred, while the majority of the filaments bore numerous
procarps and cystocarps. In another case, most of the
branches produced antheridia, but a considerable number
bore at the nodes rings of cells resembling in all partic-
xxlars tetraspore mother cells, with the involucral rays
characteristic of the tetrasporic sorus. These tetraspore-
lilce structures are described in detail on pages 77- T'^.

Grif fithsia Borne tl ana becomes conspicuous at V^ood's
Hole in the fir'st half of July, and grows rapidly until it
reaches its maximiim development ahout the first week in Au-
gust, Tov/ards the middle of August the plants of all ages
cease to produce new branches and slov/ly become disorgan-
ized, losing their rich pink color and becoming easily de-
tached fro.'.i the substratum. At this season great quantitioc
are washed up on exposed points like iTobska Point at Wood's
Hole, After being torn loose from their fastenings, the
plants float about in the water for days or even weeks, con-
tinuing to produce spores v/hich were shown by experiment to
"be capable of geriaination. At this season, the tetrasporic
plants frequently show a very robust feabit, forming spher-
ical masses upwards of 15 centimeters in diameter.

The spores develop quite rapidly in the open. Bits
of cotton cloth, tied to piles near mature plants showed
in two weeks' time sexual plants v/ith ripe antheridia and
carpospores, and tetrasporic plants with mature spores.
The largest of these plants shov/ed 3 orders of branching,
and consisted of as many as 500 cells,

A notevforthy fact about the occxurrence of the vjirious
forms is that tetrasporic plants are always more abundant,
as well as on a.n average larger tlian sexual plants. Dviring
the first two weeks of August, 1907, mcire than 500 plants

were collected at random, care being taken to collect every
plant seen, and not to select the larger specimens. On one
occasion 352 individuals v;ere "brought into the la^boratory
and sorted carefully. Of these 321 were found to be tet-
rasporic, 15 cystocarpic, and 16 antheridial. At another
time more than 200 plants shov/ed ahout the same relative
proportions. In other \7ords, there is on an average an
equal number of antheridial and cystocarpic plants, and
for each sexual plant about ten tetrasporic ones. An exact
count v/as not kept of plants collected earlier in the sea-
son, but there seems to be no doubt tha,t tetrasporic plants
greatly predominate in nvunber at all seasons.

The sajne relations are sliown quite strikingly by
Champ i a yarvula and Chondria 'tonuissima at Wood's Hole,
and will probably be found to obtain in many other red al-
gae. Similar numerical prepondera-nce of tetrasporic plants
has been noted in Laurencia by PHILLIPS C96), in Polysi-
phonia at ITaples by OLTMAIJUS C04, p. 650), and in Corallina
by SOLMS ('01). Professor Pa-rlow states that among the
red algae, "tetrasporic plants are a good deal more co/nmon
than sexual plants, and, in decidedly tlie majority of spe-
cies which I have examined, I have had to look t}irough a
mass of tetrasporic plants before coming to any bearing sex-

ual organs. In the great .Majority of Jlorideae the chances
are decidedly in favor of finding tetraspores rather than
sexual organs . "

IffiTHODS .

Of various fixing fluids e.uployed, the v/oalc clirom-acetO'
osrnic acid mixture was found to "be oest for cytological de-
tails (YA]\tA.TiOUCHI, '06^, p. 425). The time of fixation va-
ried from one to ten hours. Paraffin sections 3 and 5 1^
thick v/ere used almost exclusively for the finer details
of cell structure, G:."osser anatomical features were foiuid
to "he "best made out from mounts in toto . The most success-
ful stain employed was Heidenlaain*s iron alum haematoxylin
{2 hours in the alum solution, 4 hours in the stain) , fol-
lowed "by eosin in QlOve oil, as recomiaended "by Miss Praser
(*07). Por some purposes, good differentiation v/as obtain-
ed "by following the haematoxylin with gentian violet and
extracting the violet with the eosin-clove oil solution.

Difficulty was experienced in obtaining material show-
ing abundant nuclear figures. Plants brought intothe labora-
tory and fixed at all hours after leaving been kept in run-
ning v/ater showed almost no mitoses. J^avorable ma-
dt Ust

terial was. obtained by fixing in the field at eleven or
tv/elve o'clock at night.

VEGETATIVE CHARACTERS.

The thallus forms a hemispherical tuft, and is con-
posed of mizch "branched filsjiients, which are made up of large
swollen cells placed end to end in series. The filaments
radiate from a common point of at ta clime nt, the holdfast. In
a plant of average size, from the "base to the apeoc of a
single filaiaent, exclusive of the iDranches, there are twen-
ty to thirty cells; in large specimens the numter of cells
in a single filament may te tv;ice as great.

The cells differ greatly in 3ha,pe and size in differ-
ent parts of the filaments (fig, 4). Toward the "base of
the plant they are approximately cylindrical below and much
swollen toward the upper end. The cells nearer the tip of
the filament become shorter and relatively thicker, of an
obovate shape, and of a deeper color. Those cells of the
female plant which bear the older cystocarps become very
much swollen toward t}ieir upper ends. In the male pla.nts
the terminal cells bearing the antheridia are almost glo-
bose. The following table gives a general idea of the
sizes of the various cells in a filajnent composed of 20
cells:

llo.'fceU

0

t5

/\t,U

g^i/ia/-

5l».>llS

A\>t>

nitcit- T'-'-'^

7>Un. 1)10.111.

/en.jt/i

TTid-X iriVm

.fl,kn i!i.fcTi>

lexgtK

1

.()t, mm.

.^t -miT

^i TTWl-

. t i'''»»i irA

^

./3

•/9

•r

.i/i''7J11Il

/.2.b^

3

.0.5"

■ H

.*/

.^.r

/.ii"

H

•3a.

. It -mm-

•f

.0^

y.^r

y

.32/

/.3^

Ip

•f

■ A.

IAT

•'/i"

5,.^

7

.4

• Oj

y•^

• Vi'

■ a

CL.O

i

.^i'

•a.

;l.o

.3-

■1^

.li'

2.X5"

7

i-i"

10

.lb"

a.i"

3.0

IS"

•i-2

. /b

3.1.

.6

■li^

3. 0

0.0

•fef

a

i,.^'

.^r

-X.

x.r

The cell wall responds to the usual tests for cellu-
lose. After the death of the cell, the v/all swells greatly
in aqueous fluids. ^iThen so swollen, it shov/^s a plainly
lamellate structure (fig. 5), similar to that described for
Bornetia "by CORREITS ( ) .

The cytoplasm forms a thin layer over the inner face
of the free portion of the cell wall, averaging 61^ in
thickness in the older cells. On the cross vra.ll it forms
a thickened circular pad; adjoining pads are in communi-
cation through the intercellular pores. The cytoplasmic
pad over the upper cross v/all averages in the laj:'ger cells
10 K*^ in thickness in the center, beseeming thinner toward
the edges. On the lov/er cross wall the pad is much thin-
ner, averaging about 3 M^ in thickness. The pads are about
evenly divided into a granula.r layer adjoining the sap-

8

vacuole, and a denser homogeneous layer next the cross v/all
(fig. o) . Nuclei are quite abundant in the granular layer
"but are not of usual occurrence in the homogeneous portion
of the pad. Spherical bodies, probably of a proteid na-
ture, of various sizes, occur commonly in both layers of
the cytoplasm over the cross walls.

In Griffiths ia barbata, BERTHOLD {*Q6) found the cyto-
plasm divided into a clear outer la.yer and a granular in-
ner layer, the latter containing the nuclei and the chro-
me-tophores. This seems to be the usual arrangement of the
protoplasmic elements in the coenocytes of algae. In Grif~
fithsia Sornetiana. the cytoplasm becomes plainly differ-
entiated into two layers only v/here it reaches a consider-
able thickness, as in the thic]<:ened pads mentioned above
and in very young cells in which the sap-vacuole is still
small.

Intercellular connections are conspicuous in living
as in stained specimens by reason of the peculiar plugs
which close the otherv/ise open pit betv/een the cells,
Griffithsia is an unusually favorable form in whic]i to
observe the interce]lul?:r connections because of tlieir
large size. Evidence presented below is believed to be
strongly in favor of the view, doubted by many workers,

that even tlie older cells are c-ctuflly in physical and
organic connection tlirongh the large open pores in the
cross v;alls.

A tj^ical intercellular connection is shown in fig, 6.
The pore in tiie cross walls is closed on each side by a
disk, v.hich is the "stopper", or "plug" of ARCKER ('80).
This disk is in direct conta.ct with the thickened pad of
cytopla-sm lying on the cross wall. Connecting the disks
is a "broad strand of tldn clear cytoplasm, or, in some
cases, several smaller strs.nds (fig. 7.}, In several in-
stci,nces, hits of tlie proteid suhstance norms.llj'- present in
the pad liave "been found in the cytoplasmic strund which
connects neighboring disks, apparently having been fixed in
transit from one cell to another (fig. 8). The middle la-
mella mentioned later as being formed in some cases
in cell divisions lias not been demonstrated in the older
intercellular connections.

Tlie size of the pore varies wit]), the size of the cells
whicli it connects. The average diyjneter of the disks, whitK
is the sojae as that of tl;e pore, is about 11 1^ in tiie large
cells at some dista,nce from the apex.

In living and in unstained fixed me.terial the disks
are refractive colorless bodies. They stain }ieavily v/ith

10

nuclear dyes, rarticulr.rl:' with Heidenhain's haematoxyiin.
Cytoplasmic stains, such as eosin, color them much less
intensely. They are soluhle in Javelle watei-, as was
pointed out by Kienitz-GEKLOFP ('02). This fact, coupled
with the fact that the disks exe continuous on both sides
Y/ith unaltered cytoplasm, gives support to the view, first
ej:pressed by SCHIvIITZ (*83) tlmt they axe protoplasmic in
nature.

The results obtained by vexious workers on intercel-
lulair connections in the red algae are conflicting, ARCHER
(»80) described in Ballia.. pits which are at first open
and later become closed by a plug, or "stopper". SCHLUTz
(•83) came to the conclusion that the pit is closed by a
delicate meinbrane, which is pierced by ma,ny or several
protoplasmic stra,nds, HICK (*84 , a,b) thought he had dem-
onstrated a simple protoplasmic strand passing tlirough the
open pit, MOORE (*85) found that a pit-closing membrane
is pierced by one or several protopleismic strands. WILLE
( »86) described and figured a sort of sieve tube in Cysto-
cloniuin. HARVEY-GIBSON (»91) fo\md that in Polysip}ionia
fastigiata an actual jjrotoi'lasmic connection is present
only in young stages and that later a plug closes the pore-
canal. However, he mentions that from the edges of tlie
plug fibrillsr thickenings connect the neighboring proto-

11

plasts. KOHL in 1902 regarded tlie matter of protoplasmic
continuity in the riorideae as still unsettled. KI3;iTITZ-
GERLOPF in the sanie year found that the pit is closed by
a delicate membrane, and reached the conclusion that an un-
broken connection cannot be said v/ith certa.inty to exist
in the form studied (Polysiphonia) .

The number of nuclei in a. single vegetative cell is
always large. Since the nuclei are a,pproxiiria.tely equidist-
tant in each cell below the apex, it is evident that the
ni;imber in a cell varies directly witJi tlie size of the cell.
Estimates me-de from several prepa,rations show that the
lsj:'ge cells neai" the base of the plant conta,in, on an av-
erage, 3,000-4,000 nuclei. As the cells become smellier
toward the apex of the filajnent, the number of nuclei be-
comes correspondingly less. A subterminal cell of average
size contains about 100 nuclei; a.n exceptionally large sub-
terminal cell may contain a,s me.ny as 500 nuclei. In the
newly formed terminal cell the nuiaber is much less, varying
from 12 or 15 to 50, or even 75. The terminal cells, how-
ever, like the other vegetative cells, are always multi-
nucleate.

The occurrence of multinucleate cells is rather gener-
al in the older portions of the thallus of other Florideae,

12

while the terminal cell is usually uninucleate (DAVIS, »98;
0I.TI'.!A2T1T{? , 'OS; r. «9). SCmaTZ, who first called attention
to this fact (»79a), showed also that even in the different
species of a single genus the numter of nuclei in the cells
iarles greatlj', For example, in the genus Callitliamnicn,
all the cells in the thallus of C . plujiula are uninucleate;
in C. cor^Tnoosum the older cells sjre multinucleate; in
C« Borreri even the youngest cells liave tv/o or more nuclei.
Obviously, then, the nuuiber of nuclei in the cell is no in-
dex of relationship in the red algae.

The nuclei of Griffithsia Eornetiana are pretty uni-
fci^mly distrihuted tlirough the cytopla.sm, THiile the dis-
tance sep8j:'ating them vaj^ies somev/he.t with the a,ge a,nd con-
dition of the cell, usually it is 25-30 f»^ . Hot infrequent-
ly several nuclei, with the cytoplasm imjnedicitely sur-
rounding them, form sma.ll clumps which project into the
central vacuole (fig. 9 ) . In the cytoplasmic pad on the
CEOss wall 10-15 nuclei usually form a ring around t}ie
intercellxilar pore (fig. 10) ,

The size of the nuclei vejries considerably with the
age of the cells, as has been shown by BERTHOLD to be the
case in the coenocytes of Codium (81) . In the young cells
the resting nucleus is, on an average, about 4 f^ in diaroeter

13

just "before nuclear division and less than half that
just after mitosis. In the older cells tlie average diajn-
eter of the nucleus is 2-3 1^ . In tlie young sporelings
the nuclei are very small, measuring 1-2 M^ in diajneter.

The resting nucleus is nearly spherical or smmewliat
flattened against the cell wall. It shows a large, densely
staining chroxn£itin-nucleobs ir the center. The size of the
nucleolus varies from l/5 to 2/3 the diajneter of the nu-
cleus. It is smallest at the time of complete rest of the
nucleus, and grows larger a.s the time for mitosis draws
nea-r. Around the peripher;^ of the nucleus a faint linin
network is visible. This is connected \7ith the nucleolus
by faint radiating strands (fig. 11). Immediately envel-
oping each nucleus is a zone of cytoplasm, which appears
denser than the cytoplasm elsev/here, ajad which is probably
to be considered of kinoplasmic na,ture. The thickness of
tliis zone is omite va.ria.ble. It often becomes a.bout one-
third the diajneter of the hucleus.

ITuclear division occurs regulaj:'!^'' by mitosis, being
found most frequently in the terminal cell. It occurs also
commonly in the subterminal cell, less comiaonlj' in the
third cell from the apex, and ratlier infrequently in cells
older than this.

14

The divisions of the nuclei of a sincle cell near the
apex are almost, tliough not quite simultaneous (fig. 12 ).
In general, the nuclei near the apex of the cell are at a
slightly more advaaiced stage of division than those near
the base .' . For instance, tlie nuclei near the ape:.
ma.y shov/ stages of anaphase, or even of telopiiase, v/hile
those in the middle region of tlie cell are at metaphase,
when the nuclei near the base have reached only the condi-
tion of prophase, (fig, 13). V/iaen tliere is an accujnula.tion
of protoplasm in the apex of the terminal cell preparatory
to cell division, the ni^clei in this protoplasmic mass may
divide considerahlj- before the nuclei of the lower part of
-the cell. In the older cells the nuclei do not shov,' the
same simultaneity of division. Here small groups of nuclei
may undergo mitosis while the majority of nuclei are in the
resting condition. In the younger cells, however, when one
nucleus divides, all divide, though not exactly synchronous-
ly.

In this connection, it is interesting to note the be-
havior of the nuclei in the multinucleate cells of other
plants. In the sexua.1 organs of various Phycom^z-cetes, the
numerous nuclei divide a.t the Scune time, as in the oSgonium
of Saprolegnia (DAVIS, *C3), in the oOgonia and antheridia

15

of Pythliun (MIYAKE, '01), Al"bugo (STEVENS, '99, •00), and
Peronospora (WAGER, '89). Simtiltaneous nuclear division
is reported also in the plasiaodia of PULIGO (IIARPER, '00),
in Plasrnodiophora (UAY/ASCHIIT, '99), in the "ascus" of
Hemiasci (JUEL, '02, POPTA, '99), in the ascus of Ascomy-
cetes (PIARPER, '97) , in the hasiditun of Basidionycetes
(I'lAIRE, '02), and in the "binucleei.te cells of IJredineae
(SAPPIN-TROIIPPY, '96). Aiaong algae, approximately simul-
taneous nuclear division is knovvn in the germinating zy-
gotes of desmids (KLICBAHIT, '91) in the young colonies of
Volvox (OVERTON, '89), in Sphaeroplea (ICLEBAHN, »99), in
Hydrodlctyon ( TIMBERLAJvE , '02) in the anther idia of Fucus
(GUIGNARD, »89). In the vegetative cells of Cladophora,
SERASSL'EGER found that nuclear division is not simultaneous
though he reports that several sta.ges of mitosis s.re to be
found in a cell at the same time. TkiS seems to indicate
that the stimulus to division affects more than one nucleus
at a time. Among the Arclie^oniates , simultaneous nuclear
division is figured hy Miss Lj'^on for [Telaginella ('01), and
seeEis to he the rule in the developing endosperm a,nd in the
early divisions in the fertilized egg of Gymnosperms
(COULTER and CHAIvlBERLAIIT , '01, pp. 20, 31, 41, 83, 98); in
the free cell formation of the endosperm of many Angio-

16

sperms (COULTER and CJIAltBl^LAIF, '03, pp. 165-6, 172), and
in the developing embryo sac (ibid,, p. 67), In certain
Legiuninosae, GUIGITARD ('81) reports simultaneous nuclear
division in the cells of the suspensor. The second mitosis
in the gonotokonts of Arche^oniates is simultaneous in the
tv/o nuclei,

SCHjvIITZ, in his studies on the nuclei of Siphono-
cliadacea.e (*79, b,c) does not discuss the question of sim-
ultaneity of nuclear division, but leaves the reader to in-
fer that the mitoses in a cell do not occur a.t the same
time. The same is true of B^THeiB'S work on Godium ('81),
and of FAIRCEILD'S account of Valonia (^4) ,

Approximate simultaneity of nuclear division may be
said to be a very general phenomenon in multinucleate plant
cells.

The small si/.e of the nuclei renders Griff ithsia a
rather Unfavorable object for the study of the details of
mitosis. The follov/ing account is based on observation of
the nuclei in vegetative cells of the tetrasporic plants.

The nuclei are tliroughout their history very poor in
linin. The chromatin of the resting nucleus is not, there-
fore, distributed on a linin reticulura, but is contained in
a centrally placed, homogeneous nucleolus, or karyosome

^

17

(fig. 11), It seems posaitle that a siiia.ll aj/iount of cliro-
matin is distributed on the peripheral linin net.vork, but
the "bulk of it is certainly in the nucleolus. As the nucle-
us prepares for mitosis, it increases somev/hat in size, "be-
coming about 4,5 1^ in diameter; the nucleolus also enlarges,
ehroiiiatin from the nucleolus, in the form of rather large
granules, passes out to the periphery of the nucleus along
faint linin strands (figs, 14, 15), very much as v/as de-
scribed \)y V/olfe in Nemalion (»04), At the same time, the
nucleolus becomes differentiated into faintly and darkly
staining areas, the latter probably representing cioroi-ie-tin.
The cliromatin continues to pass out of the nucleolus until
the whole cliroma,tin content is distributed throiigh the nu-
clear cavity in the form of granules, some of which are
connected with each other by linin tiireads (figs. 16,17,13,
19,20) . The nui.iber of these granules seems in every case
examined to be more thy.n tv/ice the niuaber of chromosomes
and in some insta^nces the graniiles bedome much more numer-
ous (fig. 17). The granules now a-pproach the centre of the
nucleus, at the saiae time becoming fev/er in number, prob-
ably b^"" the fusion of separate granules, (fig, 21). As
they move tov/cj?d the centre, they become arranged roughly
in a flat plate, t}iougli al i. the granules do not lie in ex-
actly one plane (fig. 2,1 ). VHiile this is going on, a

13

faint spindle is formed, apjiarently by tlie rearrangejnent of
tlie linin tJu'eads (figs, 21, 22). The spindle fibres are
connected with small, darkly staining kinoplasmic caps,
which lie on the nuclear luembrane at opposite poles of the
nucleus (fig. 22) .

At metaphcise the spindle is seen to he short and broad
and more or less truncated at the ends (fig. 22), The nu-
cleus is flattened at right angles to the axis of the spin-
dle, so that it is much broader than long. The cliroirio somes
are closely packed on the equatorial pls,te, which is nearly
as broad as the nuclear cavity. The nuclear meiabratie is
intact, so that the v;hole spindle is intranuclear.

In addition to being closely packed, the chromosomes
do not lie in exactly the same plane, and it has been dif-
ficult to count them witli certainty. Between the chromo-
somes lie a darkly staining substance that renders coxmting
still more uncertain, numerous estimates made from polar
views of the equatorial plates, vary from 11 to 14, The
normal number of chromosomes in the nucleus of the vegeta-
tive cell of the tetrasporic plant seeias pretty certainly
to be 14 (fig, 23) ,

The nucleiir cavity is largest at time of propliase,
jasuring as much as 5,5 M^ in diameter. At metaphase it

me?

19

is c on sideral)ly smaller, averaging 5.5^ broad "by 3 ^
long, A similar decrease in the content of the nuclear
cavity has been noted by YAMAITOUCHI in Polysiphonia ('Oeb).

At metai-ihase the group of cliroiao somes splits into two,
\/hich v/ithdra.v/ tov/ard the opposite poles of the spindle
(figs. 24-28), In anaphase the daughter cliromosomes of
each ga'^oup are seen to be arranged some\/hat in tlie sliape of
a v/atch crystal, with the concave surface tov/a,rd the pole
of the spindle (fig. 26), as v/as figiired in certain nuclear
divisions in ITemalion by TOLPE ('04). The tv/o groups of
chromosomes a,re connected hy a fev/ spindle fibres (fig. 26)

As tJie daugliter groups of chromosomes approach the
kinoplasmic caps, the outlines of the individual chromo-
somes become lost in a dense mass of chromatin, v/hich is to
give rise to the nucleolus of the daughter nucleus (fig.
27) , At telophase the me-ss of cliromatin is in immedio-te
proximity v/ith the kinoplasmic cap. In the meantime the
nuclear membrane, v;hich becomes fainter during the coiirse
of mitosis, disappears, the origina.l nuclear cavity becom-
ing filled v/ith cytoplasm, only a few faint striae remain-
ing of the spindle (fig. 20) . The maas of cliromatin re-
sulting from each group of daugliter cliromosomes becomes
surrounded by a clear area, v/hich is bounded by a faint nu-

20

clear memlDrcaie (fie, 29). The kinoplasi.iic cap grov/s around
tlie dau^liter nucleus, whose organization is now complete.

The a>'.es of the mitotic figures seeias to "bear no rela-
tion to the axis of tlie cell not to the position of the cell
v/all (fig. 21), Y/lien the axis of the spindle is at right
angles to the cell wall, however, the daughter nuclei shift
tlieir position at telophase so that a line connecting them
is parallel to the cell wall.

In the vegetative cells of the sexual individuals the
behavior of the nuclei in mitosis is in general similar to
that in tetrasporic individuals. The nujiiber of cliromatin
granules v/hich pass out from the nucleolus and become dis-
tribu.ted through the nuclea.r cavity, v/hile variable, is
always much less than in the nuclei of the tetrasporic
plant (figs, 30, 31). The size of the nucleus at prophase
is about the same in the two cases. At the time of meta-
phase, however, the cavity of the bucleus of the sexual
plant is somewhat smaller than that of the tetrasporic
pla,nt , The number of chromosomes on the equatorial plate
in the sexual plant is 7, though liere , too, the counting is
made difficult by the presence of a darkly staining sub-
stance between the chromosomes (figs, 32, 33).

Mitoses of the type described above liave been observed

21

in vegetative cells of vsjfious ages, in the hair-cells of
the procarp and cystocarp, in the primary tetrasporic cells,
in the stalk cells of the tetrasrorangia, in the involv.cral
cells of the tetraspore-sorus, in the sporelinrs froM tet-
raspores, and in the sporelings from carpospores.

No undoubted cases of amitosis have "been observed. An
appeara-nce suggesting amitosis ha.s been noted in the stalk
cells of the tetrasporangium and in the cells of the spor-
ogenous lobes, and ma.y possiblj^ occiu^ in t}ie vegetative
cells; but the small size of the nuclei renders exact ob-
servation on tjiis point very "difficult . It may be said
v/it?i confide-n.ce, hov/ever, that the usual mode of nuclear di-
vision is by mitosis.

It may be well to compare at this point the beha,vior
of the nuclei in division v.dth those of Polys iphonia (yAI.TA-
jSrOUCHI) and ITemalion (WOLPE) , the t./o other red algae v/hich
have been carefully studied from the cytological stand-
point ,

22

The cliroriiatopliores are numerous, siiiall, oval or round,
flattened "bodies of rosy pink color l;/ing in the ;;;ranular
cytoplasm next the cell wall. They vary considerahlj'- in
size; on an average, each is about 3.5^ long, 2.5 f^ broad,
and 1,2 (^ tliick. Usually the outline of the chromatophere
is smooth (fig. 11), hut occasionally it is toothed, as is
true in other species of Griffithsia (BERTHOLD, '86) . In
the ;'"ounger cells the chromatophores are crov/ded together
without definite arrangement. In the older cells they of-
ten occin' in curved rows, which are arranged in the form of
an irregular network (fig. 34), as was described for other
species of Griffithsia by BERTHOLD (*06), and for ina,ny gen-
era of the Siphonocladiaceae by SCHOTZ ('79).

In ifiaterial preserved in formalin, the distribution
of the coloring xn^.tter in the cliroma-tophore is similar to
that described by PRIESTLIT/ and IRVING (»07) in the chroaa-
tophores of r:'elaginella and Chlorophytiim. The central por-
tion of the chromatophore is colorless; the coloring matter
occurs in a peripheral layer. V/hen transferred from a 5^
solution of formalin in sea water to distilled water, some
of the chromatophores were seen to have split ini.o two
halves (fig. 35), as was shovm by HAEGELI to occur when
certain chloroplasts were subjected to solutions of less

23

osmotic strength (see PRIESTLJTY" and IRVIITG, '07, p. ).

The nuiaber of chroinatophores in a cell is very lojr^e.
In an older cell of average size, ahoiit 400,000v/ere esti-
mated to "be present.

The chroinatophores in the protoplasi.iic pc^ds lying on ^
t]ie cross walls are Liuch fewer and smaller than

in other portions of the cytoplasm,

Chroinatophores apparently a.re a,hsent from some lateral
cells when first cut off; nor have they been seen in the
young procarps, in the hair cells, in the stalk cells of the
tetrasporangia, or in the young tetrasporangia. While no
leucoplasts have been deraonstrs^ted in these cells, it is
possible that they are present.

In dividing, the chroma.tophores simply pull apart.
They first elonga.te and the pigment collects in each end;
then they assume approximately a dumb-bell shape; and fi-
nally either separate completely, or more usiially remain
connected by a fine strand, as tJiough tJie division v/ere not
quite complete (fig , 36) .

Starch is normally present in the vegetative cells,
as has been found to be true of Florideae generally by
BITTSCHLI ('03), BRUNS ('94), KOLKWITZ (»99), and others.
It occurs as very small granules in circular groups
or as larger granules lying in tlie cytoplasia between the

24

chronatophores , Each stjirch erain is rounded o • oval, usu-
ally with a dark center i no signs of lamination have been
observed (fig. ). Gt;,rch is especially abundant in
sporelings, and in the cells of the attaching orgcui (fig.

).

Besides the starch grains, there are nornially present
in the cytoplasm rounded masses of various sizes of v/h3,t
seems to be proteid inaterial. These spheres usually occtir
in Gmall groups, each group being surrounded by a clear
area. The groups seem to be especially abundant in the
cells at the time of nuclea.r division, and often simulate
miclei (fig. 12, ). Spheres of v/hat seems to be the same
ma,terial are usually present in the pads of protoplasm ly-
ing on the cross \/alls, a,nd small bits have been observed
lying in tjie cytoplasmic strands connecting neighboring
cells (fig. 6) .

Cell division in Griffithsia is reiaarkable for the
disparity in the size of the daughter cells. It .'as first
described by 'WRIGHT (»79), whose account was supplemented
by the observations of BERTHOLD (»06).

In the vegetative cells, division occiu's (1) by the
cutting off of daughter cells from the teruiinal cell of the
filament, (2) by the cutting off of small dome-shaped seg-

25

merits from the upper borders of cells below the apex. The
first tinpe of division simply increases the length of the
filament, the second results in the formation of a nev/
branch.

There appear to be two methods of cell division. The
first occurs most comrnonly in the larger cells, and is al-
ways preceded by an acciunulation of cytoplasm, nuclei, and
to a less extent of ciiromatophores, which forms a dense,
more or less homogeneous irictss in the terminal portion of
the apical cell (fig. 37), A thin dome-shaped membrane is
now laid dovm, v/ith its convexity to.Tard the apex, cutting
a solid accumulation of protoplasm from the tip of the cell
(fig, 38) . This membrane is formed simultaneously over its
v.hole extent. There is no trace of cleavage in connection
with its formation, the protoplasm being in contact \vith
it on each side. The nuclei appear to have nothing to do
with its formation, nor is its formation visibly associated
in ajiy way v/ith nuclear division. The membrane is, howev-
er, neyer formed until there is an accvunulation of nuclei
and cytoplasm in the tip of the cell. The young cell, at
first a solid cap over the tip of the apical cell, grows
rapidly, soon acquiring a central vacuole (fig. 39), and
forming a tirpica.1 vegetative cell. During the growth of

26

the young cell, the cross partition loses its convex ap-
pearance and becomes flattened. It becomes overlaid on
each side \)y a cellulose ^all, which does not cover the
membrane completely. There is left in the center o. circu-
lar area or pit, v/hich is noticeable because of the early
development of the cytoplasmic pliTgs described on page 3
(fig. 40).

This is a very unusual method of cell division in
coenocytes (DAVIS, '04, p. 452-3). So far as I knov/, it
has been described in detail in no other form, though a
somev/hat similar process appears to take jla-ce in the large
vesicles of Valonia (SCHIIITZ, '79).

The details of the division v/hich gives rise to a, lat-
eral branch are very similar to those of the division of
the apical cell. The daughtersegments of the subterminal
cells are formed on the side of the upper border of the

cell, usually four or five cells from the apex. There is
first a solid axcuraulation of protoplasm (fig. 41), then
the adjoining cell .nil bulges outward, and a dome-sliaped
merobrane cuts the outer part of the protoplasmic mass from
the inner, precisely as in the division of the apical cell.
The young segment pushes out and becomes cylindrical, a
vaxuole eoTly appei^tring in its middle, and forms tlie apical
cell of a new brunch (fig. 42). The branch tlms initiated

27

grows for a time more rapidly than the main filament, until
it ahout equals it in size. Thus an apparent, or false,
dichotomy results. True dichotomy'' appears never to occur,
as in no case has a braiTidh ke.^ foxind to divide longitudinal-
ly. Frequently more than one "branch is laid dovm at a node
so that trichotomy results, and in the larger cells near
the oase of the plant, four branches have been observed to
proceed from the suminit of the same cell,

A second laethod of cell division occurs commonly in
the apical cells of the smaller branches, and sometimes in
the division of the larger cells below the apex. A ring of
cellulose projects inward from the cell wall a short dis-
tance from the apeii, very much as was described for Cladoph-
ora oy Strasburger {'88), (figs. 43,44). The ring grows in-
v/ard, but not so as to cut off the new cell completely. An
open circular pore is left in the center, across which the
protoplasioic plugs are soon formed in the usual v/ay. ITo
pit-closing membrane is formed between cells separated in
this manner.

The second mctliod of division, which is the usual one
in coenocytes, differs from the first in that the daughter
cell is from tjie beginning mud. more nearly equal in size
to the cell fro.'.i which it is cu.t tlian is the case in the

28

first :iiethod of division.

ViTierever the second method of cell division occurs,
the partition is in one plane, never arched. In all the
cases exaiained in v/hich division occurred by the ingrov/th
of a cellulose ring, the partition cnt into the central
vacuole, so as to cut oifff a segment containing part of the
vacuole of the mother cell; in the first method of division
however, the segment cut off is at first solid. The rela-
tion of the cleavage plci,ne to the vacii.ole seems to deter-
mine the method of cell division; in the division of the
vegetative cells, where the cleavage pls-ne occurs so as to
cut into the solid protoplasmic accumulation in the a,pex
of the cell, division talces place "by the first method.
^Vhere the plane of division is sufficiently removed from
the a,pex to allov/' the partition to cut into the central
vacuole, division is by the second method.

As mentioned above the nuclei appear to take no part
in cell division. This seems to be the rule in coenocytic
cells (St , '88), tliough '.Ville lias noted an appa-
rent exception in Acrosiphonis ('00).

The number of nuclei in t}ie smaller daugliter cell just
after its formation i.s various. The average number is be-
tween 12 and 20, but in some causes, and always following

the second method cf cell division, the numter is consid-
erably grea-ter. The cell next below the apex ina,y show 30-
250 nuclei.

Branched lia.irs are fretiiiently home on the upjier "bor-
ders of tlie younger cells. There are usually 6 or 7 of
these ra-ound each node on which they occur.

Their mode of origin is very similar to that of the
tetrasr^oric file^ments to be described Iciter. Small papil-
lae arise nearly simultaneously around the upper border of
a cell near tlie cross partition (fig. ), each papilla con'
taining a single nucleus and dense, homogeneous cytoi'lasm.
The papillae are cut off from the protoplast by an arched
membrane (fig. ), similar to tha.t foi-med in the division
of some of the vegetative cells. The nucleus now divided
(fig, ), and one of the da,ugliter nuclei wa.nders into a
bud from the papilla, the bud, with its nucleus, now becom-
ing cut off (fig. ). A second, a third, and somelriimfes
a fourth bud are formed and cut off like the first (fig.
Each of these daughter cells behalves in the same way, cut-
ting off three or more buds, and in t]iis way a tlirice com-
poimd hair is formed (fig. 47). Plach of the terminal cells
divides into two. The basal cell of a hair becomes multi-
nucleate, as do the cells of the firr-t order of branching;

30

the cells of the second order of branching rein3.in uninucle-
ate. The hasal cell is connected wit!:, the cell on v/iiich it
is Lome "b;- an intercellular connectioh of the same t^^e as
that v/hich occvrs bet\/een neighboring vegetative cells.

After they sxe fully formed, tlie hairs elongate great-
ly and become hyaline. Each cell takes part in the elonga-
tion. A vacuole is formed in the cytoi'la-sm v/hich increa^ses
in size as the cells elongate. A very thin layer of cyto-
plasm lies bet'./een this vacuole and the cell wall:; in the
outer ends of the long cells are seen accuiuulations of cy-
toplasm, in v/hich most of the nuclei lie.

The f\illy formed hairs mi;,y rema,in a considerable time
before elongating. Elongation occurs in all the hairs of
a single node at the same time, and seems to take place
rather suddenly. Tlie total length of a hair of average
size before elongation is about 40 >«^ , and after elonga-
tion about 550 u/ . Such great increase in size in a short
time seems to be rendered possible by the fact that the
cells of a hair do not secrete a cellulose v;all until after
elongation has teicen place.

After elongating, the hairs remain for a while on the
plfijit, but fintilly the connection bet;/een the basal cells
and the vegetative call breaks, and the hairs fall off as
a vvhole . ITot infreotientl;- a second crop of liairs is form-

31

ed before t}ie first crop falls off, go tha.t there appear to
be "t\.o sets of hair-like organs" (FARLOV/, '79, p. 132).
Bi'- the time the second f;ot is foriiied, the first set is car-
ried uy the grov/J?h of t}ie vegetative cell to some distance
from the cross paa^tition bet\/een the vegetative colls, and
the sfecond set of hairs is alv/a.j'-s formed bet..een the cross
pBj'tition and the first set (figs. 45, 46).

Individuals vary greatlj'- in the nuxiber of hairs pro-
d^^ced. In some specimens, hairs a.re found on almost every
node in the younger portion of the plant. Again, one may
look over a great ma.nj'- shoots before encountering e. single
set of hairs. V/hat e:<:ternal conditions ejre favorable to
the production of hairs in Griff ithsia is not known.

In the m£.terial examined by litESS SIvHTH ('96), hairs
occurred on the fem^ile plant only on nodes bearing cysto-
carps. Such a restricted distribution is not general,.
Any of the young vegetative cells seems to be capable of
producing liairs, and v;hile hairs occur usually in tlie vi-
cinity of reproductive oi-gans, there seems to be no neces-
sary connection between the t ,o.

The function of the hairs is quite unknovm. They un-
doubtedly ij>crease,.^^_^^ tlie si'rface exposed to the v/a-
ter, and inasmuch as tliey occur especially abundantly in

the neighborhood of the reproductive organs, ..here t}ie pro-
cesses of metabolism may be assumed to be most active, and
are usually absent on the sterile portions of the plant,
it seems likely that they perform the functions of absorp-
tion and respiration, as is believed by ROSEUVIITGE ('Oo) to
be the functions of similar organs in tlie R}iodomelaceae,

Rhizoids are frequently formed from the older vegeta-
tive cells. Protoplasm accujnulates at a spot on the lov/er
half, or near the middle, or even at tiines on the upper
half of the cell (fig. ), and pushes out as a hollow tube
v/ith a plug of protoplasm at its tip (fig. t8),. in the
cytoplasm of a. rhizoid the chromc-^tophores e:re rather few
in number, and the nuclei are smaller than usual in vegeta-
tive cells. The average diameter of a rhizoid is about
80 1^ ; the length 2 irnn or more. The rhizoid secretes a
rather thick cellulose wall. The longer rhizoids become
divided into two or t}-iree long cylindrical cells by tlie
centripetal growth of a cellulose ring such as occurs in
the division of certain vegetative cells.

The rhizoids so formed attaxh themselves tc^ any neigh-
boring object^, curving around it in tlie ma.nner of a tendril
(fig. 49) . In this v/ay the plant is more securely anchored
tlian it v/ould be by a holdfast alone. Further, rhizoids

33

freqxiently 'become entangled aiaong neighlDoring filaments of
the same plant, tlms binding the lov/er parts of the fil-
aments more closely together and rendering them less easily
torn apart (fig. 50). This is especially true of the an-
theridial plants, in which the rhizoids are very richly
developed.

After the rhizoids "become a.ttached, new shoots may
arise from tliem in the same way that lateral branches c.rise
from tlie vegetative cells (fig. ).

Tendrils have been known in the red algae since AGARDH
(*80) fir!3t described them in Hypnea , I'l^xhodea , Rhabdonia,
and other genera. SETCHELL ('96) described tendrils in
laizrencia and Cystoclonium, and stated that they may also
serve for vegeta.tive proiiag3.tion. ITOKDHAUSEIJ (*00) de-
scribed tendrils in Kypnea, Sp;rridia, 8.nd Uitophyllum and
showed that new plants may arise from root tendrils in
Hvjmea.

A process of regeneration occurs in the filament, v/hen,
as is often the case, one of the old cells perishes. Con-
tinuity of the filament is r^^Sstablished in the following
way. An outgrowth from the cell ne>:t above pushes tlirough
the intercellular i>ore, and grows down into the cavity of
the dead cell. The outgrowth is a tube, similar in appear-

34

ance and mode of foriii&-tion to a rhizoid. The cavity of
the outgrowth is perfectly contimious with tjie cavity of
tlie cell froifl which it originates (fig. 51). A similar
tube grows up more slowly from t}ie cell helow and the two
meet near the centre of the old cell cavity (fig. 52).
They fuse at their tips (fig. S3) to form a continuous hol-
low cylinder; the cylinder increases in size and comes to
replace the dead cell exactly. The usual intercellular con
nection is formed at the junction of the nev/ cell witli each
of the t\.o old cells v/hich contributed to its formation.
A similar process of regeneration was described for Grif-
fithsia Corallina by JA1ICZEV7SKI ('76), with this difference
ho'.vever, that in G, Corallina only the cell above txie dead
cell plays a part in the formation of tlie new cell. TOBLER
('03) has shown that a simil&.r process takes place in oth-
er species of Griffithsia and in Bornetia.

This process leads at times to the production of a cell
of very peculiar appearance. When the cell next belov/ tv^o
branches perishes, the lowest member of each branch puts
out a tube (fig. 51) which meets tlie tube from the cell be-
low. The tliree fuse at the point of contact and a Y-shaped
cell results, which is a product of tlie fusion of three dis^
tinct cells (fig. 54) .

35

Griff ithsia niay be anchorec'. to the sulDstratuin either
"by a special attaching disk, or more usually by a tangled
mass of rhizoids. An e-ttaching disk has been noted in
plants growing on Zostera and smooth rocks, but v/hen, as
is often the case. Griff ithsia as attached to other algae
of cylindrical habit, the disk is replaced by a ta,ngled
mass of rhizoids, which are short, thick-wa-lled, and filled
v/ith starch. Inspection of a number of specimens sliov/s all
stages of transition from a mass of rhizoids to a. well de-
veloped att&ching disk. The disk xna,y be said to be formed
of rhizoids in contact laterally. The development of the
attaching organ is described on page Q>Q •

The attaching disk when present is formed of a single
layer of heavy-walled cells, bright pirJc in color owing to
the presence of numerous cliromatophores, densely filled
with protopla,sm, and packed with laj?ge starch grains (figs.
55,56). ?rom it new shoots may arise.

Judging from analogy v/ith other forms (see OLTKAOTS,
•04, p, 64B, a,nd »05, p. 212) v/e may assume tiiat the plant
winters over by means of the atta.ching disks or the mass
of rhizoids at its base. In the spring these give rise to
the plants which reach perfection in tlie summer. Tlie ev-
idence for tiiis is rather negative. 1, The first plants

36

found in 1907 possessed the attaching disk already well
developed, v/hereas v/e know tliat in the sporelings
the tasal disk is developed quite slov/ly, 2, From the
time Griffithsia v/as first found in July development V70.s
e:-tremely rapid, and this seems to jioint to the conclusion
that t}ie pl8.nts drew on some store of. food,

SEXUAL REPRODUCTIOIT.

The antheridie, are distributed as caps over the upper
ends of tlie somev/hat glolDose termina,l cells of the male
plants (fig. 1). They are formed e.s the terminal cells of
short, much branched antheridial filaments. On a cell of
average size there were found to "be ahout 500 of these fil-
aments, each of which produce about 50-75 a.ntheridia, a
total of 25000-37500 antheridia for each fertile cell of
the male plant. The number of antheridia on a single an-
theridial filament, as well as the number of fila,ments pro-
duced on a single cell, varies greatly.

The mode of origin of the antheridial filaments is as
follows: while the terminal cell of the male plant is
still small and not much swollen, (measuring on an average
0.2mm. long and 0.15 in broad at this stage) about 100-200
protuberances arise simultaneously on its apical surface

37

(fig. 57), Each protuberance is at first hemispherical and
about 20-25 »«^ in dieoneter. There ie in each a single nu-
cleus, surrounded "by dense, clear cytoplasm, which is in
free coirununication v/itli tliat of the mother cell (fig. 5C) .
Their formation is not connected \/ith nuclear division, "but
takes place v;hile the nuclei are in the resting condition.
The withdrawal of so manj- nuclei from the upper portion of
the parent cell leaves this region almost free of nuclei
(fig, 57). As grcv/th proceeds, however, nuclei v/ander up
from the "basal region, and become aga-in evenly distributed
in the cytoplasm.

Each primari' protubers.nce is soon cut off from the
motlier cell b;'' b. delicate partition, which is laid dov-m by
the protoplasm in the same way as in the first method of
cell division described on pages 25-2fe{ fig. 59). When its
formation is complete, the prima,ry protuberance divides
several times vertically (figs. 61 ^60), A lateral cyto-
plasmic process is formed, then the nucleus divides by mi-
tosis (figs. 61,60). One daughter nucleus remains in the
body of the primary protuberance, the otlier passes into
the cytoplasmic process. The two protoplasts nov/ become
separated by a constriction of the Iiautschiclit , whose in.
grov/th appears to be aided by the formation of a vacuole

38

at the point of constriction (figs, 60,61). This process
of division continues until there are ahou.t P,00-500 second-
airy protuberances on the apical portion of the terminal
cell.

The protuberances appeaj:* not to develop cellulose cell
walls of their ovm, hut lie in the swollen wall of the
mother cell (fig. 62).

Each of these secondaj?y protuberances gives rise to a
branched antheridial filament. The single nuclei's in each
divides by mitosis, and a partition is formed separating
the two nuclei, and cutting the protuberance into an upper
and a lower cell (fig. 63). The lower or basal cell buds
off other cells above, each uninucleate, to the number of
five or six (fig. 64). These, along with the upper of the
t\/o cells first formed, in tu.rn bud off groups of uninu-
cleate cells, which become the sperma,tia directly (fig.GG .
Thus the antheridial filivnent is a twice compound structure
like a small bush, the teraiinal twigs of which become the
spermatia.

The basal cell of the antheridial filament is some-
what cubical in shape, and ms.y contain ultima-tely several
nuclei. The cells of the first and second order of branch-
ing ai'e uninucleate and are reuiarkable for their sliape.

39

each resemljling a pear with a very long stem (fig, 65).
These cells early "become filled with a large vacuole, the
cytoplasm forming a very tliin film next the hautschicht ,
and the nucleus lying in the apical portion. None of the
cells of the antheridial filament appear to form a cellu-
lose wall. The whole filament is covered hy tlie swollen
wall of the mother cell (fig, 63). X^Ihen the spermatia are
matvu'e, they simply treak loose from the cells on which
they are borne, and float freel2^ out into the water, (fig.
66) . As they become free, the long neck v/hich attached
them to the cell next beloY/ becomes dravm into the body of
the spermatium, which assiuaes an oval sha.pe .

The mature sperma.tium is about 3 M^ long and 2 k/ in
diameter. Its bulk is occupied by a large vacuole, which
is bounded by a thin film of cytoplasm. The single nu-
cleus lies in the end which pointed away from the antherid-
ial filament. ITo chroma, tophores haive been discovered in
any of the antheridial cells. The living sperme,tium is
quite cleitr and somewhat refractive.

It seems of interest to note the fact that the living
spermt-tia appear not to be e:--.truded unless a slight pres-
sixre is exerted on the cells of the thallus. If branches
of an antheridial plant are transferred carefully from sea-

40

water to a slide and left undistiu'Taed, fev/ antheridia be-
come extruded. However, the pressure of a cover glass or
even a mere touch with a needle causes the extrusion of
hundreds of i.ntheridia from each large antheridial cap.
This, coupled v/ith tlie fact that the tufts of Griff ithsia
are feeding grounds for several species of minute Crustacea,
especially of a species of Caprella , seems to lend prob-
ability to the suggestion that the antheridia of the red
algae are sometimes transloca,ted tlirough the agency of an-
imals.

xTo evidence was secured as to \7hether the antheridial
filaments produce successive crops of speraiatia. It is cer-
tain, however, that after a time the antheridial cap ceases
to producespermatia, and the antheridial filaments become
disorganized and brealc away, leaving the globose terminal
cell of the thallus free of any antheridial cells. Vlien
this occtirs, it is usual for tv/o or more side branches to
arise from the svibterminal cell and to begin to produce an-
theridia when 3 or 4 cells long (fig. 67).

The globose terminal cell which has produced one crop
of spermatia frequently prod\ices one or more new branches
from its svimmit, so that this cell becomes again a function-
al apical cell (figs. 68. ).

The procarps occur laterally on tlie nodes near the

41

tips of the filaments of the female plants (fig. 3), They
are produced successively, sc tliat on a fertile branch a
procarp is formed on nearly every node. Their origin and
development has "been described in detail hy Lliss SMITH
(»98), whose account supplements that of PARLOW ('yg), and
SPALDIIIG ('90) .

The procarps are formed from the s.-iiall terminal veg-
etative cells. V/hen a procarp is to be initiated, the ter-
minal cell, instead of dividing in the v/a,y usual in termi-
nal vegetative cells, becomes pushed to one side by a lat-
eral branch of the subterminal cell (fig. 43), v/hich be-
comes the main axis of the filament. The terminal cell
which is to give rise to the proca-rp contains several nu-
clei; it divides into two cells in such a way that one cell
lies partly over the other. The pla.ne of division is
oblique, the inner edge of the partition being somewhat low-
er than the outer (fig. 70) . Tlie lower of the two cells
so formed is the basal cell of the procarp. The upper cell
divides again by a transverse wall to form tlie central cell
of the procarp and the first peripheral cell (fig. 71) ,
The first peripheral cell is cut off from tlie central cell
on the axial side. A second and third peripheral cell be-
come cut off from the upper border of tjie central cell,
with no discernible regularity of position (figs. 72,73).

The fourtli peripheral cell mentioned "by FARLOV/ has not been

only
seen, and must occur.occasionally. Of the peripheral

cells only one has any furtjier part in the production of

edtJ\ of

carpo spores. Often, though not alv/ays,Athe peripheral
cells cuts off terminally a sins.ll sterile cell (fig. 74).

Up to this point tiie niunher of nuclei in e.'ch of the
cells of the procarp varies from 4 or 5 to 10 or 12 or more*
there appear to be alwa.ys more than one. As the procarp
develops, the nuiaber of nuclei increases by mitosis, luitil
in the cells of the matiu:'e procarp the nuclei become quite
numerous. The average numbers are about as follows: basal
cell, 50 , central cell, 45, peripheral cells, 3-30 ,
each sterile cell, 4

The cytoplasm in the cells of the Jroung procarp is ho-.
mogenous and rather dense. A vacuole of considerable size
occupies the center of the basa.l cell, and of the central
cell. Ho chromatophores or leucoplasts have been seen in
the cells of the procarp, though cliromatophores are devel-
oped in the cells of the cystocarp.

It sometimes happens that after the first tliree cells
of the procarp c._re formed the procarpic branch becomes laet-
ounorphosed into a vegetative shoot of tiie usual type, but
distinguishable from otiier vegetative shoots by the fact

43

that the cell at its base reind.ins broad a,nd flat, retaining
the appearance of a basal cell of a procari) (fig. 75) . In
sucii cases the first peripheral cell, v/hich is really a
terminal cell, functions as an apical cell of a vegetative
branch,

Prom the second or third peripheral cell the carpo-
genie branch is formed laterally. The peripheral cell from
which the carpogenic branch is produced is the "supporting
cell" of Miss SfflTH, and is equivalent to the "avDciliary
cell" of HASSEI'iCAI'.CP ('02). In this account I shall adopt
the term auj>:iliary cell, H^Froii i"t a small ununlcleate
cell is produced laterally, which is the basal cell of the
carpogenic branch. This cuts off a terminal cell, v/hich in
turn divides (fig. 74) . The upper of the tiiree cells so
formed divides again, thus fonaing a carpogenic branch of
foui" cells (fig. 76) . The carpogenic branch is bent at
right angles in such a way that the terminal cell, from
\.hich the carpogonium and trichogyne are formed, is usu-
ally in contact with the auxiliary cell. Each cell of the
carpogenic branch is at first uninucleate. From the free
border of th'^ terminal cell the trichogi/Tie is produced as
a club-shaped projection (fig. 76 a) . wlaether a diijision
of the nucleus of the carpogoniiim accompanies the formation

44

of the trichogyne, as has been found to "be the case in
Batracho speriaum (DAVIS, '96), ITeinalion (V/eiFE, '04), and
Polysiphonia ( YAivIAJTOUCHI , »06b), has not "been determined.
The trichOE^me ireaLCnes d. length of 35-40 A*<>
and ■becomes quite slender, with a diajneter of about 1 K-
In the mature trichogyne several granules, which stain like
chroms.tin, are to be observed (fig. 85 ) .

The mature trichogyne is straight or somewhat curved
and sometimes , though not alvmys, slightly swollen at the
free end.

Before and during the forma.tion of the carpogenic
branch, noteworthy changes take place in the auxilary cell.
The cytoplasm becomes denser and the nucleus nearest the
centre of the cell increases very greatly in size. Before
the differentiation of the aiuciliari/- cell, the nuclei may
average 1.5 m/ in diameter. I-Fhen the carpogenic branch is
fully formed, the central nucleus of the auxiliary cell may
reach a diameter of 6.5 Av , A similar change ma,y take
place in one of the nuclei of the other peripheral cells.

The structure of the mature procarp is as follows:
(figs, 79,80). The broad, flat ba.sal cell bears on its
upper border the central cell, which is also broad and ra-
ther flat. The central cell bears usually tiiree peripheral

45

cells, which may or may not cut off sterile cells at their
tips. One of the peripheral cells hears laterally the car-
pogenic branch, which consists of a basal cell, tv/o inter-
mediate cells, and the carpogoniuin v/ith its trichogyne.
The intermediate cells and the carpogonitun are disposed in
a straight line, which lies at right angles to a line pass;*
ing tlirough the basal and the aiixiliary cells .

A.11 the cells of the procarp 8jre multinucleate e>;cept
those of the carpogenic branch. Of these the terminal cell
is at an early stage of development binucleate, one of the
nuclei passing into the trichogyne and later disintegrat-
ing, the other remaining to form the nucleus of the carpogo—
nium. The two intermediate cells are uninucleate, and the
basal cell of the carpogenic branch is usually binucleate.
The connections between the cells of the procarp appear to
be similar in general to the connections between neighbor-
ing vegetative cells.

Mention has been made above of the hairs which usually
occur in the vicinity of procarps.

The small size of the trichogyne and of the carpogoni'*
uxa renders Griff ithsia Bornetiana a rather unfavorable ob-
ject for the study of the details of fertilization, but it
has been possible to madze out the essential facts. A cpcr-
matium becomes attached to the trichog^Tie near its tip (figs.

46

01,33) . The spermatium is either applied directly to the
surface of t}ie trichogyne (fig. 32), or there /nay "be a
short tube connecting the t..o (fig. 80). V/hether the nu-
cleus of the sperinatium divides at this stage, as is the
case in ITeinalion (V/OLPE, '04) v/as not determined though
such an appearance as is presented in fig. 84 suggests tliat
division may occur. A sufficient numher of stages v/as not
obtained to enable ine to speak with certainty on the sub-
ject, but the stages tha,t v/ere obtained seem to render it
higlily probable that the nucleus from the spermatium passes
down the trichogyne, enters the carpogonium, and there fu-
ses with the nucleus of the carpogonium (figs. 82,85).

Immediately after fertilization the trichogjTie becomes
much tv/isted and falls off, leaving a short stump on the
carpogonium (fig. 35) . Since the fusion nucleus stains
very heavily, the details of its structure were not made
out. However, because of this very capacity for taking up
dyes, it is easily distinguished from the other nuclei of
the procarp.

Very soon after fertilization, the carpogenic branch
begins to be withered, and the fusion nvicleus is seen to
be present in the auxiliary cell (fig. 87). TJie actual
passage of the fusion -niicleus into tlie auxiliary cell was

47

not observed. In no case har^ a fusion nucleus been seen in
any of the cells of tlie carpogenic branch other than the
carpo^oniiun, ajid it seems unlikely tjiat it passes into any
of these other cells. The position of the carpogoniu-in in
contact with the auxiliary cell renders it possible that
the tv/o become connected by resorption of the v/alls at the
point of contact, and tliat the fusion nucleus passes di-
rectly into the auxiliary cell, as was suggested by SCKLCTZ
for other species of Griff ithsia, anr' . demonstrated
in Thuretella and Chylocladia by HASSElTCMtP {»02) . In
Polysiphonia violacea the coinmii.nication between the carpo-
gonium and the aujciliary cell is transient ( YAI.IAIIOUCHI ,
*06b), so thii.t it might well be difficult to demonstrate
in such a form as Griff ithsia.

The cells of the ca,rpogenic branch after fertilization
and the passage of the fusion nucleus into the aiuciliary
cell usually degenerate simultaneously, and often the v/hole
carpogenic branch breaks av/ay from its attacliment to the
auxiliary cell, and lies free among the cells of the pro-
carp (fig, Oa) , In one case the lower cells of the carpo-
genic branch vfere seen to have withered before the passage

-nu.cleu-s

of the fusion^into the auxiliary'' cell, villi ch lends support
to the view that t]ie fusion nucleus passes directly into

48

the atociliary cell and not tlirougji tlie cells belov/ the car-
pogonium. Hiss f'TaTH'S account (»96, p. 41) of the wither-
ing of the carpogenic "branch v/as not corroborated in the
present study. She sjrates that the carpogonium first "be-
comes disorganized, "the adjacent cell at the saiae time
apparently increases in size, "but it atlso soon loses its
contents, and in some cases a-ppee.rs to "become disorganized,
while the tv/o lower cells take a deeper stain t^n "before".
As stated a"bove, the carpogenic "branch us^ially withers as
a whole, and hot cell "by cell.

At the time of the passage of tlie fusion nucleus into
the atixiliary cell, there is in the center of the latter a
very large clear nucleus. This is one of the nuclei orig-
inally present in the auxiliary cell. Besides this, two
or tiiree small nuclei are freqi^.ently seen in the peripheral
portion of the cell (fig. 87). These are the reim^ining
nuclei present at the time of the organization of the aiuc-
iliary cell. They seem to disappear during tlie course of
the further development of the aiociliary cell.

The fusion nucleus in the avuxiliary cell is of very
characteristic appearance. It differs from the us\Jial t^^pe
of nucleus in possessing tv/o chromatin-nucleoli instead of
one. It would seem as if the cliromatin from the male and

49

the female piurent does not fuse completely, and that the
nucleoli of different origin reiaain distinct for some time
after nuclear fusion. The "beliavior of the chroi.iosomes in
the eeirly di:?isions of the fusion nucleus was not observed,
though it ',/ould "be of considerable interest to know v/hether
two distinct groups of cliromosomes are formed at this
stage.

The fusion nucleus divides once in the auxiliary cell,
and the two nuclei come to lie in the opposite ends of the
now some-vvhat elongated cell (fig, 89), Between them lies
the greatly enlarged central nucleus originally present.
Each of the nuclei resulting from the division of the fu-
sion nucleus usually shows the characteristic double nu-
cleolus. The aiuciliary cell nov.- divides^ ©ne daugter
cell containing the enlarged central nucleus and a single
fusion nucleus, and the other containiTig^only a fusion nu-
cleus (fig. 90) . The latter may be called the placental
cell* from it the sporogenous lobes usually arise. Figtire
92 shows clearly that sporogenous lobes may also be formed
from the auxiliary cell after tlie placental cell lias been
found; the nuclei entering these lobes are derived from
the fusion nucleus. Very similar behavior ha.s been observ-
ed by HASSEIICAI.IP ('02) in the auxiliary cells of Thuretella

50

and Chylocladia.

During these changes, the large nucleus of the aux-
iliary cell continues to increase in six.e. It becomes al-
most empty of contents, the nuclear outline "becoming less
and less distinct, and finally the nucleus disappears in
the cytoplasm (fig. 91),

Changes also take place in the other elements of the
young cystocarp. The central cell, which at first contains
a large central vacuole, "becomes filled with homogeneous cy^
toplasm with numerous nuclei formed "by the multiplication
of those originally present (fig. o7), Prom the Hides of
tlie basal cell of the procpjr'p soon after fertilization sev-
eral sina.ll cells are cut off successively'' (fig. 91).
These in turn divide, ^xid tiie outer cell becomes an invol-

i;cral ray (f igs.87,?/) , Three to seven involucral rays are
formed; they are of various six-es and ages, and curve up
over t}.e cystocarp so fc,s to cover it almost completely, ex-
cept e.t the top. The structm^e of tlie involucral rays of-
fers nothing especially remi.,rkable . They are distended
sacs, pe-le pink in color ov/ing to the presence of a small
number of cliroma.tophores, A tliiTi layer of protoplasm
bounds a very large vacuole. The nuclei may become quite

numerous, 07 having been counted in a ray of average size.

51

The rays do not form a definite pericarjj, as tliey are not
united at tlxe sides. ITone are produced "between tlie cysto-
carp and the vegetc;,tive cell .'hich bears it.

The placental cell formed at tlie division of the aujc-
iliary cell increases in size and in nuiaher of nuclei, all
of v/hich are the product of tlie division of the fusion nu-
cleus. Small protuber8,nces are formed on its free border,
each of v/hich conta,ins a, single nucleiis (figs. 91,92).
Each protuberance is cut off from the mother cell by an
arched membrane, and the cells so formed give rise by re-
peated division to the sporogenous cells from v/hich tlie
carpospores are formed.

While this is taking place, there is a general fusion
of cells in tlie center oi" the cystocarp (figs. 91,93,94).
The follov/ing cells take part in this fusion; the placental
cell, the auxiliai^y cell, the central cell, and sometimes
the peripheral cells. The result of the fusion is the pro--
duction of a very It^rge, iri-egulairly sharped placenta., on
the upper sur-face of v/hich the sporogenous lobes cire form-
ed (figs. 94,95) .

The placenta contains nuclei from tliree sources: (1)
the original nuclei of the peripheral and au:<:iliary cells,
v/hich appeitr to take no part in spore foi^mation, (2) the

numerous nuclei of tlie central cell, v/Ijich lie in t}ie "base
of the placenta, a region from v;hich no sporagenous lotes
are formed, c.nd (3) tlie numerous nuclei resultinc from tlie
division of tlie fusion nucleus. These last lie in the up-
per region of tlie plcicenta, where the sporogenous lolDes
are being formed, f(,nd appear to "be the on.ly nuclei to enter
tliese. A sii'iilci-r placenta, v/itli niuiierous nuclei of diverse
origin, has been described in Chylocladia by HASnENCAlIP
»02.

At least in some cases, tlie nuclei from the central
cell appesj? to become abnorms,! and break dovm. The cliro-
matin forms a ci^esent -shaped mass applied to the nuclear
membrane on one side (fig. 96), the nuclei swell, their
outlines become fa,int, and finally their contents mingle
with the cytoplasm. ITot all of the nuclei from the central
cell degenerate, and it is often difficult to distinguisli
tliose which remain normr.l from the sporogenous nuclei, es-
pecially in the older cystoca.rps, except by their position
in the cell.

The mode of division of tlie sporogenous lobes seems
to vary considerably in different lobes. The series of
figures fro;.i 97 to 101 gives a fair idea of v/hat usually
takes I lace. Following the division of tlie nucleus of one

53

of tlie protuberances mentioned as being foi^med on the free

surface of t}ie placentail cell cr on the I'pper part of t}ie

placenta, small curved segments are cut off from the outer

t he-
surfaces of the protuberance in miicli . same way as a segment

is cut off from the apical vegetative cell, v/ith this dif-
ference, however, t}iat tlie cells of the sporogenous lobe
are usie-lly uninucleate. In t}iis way a compact tissue is
formed, the cells of which round themselves off, each cell
producing a carpospore. As the cells round off, the spor-
ogenous lobe becomes converted into branched chains of oval
cells. The links of a cha.in are free at the sides, but
connected with eexh other above and below by a narrov/
strand of cytopla.sm; midway bet\;een connected spores occur
callus-like plugs similar to tliose Ijang in tlie pits be-
tween adjacent vegetative cells. From the time when they
first round off, the spores increase greatly in size, v.nti-3:
imuil^y^^hey ore rocidy to bO'Shc^-, -4J%e- -spores ^noac^ro- about
in diameter, the niioloi moaouring ,V/lien the spores are
ready to be shed, their disjneter is about 30-35 Aa^ , tlieir
length 40-54 M/ , the diameter of the nuclei 8.5-9 M^.

This is the usual history of a sporogenous lobe. In
some cases a difference is presented because of tlie fact
tliat the spore mother cells are rounded off at a ver"" early

54

age, so tliat tlie sporogenous lobe is not a compact tissue
when young, "but a group of rounded cells (figs. 95,102),
There is aoiae evidence tliat in this case tlie method of cell
division in the sporogenous lobe is by constriction, some-
what after tlie manner of the second metliod of cell division
described on page 2-7 . The final result is the same in
the t\;0 cases, i.e., the production of branched chains of
carpospores.

The sporogenous lobes of a, single c^'^stocarp e.re of va-
rious £.ges. Lobes with ma,ture spores may be seen by the
side of unicelli;lar sporogenous lobes, and usually all
stages of development may be seen ir a single cystocarp
(fig. 103).

Each sjorogenous lobe is covered v/ith a gelatinous en-
velope, not easilj^ seen until sv/ollen v/itli glycerine or a
v/atery fluid. The individua,l spores seem to be ./itiiout a
cellulose v/all , being enclosed only by the liautscjiiclit .

As a rule, s,ll tlie spores of a single lobe become ma-
ture at the same time. The IjtiKs of the chains break at
the point where the calli.;s-like plugs are developed, the
cor;necting strands B^e drav.Ti into the body of tlie spores,

and the -spores slip oi't of the gelatinous envelope and
float av/ay into the v/ater.

55

The nvm'ber of spores produced ir; a cystocarp can }iard-
ly "be estimated witli certairilty, "because v/hile the ma-ture
spores are being slied, new sporogenous lobes are being in-
aueurated. Prom a study of several cystocarps of average
size, it v^-as found tliat about 6 lobes are present at one
time, with an average of about 40 spores in each lobe, giv.
ing a total of a,bout 240 spores in a norma,! cystocarp.
Undoubtedly the nxiwber of spores produced during the life
of a cystocarp may often be greater than this .

When set free, the spores present much the same ap-
pearance as the tetraspores described on page T3 , They
are oval in shape. Around tlie periphery' is a ;^one contain-
ing rat}ier dense cytoplasm and numerous flattened cliromato-
phores, which sometimes present their edges, but usually
their flat surfaces, to the outside. In the centre is the
la.rge nucleus, enveloped in a zone of homogeneous cyto-
plasm. The nucleulus, which contains the cliromatin, is
usually in the form of 12-14 rounded bodies in tlie centre
of the nuclear cavity. The linin is scanty- in ajnount, be-
ing barely visible axound the periphery of the nucleus.
Between the nucleus and the peripheral zone of chroiiiato-
phores, the cytoplasm is very coarsely vacuolar (fig. 104).

In a few cases, spores liave been noted wiiich liave germ-

56

inated in situ, £,nd v/hich contain tv/o nuclei (fie. 105) .

ITuclear divisions in the cystocarp are of the usual
type. In the divisions of the sporogenous nuclei, tiie niun-
iDer of chromosoraes is ah out 14 (fig. 105a) .

Pood in5,terial appears to he passed into the cystocarp
from the vegetative cell on v/hich it is horne . Figure 105b
shows a section tlirough tlie connection het\/een the he.sa,l
cell of the cystocarp and the adjacent vegetative cell. In
the cytoplasmic pad occtirs a great abundance of the spheres
of food material mentioned on page 2,4 •

ASEXUAL RKPRODUCTION (Tetraspores) .

The tetraspores are formed in a ring around the upper
border of any cell belov/ the a,pex of tlie filament (fig. 4).
The ring of tetraspores appears to encircle the node, fit-
ting snugli" in tlie constriction bet\/een neighboring vegeta-
tive cells. On the outside of the tetraspores a circle of
involucral rays grows up around the sorus. The munber of
these is varie.ble; there are often only 6 or 8; sometimes
there sTe as many as 20. They appear rather late in the
development of the sorus, in some cases the most advanced

57

tetraspores having already ms.tured before the involvicro,!
rays are formed. The rp.ys are expanded curved plates, con-
nected £i,t the hcise v/it]i the vegetative cell, and free lat-
erally and terminally. Usually nei^^hboring rays are in
contact at the sides, so tliat the circle of tetrasjiores is
well screened from without. Each ray is a single cell sim-
ilc-T in appeccrance a,nd in structiu'e to the outer cell of
the involucral ray of the cyptocarp, Y/here the rays sjre
in connection witli the vegetative cell, the pliigs character'
i^tic of the intercelliilar connections elsev/here a.re form-
ed.

The tetraspores £ire formed a,s follows: Around the up-
per border of a young cell below the apex, protoplasm ac-
cuiTiUlates in sma.ll rounded ma.sses, each containing a single
nucleus (fig. 106) . The cell wall near each protoplasmic
accumulc-tion becomes gelatinous, which a,llows the acciunula-
tions to protrude as small papille.e (fig, 107), Each of
these papillae early becomes cut off from tlie mother cell
by a delicate dome-shaped membrane (fig, 108), Each of
the cells so formed, witli its nucleus, increases in size,
and at tlie same time t]ie membrane loses its convex form and
becomes flattened (fig. 109).

The fori!ia.tion of these prima.ry tetrasporic cells seems

50

to tfiJ^e place entirely independent of nuclear division.

On tlie upper torder of each of these cells a finger-
like outgrov/th of cytoplasm is protruded (fie* HO). The
nucleus then divides "by mitosis in the way described for
vegetative nuclei of the tetrasporic plant (fig. 110). One
of the dau.gliter nuclei reiricdns in the basa-l portion of the
cell, the other passes into the cytoplas.'iiic outgrowth,
•qhich as usual becomes cut off from the basal portion by
the familiar arched membrane (fig. 112), which as usual
soon becomes f le,t , Thus is formed a small tvro-celled
branch, v/ith a single nucleus in each cell (fig. 113) . The
lower may be called the sts,lk cell, v/hile the upper is the
tetrasporangiujn or tetraspore mother cell,

Dujr'ing the growth of this structiire, the stalk cell
pushes out another cytopls,smic projection similar to the
one first formed. The nucleus divides by mitosis (fig.
114), one of the daughter nuclei remaining in the stalk
cell, and the otlier passing into the projection, which be-
comes cut off like the first. Thus a second tetraspore
mother cell is formed on tlie ste.lk cell, A tliird a.nd some-
times a fourth motlier cell may be formed in the same way.
The first motlier cell may be regc'j:'ded as terminal, the oth-
er as lateral.

In rare cases, two nuclei occur in the primary

59

tetrasporic cell from its inception, pjid in one ca.se, at
least, t\/o nuclei have been noted in t}ie very young tet-
raspoi'e mother cell. Thir. recalls tlie suggestion of KEY-
DRICH ('02) of a possible sexual significance of the tet-
raspore. Exainination of a large series of developing tet-
raspore mother cells convinces me, hov;ever, thcit there Is
here a purely accidental phenomenon, v;hich has no place in
the normal life history, s.nd which is not to be considered
as analogous in any way to the se:Kual process.

The cells of the tetrasporic branch appear at? first
not to secrete cellvxlose v;^llls of their own. The stalk
cell, with its tetraspore mother cells, remains siirroionded
by the gela.tini7,ed wall of the vegetative cell on which it
is borne. Thiff. wall, much swollen, covers the tetra,sporic
branch completely (fig, 115). It continues to swell and
by the time the spores are ready to be discharged it seems
to dissolve lejr'gely or completelj'' in the sea-wa.ter.

The stalk cell increa,ses in size, and its nucleus at
tlie same time divides b;'- successive mitoses until usually
16 da,ughter nuclei are finally produced. V.'ith tlie grov;th
of tiie cell ir size, tlie cytoplasm becomes less dense, and
vacuoles appear iii it. There may be a single large central
vacuole (fig. 115), or several sins.ller ones variously dis-

60

posed. The connection of tlie st^.lk cell ■..itij the vegeta-
tive cell and also with tlie tetraspore mother cells is of
the upual t^'pe. On the side toward tlie sta,lk cell, the
cytoplasm of tlie mother cell is produced into a rather nar-
rov/ strand, \7hich meets a similar strand from the stalk
cell at the point where the callus-like plugs are develop-
ed (fig. 116).

It sometimes happens t}iat the stalk cell produces lat-
erally a tuhtile-r process that curves up around the mother
cells cJid resemhles iri s,ppea,rance an involucral ray (fig.
117) . This recalls the condition in Griffithsia "barhata
and other species, in v/Iiich the tetraspores c'j:'e "borne lat-
erally on short involucrate reimuli •

The tetraspore mother cell increases in si^'.e, the nu-
cleus showing corresponding enlargement. The cytoplasm he-
gins to shov« nuuierous small vacuoles betv;een the rather
dense cytoplasm surrounding the nucleus and that lying in
the periphery of the cell.

The "behavior of tlie nucleolus during the period of en-
largement of the nucleus is interesting. As the nucleolus
increases in mass, it fragments into several rounded "bodies
of various sizes (fig. 113) . This process of fragmentation
continues until from 12 to 14 rounded masses of cliromatin

61

of about the same si;:e are formed (fig. 119) . These lie in
a cliunp in t}ie centre of the nucleus, staining very iieavily
with nuclecO' dyes.

At this stage the tetraspore mother cell rnay "be con-
sidered to he ns.ture. The length of a ina.tvire mother cell
is about 20 1^, the width 15 A^ , and the diameter of the nu-
cleus 7^. rwther changes in tiie mother cell are in an-
ticipation of division into tetraspores.

From tliis time the che.nges in the cytoplasm occto*
mainly in connection vitJi the vacuolar area. The vacuoles
becoine larger and the whole vacuolaj? area presents a. coarse
spongy appearance. In the meshes are deposited numerous
spheres of substance staining deeply witli haematoxylin.
There is reason to believe that t]iese bodies are derived
from the nucleus. As the time of nuclear division approach-
es, these granules become larger and fewer in muaber, so
that it is possible, by noting their size and nujiiber, to
predict in just what stage of mitosis the nucleus v/ill be
found. The granules seem to be analogous to the ciiromidial
substance of Protozoa (see GOLDS CHI.IIDT, •04) and of some
plants (see TISCHLER, '06).

The cha,nges in the nucleus pjre profound. Host strik-
ing is the decrease in staining capacit;' of tiie nucleolar

62

masses. These "become irregular in form, and at the same
time fuse with one another, so that their number is re-
duced by more than half (fig. 120) , At this stage they are
in the form of thick, curved rods, in v/hich light and dark-
ly sta.ining a,reas rae-y oe discovered. Often four dark areas
may be detected i i each rod, v/liich sviggests that this 3,tage
corresponds to the formation of tetrads in the oocytes 3-nd
spermatocytes of many aniiaals, Coincidently v/ith these
changes, small granules- are to be seen in the nucleus near
its periphery, \/hich seem to pass out into the cytoplasm
(fig. 121), to form the granules alr*eady mentioned as oc-
cuj-'ing in the vacu^tir ai^ea.

The stage just desci-'ibed is considered to be the pe-
riod of s^Tiapsis, It is of long dura,tion, it shows a con-
dition which does not occur elsev/here in the life history,
and it immediately precedes the mitoses in v/hich nvunerical
reduction of the chromosomes takes place. It differs from
the usual type of sjTiapsis in that no spirem or synaptic
tlirecid is formed; but tjiis is not to be wondered at inas-
much as nowhere in the life history of Griffithsla Bornet-
iana is a spirem produced. Perhaps tlie worm-like nuclear
masses are to be considered as replacing the usual spirem
stage.

63

While the thick, irregular rods continue to lose their
capacity for tcUcine up stains, there appear) scattered
throughout the nuclear cavity^ but riiciinly ne;\r tlie periplijry
a niunber of small spherical "bodies (figS- 12^) . In ifi£i,n2'-
instances, 14 of these bodies v/ere countod; less frequently

d.TlJ often tKe- i-vu-mber a,f[>e^re3 to la

only 13 v/ere to be seen;, in the lattor events — one of the

o-reate-r tKa^n tMs.

jp. iiulca ia luZ-.-'ccr tjian the rost, and is probably to be
con-sidered ac roiiroGonting tv.o of tho amallcr bodies — (rfirg,
iS^ , After these bodies are formed, there may be no other
trace of nucleolar ros.tei^ial in the nuclear cavity (fig/-?^)
or the acliroiaatic portion of the nucleolus may be represent-
ed by t».o or tliree faintly staining bodies of irregular
outline lying neix the centre of the nucleeir cavity (fig,7-2^^y*
This is the stage of prophase and the siiiall bodies scat-
tered through the nuclear cavity are the chromosomes.

ITot all the nucleolus goes to form tiie chromosomes .
As already mentioned, part of the nucleola,r substance pas-
ses out of the nucleus and becomes dei^osited in the vac-
uolar c^/'toplasm, and part may remain in the nuclear cavity^
where it forms irregular riU.snes. The part remfj.ining in
the nucleus ultimately disappecirs in the cytoplasm after
the nuclear membrane is dissolved.

04

The details of tlxe orcanii:ation of tlie spindle are
made out with difficulty, Dui^ing prophase, kinoplas:nic
caps e,re formed at the poles of t}ie nucleus l^y differenti-
ation of cytoplasm at these points. In most cases, at the
centre of each kinoi'ilas.vlic mc.ss is a darkly sta,ining body
(fig, 124) prohahly comparable to the "centrosphere-like
structiu'es" of Polysiphonia YALIAITOUCHI , '06h) . In some
cases these are large and prominent; in others they could
not be demonstrated e.t all. They are certainly not perms.-
nent structui^es; tliey seem rather to be t]ie expressions of
some temporary kinoplasmc activity. To them the spindle
fibres are attached. The spindle is entirely intranuclear,
and is probably differentiated from materials v/ithin the
nuclear cavity, as no evidence has been seen to indicate
that the fibres grow in from . ithout, as is the case in
Spirogi'ra (BEEOHS, '06). The spindle is truncate at the
poles and slightly broader at the equatorial plate (figs.
124,125), The chromosomes, •.:hich lay scattered in the nu-
cleajr cavity before the formation of the spindle, now move
in toHard the centre of tlie nucleus, (fig, 126a). H^re
they become arranged on the equatorial plate. Some prep-
arations (fig, 126) seem to indicate that during tliis pro-
cess they become associated in pairs, -..Mch soon sepa,rate;

65

"but on this point it is impossible for me to speak v/ith
certainty at present. The ni'inber of c}iro:nosoi!ies in the
equatorial plate is approximately 14. Tliey are small
roujided "bodies, rather closel^r crowded and not lyin^ in
exactly tl-ie saine plane (fig. 127).

The axis of the spindle seems to hear no constant re-
lation to the axis of the cell. It is .Tiore usual, hovrevor,
to find the long axis of the spind.le coincident v/itli the
long axis of the mother cell. The outline of the nucleus
at metaphase is nearly circular, or more often, slightly
elongated in the direction of the axis of the spindle (fig.
124) .

At anapha,se the chromosomes separate into tv/o groups,
probably of seven each (fig, 120), As the groups of cliro-
mosomes approach the poles of the spindle, the nuclear mem-
brane fades av/ay, and the cavity of the nucleus is oblit-
erated by the cytoplasm. In some cases, however, this does
not happen; the nuclear membrane persists throughout mito-
sis. During anaphase, it elongates and then pulls apart in

the middle (fig, 129), V.Tiether this diversity in the be-
havior of the nuclear membrane is in any way connected
with certain irregularities of development to be described
later, is not obvious.

66

As each group of cliromosojaes approaches the pole of
the spindle, the individual c lor o mo somes unite to form a
densely staining spherical mass, v/Iiich becomes the nucleolvs
of the daugjiter nucleus. V/hen the original iaciear meia-
"brane persists, tlie organization of the daugjiter nuclei is
complete on the separation of the tv;o halves of the nu-
cleus, by which time no trtice of the spindle is seen. 'Vnien.
as is more usual, tiie nuclear membrane disappears tov;ard .
telophase, the mass of chromatin in the i'.imediate vicinity
of the kinopla,smic cap becomes surrounded by a new nuclear
membrane, around v/hich the kinoplasm becomes distributed.

In s^ny event, tv.o daughter nuclei are formed, .jhich
lie at some distance from each other. Each is somev/hat
elongated, v/ith a large, spherical, uniformly staining nu-
cleolus in tlie centre, and frequently v/ith t,o or three
s:iia-ller bodies of cliroinatin in the nuclear cavity (fig. 13oV
The da.ughter nuclei are considera-bly sm£,ller tlian the nu-
cleus of the mother cell. Each is about 5/^ long oy 6K
broad, though the size varies, ITo trace of any intcrntr
clear pc-rtition has been observed between the daughter nu-
clei, which lie uuite freely in the cytoplasm.

The daugliter nuclei do not rejnain long in t}ie resting
condition. In each the nucleolus disappea.rs, and seven

67

rounded cliroino somes, ri^o^a-^ly derived from the nucleolus, ap-
pe£xr in the nuclear cavity (fig. 131) . At the sai/ie time
the nucleus elongates further, and there is to "be seen a
kinoplasiaic ca,p at each end. A sidndle is organized as be-
fore, a,nd the 7 clxroiflosomes arrange themselves in an equa-
torial plate. The division of the t-.;o nuclei is syncliro-
nous, their soies of division lying at right angles to each
other (fig. 132) . At anaphase, two groups of 7 chroLiosoraes
pass to the jioles of each spindle, the nuclear mejahranes
disappearing (fig. 133) . At teloplia-se the cliroino somes of
each group \/hich is in close proxi.niity to the kinoplasmic
cap, fuse to form the nucleolus of the daugJiter nucleus.
A new nuclear ineiaorane is formed around each mass of cliro-
matin and the kinoplasm again hecomes distributed around
the nucleus (fig. 134) .

The four nuclei thus formed lie very near tjie periph-
ery of the mother cell, and equidistant from one another
(fig. 115). Each is a definitive tetraspore nucleus.
Their arrangement in the cell is determined by the fact
that one nucleus alv/ays lies at the point from v/hich the

cytoplasmic strand passes to meet the stalk cell. The
stj^uctui^e of the nucleolus at tliis st.ige is somev/hat dif-
ferent from that of the preceding stages. The cliromatin
mass is usually pl^dnly lobulated. 0\itside and near tliis

68

is to be seen a ;:iuch DJii£.ller, regularly spherical body,
v/hose history I have heen unable to trace. Probably it is
of the same na,ture as the nucleolus, since, \/hen the nu-
cleolus fragments, as it does a little later, the sT;:aller
body is indistinguishable from the other cliro!na,tin masses.

An appearance frequently seen at this stage lends sup-
port to the view that food m^.terial is passed up froi;i below
into the tetrasporangium (see YAIvIAITOUCHI , '06b, p. 424).
The nucleus v/joicii lies near the strand of cytoplasm con-
necting the tetrasporangium v/ith the stalk cell is seen to
be surrounded by a mass of food material, v/hich is probably
derived from the stalk cell. The other nuclei at the same
time lie in clear cj^'toplasm in v/hich little stored food is
visible (fig. 135) .

During the progress of these changes in the nuclear
content of the tetrasporangium, the deeply'- staining granules
in the cytoplasm disappear, so that by the end of ^tiie first
mitosis they are no longer visible. At the same time, the
lajTge vacuoles in tlie cytoplasia give place to smaller, more
regular ones.

Cleavage of tlie cytoplasm begins al./ays when tlie four
nviclei begin to move to\/ard the centre of the tetrasporan-
givun, vjhich happens soon after their formation. The jiaut-

G9

scliicht folds inv/ard along; planes v/hich, if continued to
the centre of tjie cell, ..ould cut the protoplast into 4
equal parts, presenting the familiar tripajr^tite arrangement
(fig, 136). Hov/ever, the partitions are produced inwards
only about tv/o-tliirds of t}ie distance to the centre of the
cell (fig. 137) , T}ie central portion of the tetrasporan-
gium is occupied "by the foujr definitive spore nuclei ,/ith
their envelopes of kinoplasm v/hich ai^e in contact v/ith one
another, so the.t e, rather definite nucleo-kinoplasmic fliass
is formed. In the very centre of t]ie tetrasporangiujn lies
the portion of the undifferentiated cytoplasm enclosed \)y
the nucleo-kinoplasraic mass (fig. 138).

The nuclei at this stage are either spherical (fig.
139), or somewhat "biscuit shaped, the inner svjfface "being
less convex tha.n the outer (fig. 140). The nucleolus has
"by this time .fra-gmented into 12-14 granules, similar in
appearaiice to those in the nucleolus of the tetraspore
motlier ce;i.l before s;mapsis.

At tjiis point of development, 5-10 "^ of the tetra-
sporangia begin to sliov/ degenerative changes and do not de-
velop further. The outer sur'face of the tetrasporangium be-
comes^ery much flattened, almost v/af er-like . The entire
contents of the protoplast stain very heavily, ov/ing to the

70

j-iresence in tlie cytoplasm and in the nuclei of nuiiierous
dark crauules. Tliece deseneratinc tetro.sporangia are eas-
ilj^ distinguishable even in the living condition "b- rea,son
of their almost hlaxk opaque appearance, and some ere to
"be found in every tetraspora,ngial sorus . V/liat causes lead
to tlieir degeneration I have not "been able to determine.

In the normal tetrasporangia, the cleavage partitions,
which represent folds of the hautschicht, but which appear
in section as a single line except near the periphery
(fig. 137) , split so as to reveal clearly their double na-
ture. At the sajne time the cytoplasm, which lay in close
contact witla the partitions, separates along the line of
the partitions so that the cleavage furrov/s become wide
as well as deep (fig, 133), Even at this stage, ]io-wever,
tiiey ejKtend no further in than to the edge of the nucleo-
kinoplasmic rna.ss . Coincident witii tliese changes in the
partitions, the nuclei ..iiich were fl^ittened become again
approximately spherica,l. Vacuoles develop in the cytoplasm
in the centre oiff the nucleo-kinoplasmic me^ss . At tlie same
time, siXoll cliroj.iatophores begin to appear in the cytoplasm
along the outer border of the tetrasporangium. These in-
crease in size, and ei few extend along tiie partitions into
the body of the protoi.last . In this condition tj^e tetra-

71

si-orangium re::ic,ins for a lone time, increasing in cir.e and
in vacuolization of the cytoplasm. The significance of
this incomplete separation of the spores probably lies in
the fact that food material seems to pe.ss v.t tlirough the
"basal cell. If the spore v/ere completely separated before
ma.tiu'ity, only one of the fovr v/ould be in communication
• ith the stalk cell, the source of supplies. Kov;ever, in-
asmuch as the cliroma,tophores at this stage are v/ell devel-
oped, it seems probable tlmt the tetrasporangium is capp.ble
of ela.borating at least part of its food material for it-
self.

BERTHOLD (*86) seems to have been the first to point
out this incomplete separation of the tetraspores of red
algae after tiie division of the nucleus of the mother cell,
though SCffi.'iITZ ('79a) ha,d given an account of tv/o succes-
sive nucleajT' divisions in the tetraspore mother cell.

The tetra-sporic branches are from their inception sur-
rounded by the swollen v/all of the vegetative cell on ■, hich
they are borne. A portion of tliis \7all is carried out by
the developing tetrasporic cells. As the cells develop,
the portion of the .all surrounding them swells greatly and
appecirs to become gelatinized, ceasing to respond to tlae
tests for cellulose.

The tetrasporangivun, with the incoaipletely separated
spores, increases niarkedl;'- in size. The nuclei also en-
large and show abundant chroinatin, in tlie form of the 12-
14 masses already mentioned. Each mass is differentiated
into lightly and darkly staining areas, "ot infrequently
the niunber of these masses is greater thaji this, as many
as 20 liaving "been counted in some cases i and sometimes the
numher is considerably less than 12. This variability in

the number of cliromatin masses in the resting nucleus

e y: fi L D L ^
serves to shov; tiiat they do not flh?~^f-e the same constancy in

numbers tliat the chromosomes shov/, and therefore are not

to be relied on as an index of the condition of the nucleus^

v.heth.er haploid or diploid.

As the tetrasporangiiun enltj'ges, t'ne cytoplasm becomes
more coaxsely vacuolate, and the vacuoles in the central
protoplasmic mass become conspicuous (fig, 14) . The pcU"-
titions now grow in until tliey meet in the centre of tlie
tetrasporc'-ngilun, their ingrowth being appai'ently aided by
the i^osition of the large central vacuoles already mention-
ed (fig. 142), The spores are now completely separated,
\,it3i the nuclei in tie iuner cornei's. The nuclei v.anaer
towciJT'd tlie centre of tlie spores, tJie c]a"'omci.topliores at
the same time migrating so as to line tjie entire ierii"!liery

73

(fig. 143), and t}ie spores round off, becojning oval in
shape. The lowest sjore, up to tliis time attached to the
stalk cell hy a slender t}iread of cytoplasm, hreaJcs away
at the point of attacliment, and the strand is v/ithdra\vri
into tiie "body of tlie spore. The gelatinized cell wall, nov/
very much sv/ollen, appebjr*s to dissolve in the sea v/ater,
and tlie four spores are set free, aLmost iimnedij-.tely "be-
coming spherical. Like tiie carpospores, they are heavier
th.an sea-vvater, and slowly sink, if left undisturhed ,

The ras,ture tetraspore resembles the ca.rpospore in ap-
pearance. It is approximately spherical. In the centre
the large nucleus is conspicuous, v.ith its ch-romatin seg-
regated into 12-20 small masses. Immediatelj'^ around tlie
nucleus is a zone of rather dense cytoplasm; outside this
the cytoplasm is coa,rsely vacuolar. In the perijiheral cy-
toplasm is a single layer of chroma.tophores, outside which
is the limiting membra.ne of t}ie spore. Fo cellulose cell
wall is visible.

The average size of tlie tetrasporic structures is
sliovm in the follo-vdng table:

74

ii'-X.o;^ ;i.oXa.^/^ ^oxa-VA^ a-5--3or 40-5^^/^

The tetrasroric structures in a single Gorus are of
very various r.ges. While the first-formed tetraspores are
developing, nev/ tetre.si^ore motlier cells Bjce "being formed
nearer the cross v.a,ll betv/een the vegetative cells, tlie
older tetrasi'Ores being cea^ried av/ay from the cross parti-
tion by the grov/th and the stretching of the v/all of the
vegetative cell. A longitudinal section of a sorus shows
primary tetrasporic cells being formed very near the point
of junction of the vegetative cells; outside these are the
older tetraspore motlier cells; farther out mature spores
are to be seen; while farthest from tlie center occur the
involxicral rays (fig. 144).

The number of tetraspores produced in a single sorus
is quite laj'ge . The average number in v/ell developed sori
was found to be about 300.

The process of nucleair o.ivision in the tetrasjore
mother cell of Griff ithsia offers ^'lany striking points of
difference from the sajne stages in the life liistory of

75

PolysipLonla; it rese.'ubles much jioi-e necirly similar stii.Ees
in Corallina (DAVIP, »9a) . As tiiese tliree fprms are the
only members of the R}iodoph;ceae in v/hich the behavior of
the tetraspore mother cells hs.s been carefully studied from
a cytological standpoint, it may be well to suimnarize here
some of the points of resemblance and difference:

akroTwouin^ U^fTs, pee a.t . '

the e-t^Ts

kromosonves -r^u.f.leo La.r- boiiTes o A- -.-o -m a t t -k. ^-t'^-^u/es 5 ,j> i t^ e -x^

, (^ s e

Fe i. 6 ^ A a- s e

C entire S»>*\e or K'<-'«<'/'^'^-5'»*^''- '^'^A 4— i.l„^ i<.i^c p l ^if^^^- ca-k

i^eYul e-r

0/- Tt\9t AC I' >H«.£^C e U.5

lU.ee.^ ,'^LK^t^^r^it-e:^ n5./,;>e.r^ be-^«^e ^.r^u^U tkr.^.L

■Yy\e>»v Dr-».-He -meXoLbkii.se °/ f'^^'sr , . .' 4 fl DolK -m i ( o S &X

76

This coupaa''ison serves to emphasize one xoiJ^t. rartic-
ulai'ly. At a critical stage an the life history of rather
closely related members (Polysii<honia and Griffiths ia) of
a liighlj^ specialized ci'oup, the cytological ihenomena are
of a most varied nature. Turing the jieriod of synapsis
and up to the time of the formation of the cliromosomes,
the cytological events in Pol3''sii)}.onia are far more like
those in Liliim than those in Griffithsia or Corallina.
The behavior of the nucleus in tlie foz-mation of the tetra-
spores in Graffithsia is much more similair to that in Cor-
ailina than to that in Polysiphonia, a more nearly relented
genus. The bearing of these facts is obvious; cytological
phenomena cannot be considered trustworthy guides to rela-
tionships .

77

TETRASPORE LIKE STRUCTURES ON SEXUAL PLAITE .

As stated a"bove, one individual lias "been found v/liich
produced normal antheridia on the niajority of its f ilaiaents,
but v/}iicii produced on a considerable niunber of filauents
structures resembling the sori of tetraspores . As this case
is of considerable theoretical interest at the present
time, I sliall nov; give sorae of the details of the structure
of this plant.

The portion of the plant berring antheridia './asper-
foctly norinal in appearance. The cells v/ere of the usual

size and siiape, and bore antheridia of the normal t^^pe in
abundance. The munber of chromosomes appearing in Liitoses
in the antheridial fila::ients v/as found to be about seven
(the reduced nxnaber, characteristic of the ga-metophyte.
Mitosis has a.s yet been seen in only a single nucleus of
the portion of the plant bearing the tetraspore-like struc-
tures. In this case, in the dividing nucleus of the stalk
cell, the number of cliromosomes was seen to be seven (see
fig. 14^.

The branches bearing tetraspore-like structures are
of the saxae ti^e as those of the normal tetrasporic plant,
but are composed of cells tliat are on an average a good ^«-
deal smaller.

78

The ef,rly development of the tetras; ore-like struct\ire5
is siLiilar in every detail to corresponding sta,ges in the
development of tiie normal tetraspores. Small uniin\iclea,te
papillae are cut off from the ui>per "border of the vegeta-
tive cells. Each divides to form a short twO-celled "branchy
the lower cell repx^esenting the stalk cell of the tetra-
sporangian, the upper corresponding to the tetraspore
mother cell (figs. 145,146) . The stalk cell increases in
size and becomes vacuolate, the mother cell also becomes
larger, but reiiiains somev/hat smaller than the normal tet-
raspore mother cell. The average diameter of the fully
formed itiatiure mother cell is about 20-22 W as against 24 K
for the mother cell of the tetraspore. The nucleus in the
t.7o cases shov/s the sajne configuration, but remains smaller
in the mother cells borne on the sexual plant, (fig. 147).
ITuclear raa.terial passes out into tlie cytoplasm where it
forms small darkly staining granules,

Involucral rays are formed in the v/ays chaj'acteristic
of the tetraspore-sorus, usvially as outgrowths from the
vegetative cell outside the ring of spore mother cells,
exceptionally as lateral outgrowths from the stalk cell
(fig. 148).

The furtiier development of the mother cells on the

7S

sexual pli-mt differs strikingly from t.jiat of the norrual
tetraspore inotlier cells. In the majority of cases, the nu-
cleus divides (v/liet;ier "by mitosis or amitosis I hove not
yet been able to deterniin^^ and cleavage begins at the
periphery (fig. 149) , The cleavage furrows do not advance
far into the body of the mother cell. The surface of the
cell begins to shov/ irregular v/rinkling, and degenerative
changes set in similar to those described for certain tet-
raspore mother cells. The number of nuclei in cells in
v/hich cleavage furrows begin is usually 4-3, of which some
are very much larger tlian the rest (fig. 149).

One case deserves sjiecial mention. Sixteen nuclei
lie scattered in the cell, which shows no trace of the for-
mation of cleavage furrows (fig. 150) , The ./hole cell pre-
sents the appearance of a germinating spore. It would seem
that here the eell corresponding to the tetraspore mother
cell behaves as a monospore; tjiough whether such a cell
ever produces a norms.l plant is uncertain.

The Ciiromoso):ie -hi story of the nuclei of the cells
just described has not yet been determined.

80

GERIflKATIOlT OF PPOKFIS .

The srores germinate readily in the laboratory. If a
mature tetrasporic or cystocariiic plane be ilaced in sea-
Y.ater ocer niglit, young sporelings up to the tlur-ee-celled
stcige v.ill "be found abundantly attached to the bottom and
sides of tlie vessel tlie next morning. Many of the sta^ges
of germination here described were collected in the field
under natural coniffitionG, but the majority'' of the figtires [•;
given, especially of the younger stages, vere teJcen from
me-tei'ial cultivated in the laboratory.

The similcirity of the structure of the carpospores
and the tetraspores ha,s been noted above; the phenomena of
geriiiMation are also practically the same in the t-..o kinds
of spores. On being released, the spores become spherical
and settle slov/ly in the v^ater. They appear to become at-
tached to the surface of laij solid body they touch, such
as rocks, glass, other algae, s,nd even sucji soft bodies as
the gelatinous si-bstance enclosing c}ia,ins of diatome.

Ptiring the progress of germination, scon after the
spore lecomes attached, there is formed aroimd it a cellu-
lose -.all of the I'sual type, which becomes tolerably tliick,
especially ai'-oimd the basal region of the spoi*eling. At
the sajne time, nxjunei-ous starch grains also become visible

81

in the cytoilasm of the cjore. Several hours after the
spore is shed, tlie m^cleus divides "by mitosis. During this
time tliere is no noticeahle change of sha,pe in the sj ore .

Opportunity lias not occurred for tlie examination of
a large series of dividing nuclei in tlie sp'orelings, but
in the cases exai.ained, tlie mitoses v;ere of the type usual
in vegetative nuclei. In the dividing nuclei of sporelings
from tetraspores, about 6 or 7 cliromosomes appear on the
equatorial pla.te (fig. 155a) . In the sporelings from car-
pospores, the nizmber of cliromosomes is e.lways greater tiian
this, but appears to be less than the nvunber that might be
expected (14) . The small size of the nuclei in the spore-
lings renders exact counting of the chromosomes very dif-
ficult, but it ma^'- be stated with certaint3' that as many as
9 chromosomes appear on each spindle in the mitoses of the
nuclei of the sporelings from carpospores (fig. 155b) ; and
this is believed to be sufficient evidence for regarding
these nuclei as •©# diploid in chsjr'acter .

The daughter nuclei vvi tlidrav/s to opposite sides of
the spore . :<gig. 154), and divide again to form 4 n\iclei,
v;hich in turn divide to form 8 (fig. 156), tlien 16 (fig.
157). The increase in the number of nuclei is not accom-
panied by a corresponding increase in the size of the
spore; the size of the nuclei becomes less v/ith each sue-

82<

ceedinE division. At about this time, the sporeling
changes its shape, pusliing out, at the point of attachment,
a small rounded projection, (fig. 158) i.liich later becomes
cut off as the basal cell. Iminediately after this projec-
tion is formed, tlie spoi^eling elongates, becoming about
t\dce as long as broad, but witiiout undergoing cell divis-
ion (fig. 159) .

As these changes take place, the cytoplasm migrates
more c:n6. more to the periiihery. Tlie central part of the
sporeling is occupied by sm&,ll regular vacuoles, \.hose ex-
ceedingly thin walls are rouglily hexagonal in secfeio.M (fig.
160). A single large central vacuole such as is character-
istic of the vegetative cells of the older plants is not
usually foi^med until the sporeling reaches the 3-celled
stage.

Cell division occurs usually wiien aboprt 16 nuclei are
present. A v/all at right angles to the long axis of the
spoi^eling cuts off a small basal from a large aidcal cell
(fig. 161), Shortly after this, without further elongation
of the sporeling, tjie ti] ica.l cell divides into two by a
wall parallel with the first (fig. 162). The details of
the formf.tion of these walls v/ere not follov/ed out. As
the partitions have not been seen to assume the arched

63

shape chai'acteristic of the first type of cell division at
other points in the life history (see p , 2.5 ), it may be
infei-red tliat they are formed by the in^rov/th of a ring of
cellulose (p. 27^ ), such as is formed in the divisions of
the cells of Cladophora. From trie tv/o divisions, a small
obovate 3-celled sporeling results, which consists of a
sma.ller somewhat pointed basal cell, and t./o larger rounded
cells towards the apex, the tliree cells lying in a rov/
(fig. 165), At this stage, the ciiromatophores and proto-
plasm are so closely packed in the peripheral portion of
the cytoplasm that the sporeling appears dense and opaque.
The pointed end of the basal cell is filled only with ho-
mogeneous protoplasm, sttirch grains and chromatophores be-
ing absent from this region (fig. 162). Intercellular
connections were not deiaonstrtited in the small 3-celled
sporelings, though tliey are apparent after the enlarge.'aont
of the cells.

The description tjiven above applies to the r.uijority
of sporelings examined. Kany sporelings^ however, show va-
riations from the type described. For instance, it happens
rather frequently that cell division occurs after the for-
ras,tion of only four nuclei (fig. 161), wid before the
sporeling assumes the elongated shape represented in fig.
159.

84

The 3-celled. stage is first attained in a'bout 12 hours
from the time tlie spoi-e is slied. It persists, under labor-
atory conditions, for several days, during which time the
three cells change greatly in size, shape, and H.ppear-nce.
the ifiost striking changes occur from the second to the
third day after the spores ai^e shed. The cells increase
greatly in size, and a large central vaxuole is formed in
each. The protoplasm a.nd inclusions "being spread over a
Itjrger area form a thin filia, in which tlie cliroma.tophores
core no longer crowded, hut are separated by considera,ble
deer spaces; the sporeling therefore becomes lighter in
color and more transpBjr^ent , The following comparison of
spoi-elings of the second with those of the third day gives
some idea of the grea.t increase in the size of the cells.

G.^icu.L <-eL(. .'337 'f^^^- -I^S >''-'^- .alcjtr^^^- .073

bas 0- L ccLL -oHl , /07 .of -073

totc^L -/2j .7/

The number of nuclei in the cells, even after enlarge-
ment, is sirif.ll . Several counts indicated that there are
in the ajical cell on aii average 25-30 nuclei, in the ioid-
fle cell 20-25, in tlie basal cell 5-10. These nuclei in

6S

the restinc condition £ire very snie.ll, averacing, perhajiij , /•^'/^
in diejneter. In structm^'e they resemble the nuclei of the
older vecetative cells.

The chances in the apical and middle cells consist
ma,inly of (1) a e^'es-t increase in length, (2) slight inb-
crease in breadth, and (3) the distribution of the proto-
plasm and inclusions over a much lejfger a^rea.

In the hasal cell, besides an increase in size, the
most striking changes are those of shape. These changes
depend to a large degree on tlie substratimi. In ca,se the
sporeling is attached to some soft body, such as another
alga, the bg-sal cell remains somewhat top-shaped, with the
pointed end 5,pplied to, or in some cases vredged into the
substratum (figs. 164,165,166,167). If, however, the
sporeling is attached to a hard body, such as glass or stonc-y
the bascil cell becomes greatly elongated, in some cases
coming to equa,l in length ?,11 the rest of the sporeling
(figs, 168,169). V/lien this occurs, tlie basa,l cell resem-
bles very strikingly a rhizoid of tlie older plant.

MlTien the stage just described is readied, there seems
to come a natural pause in the life-cycle. In a state of
nature, a great majiy more sporelings etre found in the 3-
celled stcige tlian in any other, indicating that this stage

86

occupies a longer time in the course of development tlian
any other. Under laboratory conditions, the 3-celled stage
is retained at least several days, and frequently develox^-
ment goes no further. The factor determining furtlier de-
velopment seems to "be, in v-d^rt at les.st, the cha,racter of
the suhstratum. In the cultujr-es exairdned, it -..as found 4ii
that on glass or on clean, though rough stones, the basal
cell continued to elongate, though without further division
of the apical cell, until the whole sporeling lost its nat-
ural color and died. However, i^ji case the elongating basal
cell came in contact v/ith some soft substance, it fastened
itself iimnediately , and norme.l development proceeded. In
a state of nature, young sporelings of Griff ithsia have
been most coi-oinonly found at V/ood's Hole, on Champ ia pa.rvula
and on Lomentaria uncinata, though they occur on other al-
gae and on Zostera marina. Young plants were sometimes
foxmd on stones, the surface of v/hich a^ppeared clean, but
proved., on cai^'eful examination, to bear other sporelings,
to which the plantlets of Griff ithsia were probably e.t
first attached. There is no evidence of parasitism, how-
ever, in the early development of Griff ithsia. Sporelings
floiirish on any soft substratum, such as bits of cotton
cloth. From tlie observations noted above, it seems clear

88

that Griff ithsia Borne ti ana needs some other substra,tuin
than the stones on v.hich tlie inatwe plant is often found,
to pass tl'irough the early staces of its eristdnce.

When the sporeling is growing on some other alga, the
"basal cell mav simply hecoiie attached to the surface ■by-
some adhesive at the surface of contact (fig. 164), or may
grov/ in betv/een the cells of the s.lgal substratum (fig. 165)
or may even tv/ine about it in the ma.nner of the rhizoidal
tendrils,

Further development of tlie basal cell results in its
division into cells of various sizes and irregular shape.
Usually short tubular projections resenbling rhizoids, be-
come cut off (figs. 170,171) by the circular ingrov;th of
the cell v/all , Somewhat less frequently dome-shaped seg-
ments are formed on. the sides of the basal cell (fig. 172).
In either case a multinucleate lioldfast, or attaching disk
is formed, all the cells of \/hich are derived from the di-
vision of the ba.sel cell.

The apical cell cuts off daughter segments ir; the
usual manner (fig. 1'^^) . By the time tlie sporeling is
four or five cells long, lateral branches appear on the
upper borders of the cells below tlie apex (fig. 175) . The
rapid grov/th of the lateral branches gives the characterist-

89

ic false die ho torn;'- to the yovng t}iallus. Branching is pro-
fuse neur the base of the sporeling. Frequently five or
six lateral "branches are civen off from ea,ch of the lov/er
cells, so that the youn^^ ;^lant is copiously branched. In
this event, the cells bearing numerous branches becomes
thick-v/alled and almost globose in shape.

It is interesting to note that \/hen niuabers of spore -
lings ai-e found in iioinediate vicinity in natu-re, all are
often at fefe© precisely the same sta-ge of development. "For
instemce, about 15 sporelings v/ere observed on a single
branch of Lomentaria; a,ll v/ere at the stage of gennination
represented in figure 174,

Hairs tire usually v/anting in the young plants; nor
are rhizoids developed except from the basal cell.

The phenomena of germination noted above agree in all
essentia,ls v/itii the account of Griff ithsia Borne tiana. given
by Miss DERICK (»99), and ojfe in line with the pjienomena
reported in other species by TOBLER (*07).

It is of considerable interest that the coenocytic
condition characteristic of the cells of the mature plant

is attained in the snoreling before any sign of cell di-

(1)
vision or differentiation. The recapitulation theory

(1) "A highly organized plant, v/hich begins its develop-
ment v/ith the simplest stages and gradually adva,nces to a
state of higher differentiation, repeats in its ontogeny its

so

has "been shown in iiic.ny cc ses to "be applicable to otJier
plants, e.g., in the forms.tion of the mecjaspores of the Hv-
dropteridineae (STRASBIIRGRR, »02), in the forms of juvenile
leaves (BERRY, *06), in the post -embryonal otaces of t}ie
Laminariaceae (SETCHELL, *05), in the formation o£ t}ie eggs
of the Pucaceae (for the facts, see OLTIJIAITIIS , »05, pp. 47-
3) . If this theory is c-t all applicable to Griff ithsia,
.e should expect some evidence of it at the times in the
life history \;hen the plant returns to the unicellular con-
dition. If there is virtue in the conclusions drawn from
coiviparative morphology, the ancestors, and even tlie compar-
atively recent ancestors^ of Graf f ithsia, possessed iminu-
cleate cells. The coenocytic iiabit ^.as acquired late in
the history of the race, and v/e -Should exjject it, there-
fore, to appes,r late in the history of the individual, so
that the cells of the early stages v/ould be uninucleate;
yet in the germinating spore of Griff ithsia the first vis-
ible change is the attainment of v/hat we luuct regard as
the recently acquired coenocytic habit. In this respect,
then. Griff ithsia does not conform to the recapitulation
theory.

phylogenetic development." Strasbitrger, IToll, Sclienck,
and Schimper: A Text -book of Boto.ny, 2nd English edition,
1903, p. 49.

91

VEGETATI VE iTLTLTI PLI CATI ON .

Griff ithsia Borne ti ana ina.y reproduce itself ve(;eta-
tively in tv/o v/ays: first, by accidental isolation and
subsequent crov/th of single cells or small pieces of a
filament; second, by tlie production of new plants from ten-
drils.

The first method of propagation v.as described for
G. Corallina by JAITCZEV^SKI (*76) and mentioned as occurring
in G. Borne tiana oy TARLOW (•79). More recently TOBLER
('03, *03a, *02) has called attention to the fact that
Griff ithsia, Bornetia, Dasya, Polysixhonia, and other forms
may reproduce themselves under laboratory conditions by a
l)rocess of fragiaentation of the filaments and grov/th of
the resulting portions into nev7 plants. In G. Borne tiana
this process taJces place not rarely in nature. In such
cases the isolated cell produces a rhizoid from its base
and a new growing point from its apex (figs. 151,157).
The rhizoid is formed norraall;/ in the raa.nner already de-
scribed. The apical cell is produced b;-- the accumulation
of protoplasm at the tip of and subsequent unequal division
of the parent cell in tiie usual me.nner.

Vegetcitive propagation by means of tendrils lias heen
described on page J3 .

3Z

nSCUSSION OF RESTTLTS .

Prom the cytological evidence brouglit forward in this
paper it seems probable that there exists in Grif fithsia
Borne ti ana an alternation of generations similar to that
which has been suggested for Polysipiionia violacea (YAil-
A2T01TCITI , '06). The fusion nucleus, which contains 14
chromosomes, with the cooperation of the c^rtoplasm of some
of the cells of the procarp, produces the ci'stocarp, in
which are formed carpospores; the nucleus of each of these
contains 14 cliromosomes . The nuclei of the tetraspoTic
plant contain each 14 chromosomes; and it therefore seems
reasonable to assume that the tetrasporic plaints arise from
carpospores. In the first division of the nucleus of the
tetraspore mother cell, the number of cliromosomes is re-
duced one -half, so that 7 chromosomes enter the nucleus of
each tetraspore. It seems probable that on germinating,
the tetraspore gives rise to an individual, whose general
morphological relations and vegatative structure are sim-
ilar to those of the plant producing tetrasjores, with t\/o
significant exceptions, (1) the nuclei show at mitosis 7
chromosomes instead of 14, and (2) the individual bears
sexual organs instead of asexual spores. In other words,
in r^rif fithsia a sexual plant is probably succeeded by an

93

asexual plant of similar morphological relations.

The proof of this hypothesis must rest on actual cul-
tiu'al experiments, and it is much to he desired that such
experiments he ca,rried out.

Since STEASBlIRCrKR ('94) showed that in the Archegoni-
ates the douole niuiiber of claro:nosomes is characteristic of
the sporophyte and tlie single nujaber of the gametophyte,
the main facts have oeen confirmed in so many forms tha.t
many "botanists ha.ve come to consider that the cliromosome
number alone is a trustworthy guide for the identification
of the two generations: that plants showing the diploid
condition of the nucleus necessarily belong to the sporo-
phi'-te, and that where tlie haploid condition of the nucleus
obtains, the gametophyte is necessai-ilj'' involved. Bota-
nists have come to speak of the sporophyte as the "2x-gen-
oration" ^LOTSY, *04) , and of the gametophyte as the "x-
generation"; and undoubtedly within tlie Archegoniate series
such a conception is very useful. However, even in Arclie-
goniates, where the rule is so generally applicable, recent
v/ork tends to shovT that the diploid condition of the nu-
cleus is not necessary for the differentiation of the sporo-
phyte ( YAI-IAirOTJCHI , *07), nor is the I'laploid condition nec-
essary for the differentiation of the gajnetophyte (JAKuITIR

9f

and DIG3Y, fO?) .

In thallophytes, the evidence at hand indicates great
diversity in the point at which the numerical reduction of
cliroino somes takes place. Even in the single group of Rho-
dophyceae, the point of reduction occurs at different
places in the life history of different species. When one
comes, therefore, to regard the chromosome n\jjnl3er as the
sole test for the delimitation of sporophyte and gameto-
phyte, it seems prooaole that confusion vdll result. V'ith
this in mind, I shall now reviev/ briefly the opinions ex-
pressed by workers in tJiis field as to the alternation of
generations in the red algae; and shall venture to offer
some suggestions as to the meaning of the rather compli-
cated nuclear life histories of members of this group,

OLTrJ^TlTS. {*9Q) after a careful study of the develop-
ment of the cystocarp in four genera, came to the conclu-
sion that the sporogenous cells constitute a generation
similar to the sporophyte of the Archegoniate series. La-
ter, ('05, '07) he elaborated this conception and expressed
the opinion that the sporophyte is of antithetic origin,
i.e., that it became gradually intercalated in the life his-
tory by b. series of stages of increasing cojnplexity. In
those forms in v/hich the tetraspores are borne on distinct

^5

t

plants, he regai'ds tjie tetraspore-producing plant as a
"fac\iltative gaj/ietophyte" , which is a resx'lt of a process
of diff erentia.tion similar to that v/hich produced dioecism
in many of the Archegoniates . The tetraspores he consid-
ers analogous to the gemmae of certain liverv/orts . Admit-
ting the possibility that a nxuaerical reduction of the
chromosomes may take place in the tetrasporangivim, he ex-
presses the opinion that the numtier of chromosomes is not
the final test of alternation of generations, "Ich ver-
mvite, die vergleichende Untersuchiing des ganzen Entvvick-
elungsganges ftUirt eher ziun Ziel, oder aher die Kombination
header Llethoden" ('05, p. 273).

The v.'ork of "Wolfe on JHemalion multif idum (»04), in
which he found a nujnerical reduction of chromosomes just
previous to the production of carpospores, furnished a
cytological analogy betv/een tiie cystocarp of the Rhodophy-
ceae and the sporogonium of the Bryophytes, and strength-
ened the position of OLTl'JUTiTS .

YAI'JMTOUCKI ('Oeb), after a very complete cytological
study of Polysiphonia violace^, reached the following con-
clusion: "The sexual plants and the tetrasporic plants pre-
sent the t\.o distinct phases of an antithetic alternation
of generations, with the cystocarp a peirt of the sporophyt-
ic phase" (p. 433). This conclusion is based on the dis-

9^

covery by YAJ^^rOUCHI that the dividing nuclei of the tet-
raspore-producing plant tiiroughout its iiistory, as well as
those of the sporogenous cells of the cystocarp, show 40
chromosomes (t}ie 2x niuaher) , while the nuclei of the sexual
plants show 20 chromosomes (the x nvunber) . The nvunber of
the chromosomes is reduced in the divisions of the nucleus
of the tetr^spore mother cell; the double niuaber is Restor-
ed by the union of the nuclei of the gamates. In discuss-
ing the origin of the tetraspore, YALJ'jTOUCPII surmises that
in some such form as Batracnospermma, in which monospores
are borne along v.itli gamates on the sexual plants, reduction
!nay ha,ve 'been suppressed in the formation of the carpos-
pore, "so that it germinates with the sporophytic nujober of
chiTomo somes, producing a plant v;^hich consequently becomes
at once a part of the sporophytic phase. It is quite pos-
sible that the first tetraspore mother cell corresponded to
monospores on the sexual plant except that they had tlie
double niuaber of chromosomes, since such reproductive cells
v<ould very naturally become the seat of the delayed reduc-
tion phenomena,. The resejiiblance in general morphology of
the tetrasporic plants in the red algae to the sexual
plants would be expected, because they live under similar
environmental conditions" (b.435) ,

9/

The viev/s of 0LTI.IAII1TS and YAJIAITOUCHI coincide so fc>,r
as regarding the sporoeenous cells ol' the cystocarp as "be-
longing to the sporophytic phase of an antithetic alterna-
tion of generations. Tlie point of departure lies in the
interpretation of the tetraspore-producing plant. Because
of its general morphological identity v/ith the gametophyte,
OLTTIAIIMS regards this as a part of the gai/ietophyte , vhich
has become differentiated for the production of tetraspores.
Because of the diploid condition of its nuclei, Yamanouchi
regards the tetrasporic plant as a part of the sporophyte,
whose resemblance to the gametophyte is stamped on it by
"similar environmenta,l conditions".

For the purposes of the present discussion, I sliall
assume from the c^-^tological evidence what it will take cul-
tural experiments to prove, nai.iely, tlia-t in those red al-
gae in which tetraspores and gametes are regularly formed
on separate individuals there is an actual succession of
sexual and tetrasporic plants, the reproductive bodies of
one kind of plant always producing tlie other kind of plant.

YAIIAJ'OUCKI ' S suggestion ('06b) that the tetrasporic
plant may iiave arisen jihylogenetically b:' tiie postponement
of the phenomena of reduction from the formation of carpo-
spores to tlie production of asexual spores seems to be re-

S6

rendered probable "by what we know of other groups of plants.
In the simplest plants which liave been investii^ated from
Viiis standpoint, the position of reduction in the life
history seems to he at the first divisions of the fusion nu-
cleus, as is described in Coleochaete (ALLEN, '05), certain
desmids (KLEBATCI, '91), SpiroF.LT^-af CrOCEELRV/SKT^ '00), and in
I/T;ocomycetes (laiAlIZLIir, '07) . Beginning with the simple
Bryopliytes, the familiar Archegoniate series shows a pro-
gressive removal of the point of reduction from the point
of fusion of the sexual nuclei. These and other examples
seem to show that there is a general tendency thi'oughout
the planj; kingdom to prolong the diploid condition of the
nucleus through the greater part of the life history.

ITowhere is this more plainly shovm tlian in the Uredi-
neae (ELACKIIAII, '04, BLACKIIAIT and ERASER, 'Of,, CKRISTilAM,
•07) . This group is characterized by a succession of
phases, or generations, whicji httve been shown by Christman
(•06, '07) to be morphologically equivalent, thoug}i each
ends in a distinct form of spore. Now nuclear association,
which has been reg&xded as tlie equivalent of fertilization
in this group, occurs, in all foi^ms in which tlie aecidial
stage is present, in those cells of tlie mycelium which give
I'ise to tlie aecidium. The process of nwnerical reduction

99

of the chromosoEies, or, to speak ijiore accurately , of the
cliroinatin, occurs alv/ays in tlie last spore -form preceding
the I'roduction of ci.ecidia, in tlie teleutosjore -when pres-
ent. The diploid condition, extending from the aecidium
to the teleutospore, is lengtiiened hi' the intercalation of
new phases, v/hich, in some cases, seem to have tlie pov/er of
contini^ing the diploid genera.tion indefinitely. The haploid
generation, from the teleutospore to the hinucleate cells
at the "base of the a.ecidiiijn, is never lengthened hy the in-
tercalation of new phases. In other v/ords, in the Ure-
dineae the diploid generation hits 'become prolonged tlirough-
out tlie greater portion of the life history.

In the red a,lgae, it seems likely that a simila.r post-
ponement of reduction has taJcen place. In the llemalionales^
v/hicli is considered the most primitive group of the Rhodo-
phyceae, the point of reduction is retaoved from the point
of nuclear fusion only "by the few cell-generatioas in the
cystocajfp (Y/OLFE on TJemaliori, '04). In the higher forms,
such as the Rhodomelaceae ' YAI'.IAITOUCHI on Folysiplionia,
•06), and the Ceramiaceae (Griff itlisia) , nuclear reduction
is sepsj'ated from nuclear fusion, not only by tlie cell-
generations of the cystocarp, hut e-1so by all the divisions
of the vegetative cells of the tetrasporic plant. Tliat is,

700

the dii'loid j-hase has coaie to occupy tlie greater i.ortion of
the life history.

The tiological meaning of tlas appareiitlj' ^^eneral tend-
ency in the evoli;tion of pls,nt structures is hinted at "by
the experiments of GPIRASSIl'.O^, ('01). Aftei- studying the
growth of vegetative cells of Spii''0":rra in y/hich nuclei
had Ijeen induced to fuse, GERAoF;IMOW,cJ'.' -e to tlie conclusion
t}iat the grov/th of a cell which has an u.nusual amount of
nuclear material is more vigorous than that of a cell with
the USU9.1 nuclear content. The cell wall, the chrojiiato-
phores, and apparently the protopla.sin grow more vigorously.
Such cells divide only after they have reached a size nc-
ticeahly larger than norme,l. (see Bot. Ga^, . 3_5, 224-5,
»03).

If, then, the presence of nuclei v/itli the douole
chrome tin content impcirts greater vigor to the cell, we
should e:-;pect to find some evidence of this greater vigor
not only in the size, hut in the rate of growtli of the tet-
rasporic plants of Griff ithsia. A comparison of sexual
plants with tetraspoi-ic plants does not reveal any constant
difference in the size of the resting nuclei or in the size
of the cells of the two ki^ds of individuals. However, a
comparison of those cells of tlie diploid individual which

101

produce sori of tetraspores with tlie occasional cells of
the haploid plant which form similar structures shows a
striking difference in size, tlie cells with tlie diploid
nuclei "being much larger . .

More iifiportant froDi this standpoint is the fact t}:iat
not only in Griff ithsia,, but in red algae generally v;here
tetrasporic and sexual plants occur side "by side, the tet-
rasporic plajits are, as a, rule, more alaundant (p. 4" ) . In
Griff ithsia Borne ti ana the nuinber of tetraspores produced
is certainly much greater than the niimlDer of carpospores,
and v/e should e:cpect, tlierefore, if the tv/o kinds of spores
were equally vigorous, that the nuinter of sexua-l plants
v/ould grea-tly e^xeed tlie numter of tetrasporic plants;
whereas the reverse is the case. It seems possible that
the carpospores have a greater capacity for development
than the tetraspores. Cultural experiments along this line
are mucli to be desired.

If the view is correct that a postponeuient of reduc-
tion has occurred in some Rhodophyceae , it is evident tiiat,
besides the altei'nation of the gcunetophyte (the sexual
plant) with tlie antithetic sporoph^'te (tlie sporogenous
cells of the cystocarp) , there is a succession of homolo-
gous phases, inasmuch as a tetrasporic individual regularly

]0Z.

succeeds a sexual individual of identical morrhology. This
latter condition is not paralleled in tlie Archegoniate se-
ries; and since the terms gametophyte and sporophyte hiave
come to have a special significance in connection with such
conditions as are found in the Archegoniates , neither of
these terms should he applied to the tetra.sporic plants of
Griff ithsia and Polysiphonia. The tetra-sporic plant iias
prohahly been intercalated in the life history of the red
algae, hut there is no evidence for t}ie helief that it lias
been intercalated hy gradual integration and differentia-
tion of a. simple product of the germination of the zygote,
which product was at first unlike the sexua,l plant and
which represents a new departure in the life history; and
the interco.lation of an unlike phase seems to he the very
pith of the theory of antithetic alternation (see B0Y7ER-
89-91) .

According to this view, the tetrasporic plant probably
arose, when first produced, v/ith tlie complete differentia-
tion characteristic of tlie species. Tlie best evidence for
this conclusion is based on tlie morphological identity of
the tetra.sporic with the sexual plant. Similar environ-
mental conditions would hardly suffice to produce identity
of form in two individuals unless the individuals were

joz

from t}ae "beginning identica,!. The tetrasporic pla-nt of a
red alga may "be said, tlien, to be homolocous v/ith the sex-
ual plant.

That the tv/o phases are homologous is evidenced, not
only by their similarity of structtire, but b;^ the fa,ct that
either seems capable of producing the morphological equiv-
alent of the reproductive structiires of the other. It has
been known since BOroiRT first called attention to the fact
( ) that in many species of red algae structures re-
sembling tetraspores are occasionally found on the sexual
individuals. This phenomenon has been carefully investi-
gated in Polrsiphonia violacea and Griff ithsia 3ornet_iana.
In Polysiphonia, YA:IA:T0TJCHI ('06b) found that the develop-
ment of these tetraspore-like structures ceases at the moth-
er cell stage; cleavage of the cytoplasm ma;' begin, but
noi-Tiial nuclesir division is absent. In Griff ithsia, the
phenomena observed h< ve been similaj? to those noted in
Polysiphonia^, e:;cept that by the time tlie abortive cleavage
begins, the nacleus has divided into two or three. The
cleavage planes have never been observed to reach the cen-
tre of the cell, and it is quite evident that tetraspores
are not formed, since the v/liole cell becomes withered and
wrinkled, resembling the degenero ted tetrasporangia de-

10^

scribed on page G5 . In one instance, however, a xiiother

y
cell was observed in ..hich no trace of cleavage of tlie c.t-

toplasm was apparent, and in vJiich the niunber of nuclei had

increased to^the whole structure resembling very much the

early stages of germination of a normal spore. It seems

quite possible that the tetraspore mother cells borne on

the sexual plants sometimes germinate as monospores, though

tliis can be ascertained only by cultivation.

On the otlier hand, tetr^cporic plants may at times
produce structures morphologically identical with procarps.
Text-figure / is taken from a tetrasporic plant of Sper-
mothamnion Turneri collected at Wood's Hole in August,
1907, and shows tetr£tspores and procarp on opposite branch-
es of the same fil ament. Text-figure ^ shows a section
of the tetrasporangium in which the definitive tetraspores
are formed, though not yet separa^ted. Antheridia liave not
been reported as occurring on Spermo tharoni on Tiu'neri at
Wood's Hole, and cystocarps are very rarely produced.
Whether functional gametes are ever produced on an indi-
vidual v;hich bears normal tetraspores is not known.

Another species of Sp ei^mo t hamn ion. S. roseolum, is
extremely interesting from our point 5?f viev/. pRINGSIffilM
states that in this species, which he collected on the

105

coast of Helgoland, "Kapselfrflchte, Vierlinf:sfrQchte ,i'nd
Anther idi en normal ztisazmnen auf denselben Exemplaren auf-
treten" ('61, p. 26). V/hatever may te the conditions in
Poly^iphonia and Griff ithsia, in which tetraspore-like
structures occuj? only occasionally on the sexiial plants, it
seems highly unlikely that such a careful v/orker as PRIITGJ-
I-Q-IIM should have mstsJcen for tetraspores structiires vrliich
really never "became separated as tetra,spores . Assuming,
then, that the structures descri"bed "by PRIIJGSHEIM were
really normal tetraspores, it becomes at once evident that
differentiation into homologous successive sexual a.nd asex-
ual p}iases has not "been "brought about in all Florideae
above the ITemalionales, but only in those in which the tet-
raspores and sexus-1 organs are borne on distinct plants.
Cytological investigation of such a form as Spermothamnion
roseolum is much to be desired.

The theory o^ homologous alternation in the red algae
outlined above is almost identical with the viev; of PRIITGE-
KRIM as to the relations within this group (»76). PRIITGSR
HEIM states (p. 897) "die Annahme ist nachstliegende dass
bel Plorideen und Dictyoteen zwischen Exemplaren mit Kap-
selfruchten und Exemplaren init Vierlingsfruchten eine Ab-
wechselung besteht". PRU'IGSIIEIM'S view was based, however.

JOG

on a very different kind of evidence froi.i that broueht for-
v.ard in the present paper.

The evidence at liand seems to justify the following
conclusions:

1. There is in Griffithsia a.n antithetic alternation
of generations, the gajnetophyte being represented Id;'- the
sexual plants, the sporophyte by the sporogenous cells of
the cystocarp,

2, In addition to this, there is a regular succession
of tetrasporic individuals and sexual individuals. The
tetrasporic individuals resemble the sporophyte in niujiber
of chromosomes; they resemble the iametopiiyte in morpho-
logical differentiation. They are to be considered as a
phase of an homologous alternation of generations, not the
equivalent, wholly or in part, of the sporophyte of
Archegoniates ,

«L

VITA.

Ivey Foreman Lewis was born in Raleigh, North Carolina
August 31, |08f5. He received the degree of A. B. from the
University of North Carolina, in 1902, and that of K. S. in
1903, being Assistant in Biology during 1902-3. He was
Instructor in the University of ITorth Carolina Summer
School in 1904. He undertook graduate work at the Johns
Hopkins University in 1903 and continued in residence until
November, 1905, when he was appointed Acting -Professor of
Biology at Randolph-Macon College, Ashland, Virginia.

In the fall of 1906 he returned to the Jo}ins Hopkins
University, where he was the holder of a University Fellovr-
ship during 1906-7, During 1907-i.i he held the Adam T.
Bruce Fellowship in Biology a.t the same institution. In
June, 1907, he was elected Professor of Biology in Randolph-
Macon College, v/ith one year's leave of a,bsence.

V/liile at the Jolms Hopkins Univei^sity, his jirincipal
subject has been Botany, his first subordinate Zoology, and
}iis second subordinate Phj/siblogy.

Acknowledgment .

I v/ish to express my grateful appreciation of the un-
failing interest and helpful criticism of Professor D. S,
Johnson during the course of this investigation.

^"^

%^-

\

w^^

^\

\

'%

m

"m^

'%i^

■, ;i

im

■y^^mm

I

.'mm

mm

\'i\\]:f

'&

\)\

! VI

M^