^ d ^ f« ?£ ^

Ï1 m

E3 M-

JOURNAL

OF THE

COLLEGE OE SCIENCE,

IMPERIAL UNIVERSITY,

VOL. IV.

^ m :J^ m W ^n
m v^ -ti- ra ^

PUBLISHP]!) BY THE UNIVERSITY.

TOKYO, JAPAN.

1891.

iq x^

CONTENTS.

Page.

On the Fœtal Membranes of Cbelonia. (Coutributious to the Embryology

of Eeptilia II.) by K. Mitsukuri, PA. D., rti(jalcultakiishi, Professor of
Zoology, Imperial University. (With I'lates I~X.) 1

On the Development of Araneina. By Kamakichi Kishinouye, lUyaku^shi,

Science College, Imperial University. (With Plates X-XVI.) 55

Observations on Fresh-water Polyzoa. (Pectinatella gelatinosa, uov. sp.)

by A. Oka. Imperial University, Tokyo. (WitJi Plates XVIl-XX.J 89
On Diplozoon nipponicum, n. sp. By Seitaro Goto, Pigalcushi, Post-
graduate Student in Zoology, Imperial University. (With Plates XXI~

XXIII.J 151

A New Species of Hymenomycetous Fungus Injurious to the Mul-
berry Tree. By Nobujiro Tanaka. (Witli I'lates XXIV-XXVIIJ... 193
Notes on the Irritability of the Stigma. By M. Miyoshi, Piijai.-usJn.

(With Plates XXVIII-XXIX.) 205

Notes on the Development of the Suprarenal Bodies in the Mouse.

By Masamaro Inaba, Pii(jakushi. (Witli Plates XXX-XXXIJ 215

On some Fossil Plants from the Coal-bearing Series of Nagato. By

Matajiro Yokoyama. (With Plates XXXII-XXXIV.j 239

Comparison of Earthquake Measurements made in a Pit and on the

Surface Ground. By S. Sekiya, Professor, and F. Omoei, Riijalciishi,
Imperial University, Japan 249

Laboratory Notes. By C. G. Knott., D. Sc, F. E. S. E. Professor of Phy-
sics 287

Diffraction Phenomena produced by an Aperture on a Curved Sur-
face. By H. Nagaoka. Piii/akiishi 301

Effect of Magnetization on the Permanent Twist of Nickel Wire.

By H. Nagaoka. Wnakushi (With Plate XXXVIII.) 323

On Certain Thermoelectric Effects of Stress in Iron. By C. G. Knott,

D. So., F. E. S. E. Professor of Piiysics, Imperial University and S.
KiMURA, Pd(jakushi 311

On some Cretaceous Fossils from Shikoku. By Matajiro Yokoyama.

(With Plate XL.) 357

PRINTED AT THK SEISHIBUNSHA, TOKYO.

Publishing Committee.

Prof. D. KikuChi, Ri^akulmkuslii. IV!. A., Director of the College (ex oßcio).

Pi'of. J. Sakurai, Rié^kuhakushi. F. C. S.

Prof. K. Milsukuri, Rièakulmkushi, Ph. D.

Prof. C. G. Knott, D. Sc, F. R. S. E.

From July, 1891.

Prof. D. KikuChi, Rièakulmkushi, M. A., Director of the College (ex oßcio).

Prof. E. Divers, IVI. D., F. R. S., etc.

Prof. K. Yamakawa, Ri^akuliakuslii, Ph. B.

I'l-of. K. Mitsukuri, Ri^akuliakuslii, Ph. D.

On the Foetal Membranes
of Chelonia.

(Contributions to the Embryology of Reptilia H-')

by

K. lYIitSukuri, Ph. D., Riéakuhakushi.
Professor of Zoolofry, Imperial University.

With Plates I— X.

Om- k'iio\vl('<luc ()(■ Hie I'd'tnl iiicirilii-;m('> <»f Kc))tili:i is conrrsycdly
slill very impovfoci. It is o'onenilly assimu'd that they rosoiiil)lt'
luoi'c or less closely tliosc of liirds. K;)IIikor is nlt<\oothci' silcMit,
I'.aliniir oivcs very ine:»ure information, on the siihjeet in thrii' respec-
tive treatises on Enil)ryoloo-y, wliilc llert\\iii- in liis Lelirluich treats
Reptilia thronglioiit as |)resentin,ii- the sanic appearan<'es as liinls on
this point, Ueeently Strahl (Xo. ô), llollnian (Xos. i] & 7). Kavn
(Xo. 0) ami IV'ivnvi (Xo. K!) have toin-lic«! on the siihjeet hut their
ohservations are eoniined mostly to the earlier stages.

Whilst eollectiim- eml)ryos of Chelonia, I beenme aware of the
fai't that there are sonu- very notal)le ieatures presented hy the fœtal
)uem1)ranes of these animals which, so far as T am aware, liave
hitlua-to heeii entii'ely o\ crlookcil. These features a) )peared to me so
remai-kahle and interesting tiiar I thought it wortli while to inves-

* I shall consider th«^ article on " the Ponnation of tlie Germinal Layers in Chelonia ''
by Mr. Ishilcawa atul uiyself, and iiuhlishe<l in tliis .Tournai Vo|. 1 and also in Quart. .Tnnr. of
Micro. Sei. Vol. 27 as the first of this series of C(.n(rilintion.« to llio Euil.ryoh)oy of Ifeptilia.

Z K. MITSTTKUHr.

tio-ate the whole history of these nicmbraiu's in thi« oroup. The
following embodies the resiihs of my sf udy km\ tliis subject.

The species which I have iiivestigatxMl are C le mm y s (or
Emys) Japonica, Gray, and Tri onyx Ja))onicus , Schlegel.
In earlier stages, the fetal meml)ranes of these two species are very
much alike ]3ut in later stages they present differences which, in my
oj)inion, are highly significant. F«H' convenience of trentment. I shnll
divide the present article into three parts, ns follows : —
I. Earlier Stages of the Amnitin,
11. Origin of ihe Allantois.
ni. Later Stages of the Fœtal Membranes.

And in (»ach part, 1 shall tivat the two spe(äes sej)aratel y, generally
giving th<' description of CI em my s first, as that species seems to
have more primitive relations in its fœtal membranes. At the conclu-
sion, 1 have put together some suggestions on the theoretical bearings
(»(' the Cacf^ l)foii<jli( fordi inidef tlie liead of (General Considerations.

I. Earlier Stages of the Amnion.

a. Clciiwni^ Japonica.

The first stage of which 1 shall give a description is re))resented
in siirfi<e view in Figs. ]. and 1 a. Fl. I. There is at tins period a
deep horse-shoe shaped groove bounding the anterior end of the em-
bryonic region — the "Vordere (Jrenzfurche " of Cerman authors*
The posterior wall of this groove is the head of the eml)ryo, while its
anterior wall is the first rudiment of the anterior fold of the
amnion. T'he structure and relations of these parts will become
clear from the sections to be described directly. The medullary
groove is still open thoughout its length, its posterior pai't being
wi<ler apart than its anterior poi'tion. The dorsal opening of the

ON THE FOETAL MEMBKANES OF CHELOXIA. 3

l)l:iwr(jpor(' is \ery distinct. At tiie posterior end of the eniljryonio
region, there is, in the specimen figured, îi h)vv semilunnr fold bound-
ing the embryo from behind. It gives one the impression of its
being the posterior fold of the amnion. Such folds are not, however,
found by any means in all the embryos, and even when present, are
not always of the same figure and distinctness as in the figure. As
the subsequent history shows, these inconstant folds ut the posterior
end of tlie eml)ryo take no share wliatever in the formation of the
permanent amniotic sac.

Hound the head-end of tlie embryo, there is an irregularly semi-
circular transj)arent area of the blastoderm. In this area, there is
usually an opaque line also semicircular and concentric with the
cephalic grooxe (Fig. 1). Kound the postei-ior end of the embryo,
and along its sides, there is a brcjad h(jrse-sh<^e shaped opac(ue streak
which is caused by the abundant accumulation «jf yolk-granules —
the germinal wall. The mesoblast, at this stage, extends into the
head of the embry«) [)r<jper, but anteriorly, laterally, and postei'iorly
the opaque liorse-slioe shaped streak mai'ks the limit of its extent.
Hence the transparent area in front of the embryo is as yet free from
the mesoblast .

In Fig. n.S, IM. \ll. (See also Diag. 1. IM. X.), a median longi-
tudinal section of the head end of the embryo is represented. It is
evident from this section that the deep horse-shoe shaped groove at
the anterior erid (a. /../.) is bounded posteriorly by the head (H. F.) of
the embryo, while its anterior wall forms the first rudiment of the
anterior fold of the :tmnion (Anin). The anini(jn is thus laid in the
region into wliidi the mesoblast has not yet found its way, and
ihcTefore. of nercssitN , consists at lirst only of the <'|)iblas( and hypo-
blast. In Fig. 59 (Fl. VII.), a transverse section of the same region
is represented. From this and Fig. 58 (PI. VFT.). the characters

4 K. MIT«UKUKI.

of the two layers in the niniiioii will Ik- easily understood. Where
the layer« reaeh the le\el of the _ueiieral surface of the ])lastoderin,
the epiblast presents a thickened ridu'e aloni;- the whole u|>|)er ed^e of
the ii'i'oove. As it i.s hy the _ü'rowth backward of this cdu'e that the
amnion comes to co\'er the endjryo, tliis ridge of the epihlast must
he the seat of an active i^row th. There is also in the mcthan line a
thickened rid_ue (Fig. 51:), (•.)oftlie epildast which starling from the
hottcjm of the groo\e reaches as far a.s tlic lexel of the hlasloderm,
litling in its ujjward course the still oj»en medullarv !jroo\ e of the
head of the endjryo.

In the semicircular tI•an^])arent area in front (jf the anterior
liorse-shoe shaped groo\e, the epihlast consists of two layers of pave-
ment cells, of which the up|)er is especially Hat and seems to be of
stitf consistency. The hypoblast in the region directly in front of
the groove consists of polygonal cells. The o[)aque semicircular line
in the transparent area already spoken of seems to be due to a special
acGimiulation of the hypoblastic (Fig. 58.) cells. A little in front
of this line, the hyjjoblast becomes suddenly a mass of live yolk
granules with nuclei scattered among them. At the periphery of the
transpai'cnt area this passes rather abruptly into a 1)ed of large y«>lk
spherules.

The fact that the amnion in Kept ilia, consists, when Hrst laid,
only of the two primai-y layers was made known by IStrahl (No. 5),
Hoffman (No. (i), Ferenyi (No. 16), and Kavn (No. 9). Hoffman
pointed it out as a point of great difference between the amnion <jf
l\e{)tilia and that of lîirds, but it is now well kno\\n lliai, in Jîirds
also, the amnion consists at first only of the epihlast and hypoblast.
Kijlliker refers to the fact in his classical work (zweite Auf. j).l.SS. and
Fig. 85). and Ka\n (No. 8) has woi-ked out the point ehdiorately. \ an
Beneden and I'h. dulin (No. 11) also ob>erved the same fact in the

ON THE i'uETAL MEMBIIANES OE OHELONIA. •'>

|{,;il)1)ir ;iti(l lîats und iiaiacl the two-layered amniotic cap the •' Tro-
amiii(jij."' Fleischmann ha.s also found the s«ame state of things in
the eat. It seems tlierefoi'e an estabhshed faet that the head-fold of
the amnion, when iirst laid, consists throughout the Amniota only
of I he epiblast and liypoblast and is therefore of the natiuv of Pi-oam-
nion. In liei)tilia, this point is made perha])s more conspicuous hy
the sid)se<pienl histoi'y of the fœtal membranes than in other groups.

It will be seen fr«jm this stage tliat the head ot" the embryo sinks
fr(jm the Iirst I)eIow the leNel of the blastoderm. Apart from any
])hylogenetic significance, there is mechanical necessity for its sink-
ing in this manner. As soon as the development begins, the white
of the egg is rapidly absorbed from the part over the blastoderm
which becomes adherent to the innei- surface of the shell membrane.
There is therefore no space into whicli the head can grow except
towards below. In removing endjryos from the eggs, I availed
myself of the llict of tlie ijlastoderm ])ec«;ming adherent to the shell
membrajie. for. with a stout pair of scissors, I could easily cut a
watch-glass shaped piece of the shell with theshell mendjrane and the
embryo adherent io it, and inverting it, I coidd jxjur the preservative
iliiid into it, thus using it like a veritable watch glass, (jnly excelling
it ill this that it keeps the end)ryo and the Ijlastoderm stretched in
their natiu'al positions.

I am inclined to thijik that the sendlunar ridge at the posterior
end (I'ig. 1) i'*^ also caused l)y the posterior heavy end of the embryo
sinkini!' int<j the s})ace below . Its section is almost exactly like that
of the lateral foM of the [»ermanent amnion (Figs. oO and oOa, PI. V.).
Such adventitious ridges seem to be produced here and there witlnnit
any regularity (••f. also Fig. 2). They are of a transient nature and
take no part in tlie foi'ination of the anmion.

As later staijes will show, the wliole amniotic sac is

6 K. MIT.SUKUKI.

}) V ( ) duce d solely 1j y t li e g r u w t li h a e k \v a r d o f t h e an t e r i o r
fold in conj unction with the lateral folds which ri«e grad-
uall\ from before backward.

In the «taiie with two or three ]ne>s(jbla.stic somites, a> shoxNU
in Figs. 1> and 1^«, IM. 1. (see also Diag. II. and 11'. Tl. X.), the am-
niotic fold has extended nearly lialf o\er the body of the embryo
Avhose anterior pari has sunk meanwliüe more and more below ilie
level of the blastoderm. The jjosterior ii'dge of the amniotic hood
presents a horse-shoe shajjed outline, being caused by the lateral f )ld
of each side extending more posteriorly Ihan the median p;irl. fhere
are ;igain some irregidar folds (a) in the posterior parts (jt I he
embryonic region.

Figs. o0-o3 (1*1. \'.) show a series of transverse sections selected
from différent parts of this embryo, Fig. oO being the most j>ostei'ior
and Fi"". o3 the most anterior.

F^iiJ". 30 is from the region co\ered only bv the lateral lind),s of
the horse-shoe shaped posterior margin ol the amniotic hoixl. F'rom
this and Fig. oO« (PI. V.") (the latter representing the left half of
Fig. oO under a much higher powei" of magnilication) it will be seen
tliat the lateral f »Id of tlie amnion, when first laid, presents two
peculiarities: (1) it is purely epiblastic, ;uid the mesoblast has no
shave whatever in it ; (2) the fold is solid and not composed of I he
inner and outer limbs as re])resented in ordinary diagrams.

F^ig. ol is just in front of the point where the two lateral am-
ni<jtic folds have united. One half of it i> shown under a hierher
power in Fig. öl</ (PI. \ .) j'he whole amnion here is composed of
a solid sheet of the e[)il)l:i.s|, I he mc>oi)I;ist insinuating it sell' between the
epiblast cells only later on. The cells of this part of the amnioji are in
se\ eral layers, and of these, the cells of the outermost layer have under-
gone some process of hardening and their nuclei are stained dee] 'est.

O.V THE FOE'I'AL MEMBRANES OE I^HELOXIA 7

Fii;-. 'A'2 is irniii the point where the lien<l-eiid «»1' the embryo is
}uM ])Vii\ii\\']])<x ^" ><\nk Inflow tlie level of the Idàstodcriii, Tlie meso-
hlast of the liody hns .^epiij-jited from the extr:i-emhryoTiie \mri . The
amiiioii is mostly epibl.istic, ;xlthoiio-h lined hy the hypolilasl iiir n
short distnnee on en<-li side.

Imù". .').') is froiii the liead reo-ion whidi is completely snnk below
the level of tlie hinstoderm. As em|)hnsized bv Motfmnnn, the bead
aj)pears in the cr«»ss section Ix-low. instead of above, the blastoderm.
The aiiinioii is eom]>ose(l of the epiblast and hypoblast, eaeli heino-
only one-eell layered.

These sections show that the amnion at this sta^'e eonsists, in
the region ol* the sunken head, of the epil)last and hypol)last, and in
the (h)rsal ren-ion, of the epil)last only. The mesoblast as vet has no
share in it.

In the stajii'e with t! or 7 mesoblastie somites (Figs. )> and .') a PI.
I. See also Diag. TIF. and TIT. PI. X.), the amniotic hood has ex-
tended to the posterior end of the eml)i'yo, leaving (jnly the region
round the neurenteric canal exj)osed. The mesoblast has also very
much increased in its distribution and has become, thoughout, split
into the somatic and splanchnic layers. The cœlom has thus
appeared not only within the body of ihe cnibi-v«» proj)er but lias
extended itself into the cxti'a-embrNonic porti«)ji ol' the blastoderm.
Although the mesoblast has originally s])read IVom behind foi'ward,
the cœh^inic cavity appeal's first in the neck region of the embryo and
s])reads graduall}^ backward — as was pointed out by Strahl (No. 5).
In the stage represented in Pig. o, when seen through from above,
the extra-embryonic coelomic cavities of tAvo sides, extending into
the amniotic folds come close together (but are not fused) in the
median dorsal line along a considerable distance in the anterior part
of the dorsal region, but separate from each other before the posterior

8 K. MITSÜKURT.

eclo"e of tile aiiuiinii is roachod, and, oradn.'illy Icssciiiiif^' in \hv\r height,
arc lost too'ethor with tin- uratliially lowcrini;' lat^-i-al iolds of tlie
amnion. Tlius the niosohlast now has a consi(]('ral)l<' share in the
formation of the amnion.

Fio-s. 34-38 (PK V.) are a series of transverse .sections selected
from different regions of tliis embryo.

Fip-. 34 is from the reo-ion where the lateral folds of tlie amnion
are still Ioav. When we compare this with Fio-. 30. we see that in
this stage the somatic layer of the mesoblast is f<^ldetl and pushing
itself into the hitherto solid epiblastic amniotic folds.

Fig. 35 is from the region where the mesoblastic folds or, wliat
amounts to the same thing, the extra-embryonic cœlomic cavities are
still some distance from the median line. Fig. 35 a represents the
median dnfsal portion •»!' the amnion in the same section under a
liiiidier ])owei-. It is evident that here also, a somatic fold ol' the
mesoblast insinuating itself, so to speak, on each side into iho ori-
ginally solid epiblastic amnion is separating the latter into two liml)s
of which the inner is the true amnion and the outer the false amnion*
or serous envelope.

In Figs. 36 and 36 a, the mesoblastic folds have reached furtlier
dorsahvard, but the amnion aiid the serous envelo])e are unite(l in the
mc^dian line. \n Fig. 3(! /*, a few >ections forwai-d ol" Fig. ."»(!. the
mcsolilastic folds have readied still fiirthcj' dorsal ward— the most
(h)rsalward at this stage — but still there is a distinct connection be-
tween the amni<m and the serous envelo])e.

The mesoblastic folds inaintain themselves at the level given in
Fig. 36 /' for many sections forw^ard, and the comiection between the

UTl'

* In futniv, I shall avoid iisint;- tho t^riii " fais.- amnion "" to d.'iiotu the strurtnit' 1
indicated and shall call it tlip sorons .'nvelope, as tlio tonn -'fnlsi' nniiiinTr' is appli.'d \>> tu
very different structures }jy GJerman and Enolish authors.

ON TUE FOETAL MEMBRANES OF CHELONIA. 9

aiTinion and the serous enveloj)e is also iuvariahly jn-eseiit. (Coinjtare
Fig. 3, PI. [.).

In Figs. 37 .'ind o7 a, which arc froii) the region of the he;irt,
where the head-end is heuiiiniiig ti) sink heh)\v tlic sm-face of the I)histo-
derm, the mesoblastic folds have again receded fi-om each other and the
connection l)eMveen the amnion and the serous enveioi)e is a^ain liroad.

Fig. 38 is from tlie .legion of the head sunk l)e]o\v ihe le\el of the
bhist()der!n, which therefore ;i|)])ears above tlie liead in this s(^ctiorj. Tiie
amnion or proamnion consists only of the epildast and hypoblast.

The relations of different parts will become clearer, when studied
in a longitudinal section.

Fig. 41 (PI. V.) is such a section slightly out of the median dorsal
line so that the extra-embrvonic cœlomic cavity fcœl') of one side
appears in the amnion. This section shows that the epiblastic amniotic
fold reaches nearly to the neiu-enteric canal, while the hypcjblastic fold
extends only to the neck region. The triangidar space between these
two f(jlds as seen in this section is occupied, for the most part, by the
mesoblast enclosing a porticjn of the extra-embryonic cœhjmic cavity.
A little earlier there would have been no mes(jblast in the amnion
which then consisted, in the dorsal region, of the epiblast only, and
in the sunken head part, of the epiblast and hy|)oblast. The meso-
blast is now ])ushing itself into the solid epiblastic sheet of the
amnion, dividing it into an outer and an inner lindo. In Fig. 41, the
posterior ])art of the epiblastic fcjld is, however, still solid. Anteriorly
the mesoblast is insinuating itself between the epiblast and hypoblast.
The cœlomic cavity in the mesoblast is widest in the anterior part.

One of my most important results is in regard to the connec-
tion between the amnion and the serous envelope, seen in Figs.
35-37. Contrai-y to what is hithert(3 known, the extra-
em b r y o n i c c ( e 1 o ni i c c a V i t i e s of t w (j sides a r e n e v^ er united

10 K. MITSUKÜRI.

across with each other over the dorsal region of the em-
bryo. A C(mnection — quite elongated and definite in
later stages — between the amnion and the serous en-
velope separates them to the very end of the develop-
ment. That this structure causes great peculiarities in the fœtal
membranes is to be expected and will become clear as later stages
are described. This connection, I shall call hereafter the sero-
amniot i c connect] on . It does not extend to the sunken head part
where the amnion cojisi.st.s of the ejnblast and hyp(jblast, and is con-
fined to the region l)ehind the neck representing the original solid
epiblastic sheet of the amnion oi- its prolongation behind.

While Fig. 3 (PI. I.) Tio doubt represents the conunonest and
noi-nial form in which the amnion s])rea(]s backward, it seems by no
means to l)e the exclusive «jne. Fig. 14 (IM. II.) shows one in which
the posterior f)ld is present but a part of the left lateral fold is
absent, so that the hcjrse-shoe shaped postei'i<jr margin of the amnion
is open toward the left. I have also another embrjo in which a
part of the right lateral fold is absent.

Now comes the most i'emarkal)le point in the development of the
amnion in C I em my s . According to what is hitherto known al)out
the amnion, one would expect that when it has reached the stage
shown in Fig. 3 (PI. 1.) the |)osterior fold will be produced or the
lateral folds will convei-ge toward each other and thus the amniotic sac
wnll be completely closed. Such is not the case in Clemmys. The
anteri(n- and lateral Iblds which stalling from the head have gradual!}^
extended backward over the whole embryo do not stop at the
posterior end of the embryo but continue to grow back-
ward, although diminished in their width, until final-
ly there is produced a tube extending backward from
the posterior end of the embryo, almost as long as the

ON THE FOETAL MEMBRANES OF CHELONIA.

11

body o f t h e embryo itself, connecting' the amniotic sac
with the exterior. A reference to Fig".s. 4-7 (PI. I.), will make
the growth of this posterior tube clear. In Fig. 4, the folds have
extended slightly beyond the posterior end of the embryo. Beyc^nd
this point, they snddenly come near each other, and being diminished
very much in width, their continued growth backward produces a
tube (Figs. 5 and 6). It will be seen that the extreme posterior point
always presents a horse-shoe shaped outline, as it did when growing
over the body of the embryo itself. Fig. 7 shows the stage of the
greatest development of this tube in my possession. Tn three
embryos of this stage whose lengths are 8, 8, and 8 \ millimeters, the
length of the posterior tube of the amnion is respectively 6, 8, and 7 ^
millimeters. The posterior opening is s(^me distance beyond the
edge of the vascular area.

The sections of this tube show that the relations of the
different layers are in all essential respects exactly as in that part of
the amnion )»roper enclosing the einbryo as shown in Figs. 39 and 40
(PI. v., from the embryo given in Fig. 5) of which Fig. 39 is froni the
anterior part of the tube near the embryo and Fig. 40 from about the
middle of the tube. In the surface view, there is often seen a streak
along the median line of the tube, which is shown by the sections to
be a thickenino- on the floor of the tube. The structure of the similar
tube in Tri onyx is gi\en in a more enlarged scale in Figs. ô3-r)o
(PI. VI.).

What the function of this remarkable tube connecting the am-
niotic sac with the exterior is, -^whether it has any active function at
all or is only of the nature of a remnant (-»rgan, I am unable to tell.
I think it probable that it serves for conducting into the amniotic
sac the nutritive matter from the Avhite, with whose gradual
disappearance from over the embryo the backward grow^th of the .

12 K. MTTSUKURI. -j

posterior amniotic tube seems to keep pace.

The condition of the amniotic sac proper at this stage when the
posterior tube has ah'early been developed is shown in the series given
in Figs. 42-47 (1^1. VI.) from an emln-yo with twenty mesoblastic
somites. An inspection of these figures shows rliat. over the ])osterior
part of the embryo (Figs. 42 and 43), the amnion and the serous
enveh)pe are still adhei-ent to each other for a considerable space :
hence the extra-embryonic rœlomic cavities (cœl') of the two sides are
separated from each other by a wide interval over the dorsal region.
As we proceed forward, the mesoblastic folds gradually push toward
the median line se])arating the amnion from the serous envelope, until,
over the middle region of the embryo, they are separated only by a
thin partition (Figs. 44 and 44a). This partition — the sero-amniotic
connection^ — -has now become vertically somewhat elongated and unlike
Figs. ?)(vi and b (PI. V.) ])resents a string of cells in a cross-section
(Fig. 44a). This represents the greatest vertical elongation of the sero-
amniotic connection at this statte. Further forward, the mesol)]astic
folds becoîne again separated by a considerable interval (Fig. 45). An-
teriorly to the point wluu-e the head-end l)egins to sink beneath the
surface of the l)lastederni, the cœlomic cavities of two sides which
arose separately have l)econie united across, there being no sero-
amniotic connec'tion from the beginning in this part (P'ig. 46). In the
head which is freely pi-ojecting into the cavity below the Ijlastoderm,
the amnion still consists only of the e])iblast and hypoblast (Fig. 47).

Fi'oin what has been given above, it iollows that the extra-
embryonic cœlomic cavities of two sides are separated from each
other over the dorsal median line by the seio-amniotic connection
from t\\c n«M'k region to the very tip of the posterior
tube. In front of the neck region, i.e., in the siuiken head regicm,
the cavities become early united across. It is important to

ON THE POETAL MEMBRANES OP GHELONIA. 1«^

reineinb(>r this fact in order t<i understand the relations of some
parts in later stages.

In a slightly older embryo, the sero-amniotic connection has
increased more in its vertical extension. Figs. 48 and 48a (PI. VI.)
are from the tail region, Figs. 49 and 49^/ (PI. VT.) from the middle
of the body. Tn the latter, the sero-amniotic connection is of a con-
siderable length, becoming quite definite.

As to the fate of the posterior amniotic tube. At the stage
(Fig. 7, PI. I.) when it is in its highest development, the axis of the
tube is the same as that of the embryo, i.e., the embryo and the tube
are in the same straight line. Beyond this stage, the tube begins to
become curved, at first slightly, then more and more. In Fig. ISa
(PI. IT.) the curvature is very slight ; in Figs. 13/> and 8 (PI. II.) it
has increased greatly ; in Fig. 9 the distal portion of the tube is bent
at a right angle to the proximal basal part ; in Figs. 10 and 15 (PI.
[[.), the tube has become very irregidarly curved. It will be seen
that the tail end of the embryo which is at first far in front of the
horse-shoe shaped distal end of the posterior amniotic tube (Fig. 7)
gradually approaches the level of the latter (Fig. 9) until in Figs. 10
and 15 it has pushed itself far behind. It is now the distal end of the
tube that is in front. This change of the relative positions is no
doubt due to the flict that the embryo and the amniotic sac proper
grows more rapidly than the posterior amniotic tid)e which they push
aside, so to speak, in oi'der to gi-ow beyond it. As the eurvature be-
comes greater, ])arts of the tube become fainter and fainter in
appearance. For instance, in Fig. 10, a large |)art of the tube
excepting the dist;d horse-shaped end and the proximal basal jmrt,
was very difficult to recognize (being represented too distinctly in
the Figure). In Fig. 15 I could detect only faint traces of the tube,
hei-e and there excepting the proximal l)a,sal part which is always

14

K. MITSUKURI.

distinct. The oldest stage in whif^h I detected any portion of the
distal half of the posterior amniotic tube i>; that ^iven in Fig. 67 (PI.
VITL). I found there the horse-shoe shaped distal eîid of the tube and
the portion contignoiis to it, hnt after n most cnrefnl sf^arch. I could
not connect it with fhe proximal part. From tliese facts, it appears
that the largest part of the posterior amniotic tidie disappears entirely,
and that oidy the proximal |»art — the ])art nearest the amnion proper
(prox pt. Figs. 9, 10, and 15, PI. II.) — remains permanently. It will
he remembered that the scro-amniotic connection extends from over
the neck regi(^n of the embryo to the tip of the |)osterior tube. As
the proximal part of the tidoe remains permanently this marks in all
later stages the posterior end of the sero-amniotic connection. As fur-
ther growth in size of the amnion proper (accommodating itself to the
growth of the embryo within it) takes place mostly behind the rem-
nant of the posterior tube, the latter and the sero-amniotic con-
nection come to lie in the anterior part of the amnion in older
embryos. The growth in size of the amnion after being closed on<-e is
therefore due mostly to the enlargement of that part which is placed
behind the posteriiM- tube enclosing the tail end in a sî;ige like
Fig. 11 (PI. IL).

In all the stages hitherto described, the head of the eml)rvo
projected below the level of the blastoderm covered l)v the proamnion
which consists only of the epiblast and hv})oblast (Fig. 41. PI. V.).
On this accoinit, in sections of this region, the head is found Ix'low
the general level of the blastoderm (Figs. ^3, 88, PI. \'., Fig. 47, PI.
VI.). The manner in which this anomalous state of things is
brought to a close, and in which the head covered by the tunnion «-on-
sisting of the epiblast and the somatic mesoblast comes to lie above
the hypoblast as in other |)arts (^f the body, has been described by
Strahl (No. 5) and Hoffman (No. 6) and c^uite recently by Ravn (No. 9).

ON TEE FOETAL ICEMBRANES OF CHELONIA. 15

The last named author ÇS(^. 8) hns also studied the process in the
chick and found it to he alike. My own ol)servations jiuive in all
essential points with the account gi\en hy these authors. The
process briefly stated is as folhjws : As stated before, the extra-
embryonic cœlomic cavities of the two sides become early united across
iîi the he;id re^'ioii, there being no sero-amniotic (.'onnection here.
This united caxitv or its niesol)hist \\;dl. in spreading itself, in-
sinuates itself hetween the epibhist and hyo]toblast of the bhistoderm
and thus |)Ushes the hy|>oljlast forward and downward. A com-
parison of Figs. 41 and 41(/ (PI. \'.) aii<l Diags. III., I\'., \'. (I 1. X.)
will make this jxnnt cI(Mu-. In Fig. 41. the head is still entirely
covered by tlie proamnion; in Fig. 4h/,, the extra-embryoiiic
cœlomic cavity (cad') in enlarging itself, has ])ushed the hypoblast
forward and peeled it otf, so to s])eak, from the greater part of the
proamnion covering the head, so that now tlie proamnion is found only
on the ventral part of the head. Meanwhile, the emliryo turning on
its longitudinal axis comes to lie on its left side. These movements
brinR- about the state of things as shown in F'in's. 11 and 11(/(P1. II.)
In Fig. 11 the end)ryo lies evitirely on its left side, and a small
anterior part of the head is covered by the now much reduced pro-
an.nion. In th(^ ventral \ iew of the same (Fig. IP/) the proamnion
is very conspicuous, because it is transparent and without blood-
vessels. A section from the head of this embryo is shown in Fig.
85 (PI. X.). Lt shows how the proamnion extends now only for a
short extent.

The final disappearance of the proamnion is bn^ught about by
the continued extension of the mesoblast. Although the encroach-
ment of the proamnion takes place to some extent from behind and
before, it takes place most actively from tlie two sides. Pig- 86 (Id.
X.) is a se<'tion sinvilar to Fvj;. 85 Iron! a somewhat older embryo.

16 K. MITSUKURI.

How the proamnion hns been eneronched upon from both sides and
has all but disap'H'ared is \ ery clear, if we compare these two ügures.
These two figures show also that the k^ft \'itelliiie vein (Y va) becomes
much larger than the right.

h. TrioHijx Japonicus.

Earlier stages in the development of the Amnion in Tri onyx
are verv much as in Glemmys. There is in fact n<j point of any
impcn'tance which is different in the two species. As, however, tlie
Trionyx embryos in ni}' possession show very well in surface views
how the extra-embryonic cœlomic cavities arise first in the neck region
and gradually spread backward, I shall introduce here some figures
which illustrate that point among others.

Fig. 16 (PL III.) is the stage closely resembh'ng Fig. 1 of
Clemmys (PI. I.). The anterior horse-shoe sha})ed groove (" die
vordere (Trenzfruche "), the still open medullary canal, and the trans-
parent area in front of the embryo are all very similar to the
Clemmys embryo < )f the corresponding stage.

In Fig. 17 (PI. III.) the amnion has extended over the anterior
half of the embryo. When seen from the ventral side the whole anterior
end of the embiyo covered by the proamnion is projecting below the
level of the blastoderm, as shown in Fig. 17^. In the neck region where
the emljrvo gains tlie level of the blastoderm, one is able to recognize
distinctly the extra-embryonic cœlomic cavity on each side of the
embryo appearing as a vesicle which bulges out the dorsal and ventral
surfaces of the blastoderm (Figs. 17 and 17a). The level of its poster-
ior limit is the same as that of the posterior limit of the amnion, and the
growth backward of the cœlomic cavities progresses hand in hand with
the backward growth of the amnion. fhese t\vo cavities, one on each
side of the embryo, are (jf course the same as Strahl's '' Mesublastische

ON THE FOETAL MEMBliAXES OF CHELOXIA, 17

Schliiiiclie " (Of. Struhl Xo. 5). The section« of this embryo show
that in the reo-jon where the hitenil fjlds of the amnion have not yet
united in (iie median line (Fig". 50 l*L \ I.), the fold is purely
e[)iblastie and sohd. and tlie mesol)hist has no share in it at all.
In the region where (he extra-emhryonie l)udy ea\"ity is jjresent (biii".
•M). tlie mesobhistie fokls have ah'eady pushed themselves eon-
sideraljly into the epiblastie amniotic sheet, dividing it int<» two
limbs: tlie amnion proper and the serous envelope. In the embryo of
Clem my s given in Fig. 2 (PI. I. ), the mesol)iast has as yet no share
whatever in the formation of the amnion. It follows therefore that
the mesublastic folds begin to push themselves into the epiblastie
amniotic sheet somewhat earlier in Trionyx than in Clenimys.
Fig. b'2 is lr(jm the head region of the embryo given in I'ig. 17. The
head is surrounded b^• the proamnion conn)osed for the most part of
the epi blast and hypoblast, and appears beneath the blastoderm
instead of aboN e it.

Figs. 18 and U) (PI. III.) show that the amnion is gradually
spreading backward, and with it the extra-end)ryonic cœlomic cavities
are urowinn- laru'er and lart^er. In Fiu'. 18, the cavities of the two
sides are still wide apart over the doi'sal region of the embryo; in Fig.
11) they aluKjst touch each (jther along the median dorsal line in the
anterior dorsal pari of the endjryo. but are considerably apart ui the
posterior region. A section tivjni the anterior region (Fig. 57) shows
that thev are separated l)y the sero-amniotic connection which aj)j)e;u's,
however, still very short in a section.

In the embryo given in Fig. 23 (PI. 1 \ .) the amnion has
covered the eml)ryo entirely and has even extended a short distance
behind it. The c(elomic cavités are correspondingly enlarged.

In the stage gi\ en in Fig. 24 (PI. \\ .). the ])osterior amniotic
tube has become alread}' (piite elongated. Its ])os(erior opening is

IS

K. .Ml'I'SUKUi;!.

iKtw jtisl ;il flic cdiic (»1 tlic \as(;iilar area. Tli«' I'Xl fa-ciiilirvuiiic
(j(i'loj)iic cavities have now ('xteiided so iniicli thai llicv ar<' no Jouter
recogiiizahle a.s \'esicles in a sin'faee view. I'iu's. oo-of) (I'h ^'I.)
ai'e tlirce sections from (liHci'cnt parts of" llu' ])()stci'inr ainnit)tic
tiil)e (ti this emljrxo. 1' ii^s. ').") lieini!' near the jxisteriol" openini:' and
others ijcino- in trout of it, Huh \]\c c])il)hist and somatic ]iies<;h]ast
ai'e re{)resented in tlicse hüuix's. the i(cIoui. the s])hinciinic niesolihi>t
and yolk licini:' left out, as a comparison willi Fii;'». .")0 iind iO (I'l.
\ .) \\ill show, Fiu'. .")(! is the median ]>ai't ol" the auinion from o\(;r
the middle re,Lii<)n (jf the body of the emljrv«' :njd shows the 2"reîitest
encroaehment at this stiige of the extni-endjryonic co'loniie cavities,
reducini:' the sero-amniotic connection to a meiv sej»tum-]ike ])artition,

rigs. 21 a, h, c, and d. (IM. III.) she)w the [)osferior amniotic
tuhe of foui- embrvos of the staii'e a little older than that i;-i\en in
Tig. '24. hi a the tube is still straight, in h it is slightly c^r^('d,
and in r and il more curved. hi a the embi-yo is (! mm. h»i)g, while
tlie ])osterior amniotic tulie is only o -g mm. As this is no doubt the
stage of the highest development of the tube, it follows that the
])ostei-ior tube in Trionyx is not as long relatively to the l)ody of
the embry« > as i t is in C 1 e m m y s .

In f'igs. 'J'J (I and h (PI, Iir,) tlie posterior amniotic tube is
becomijjg very irregularly cur\ed.

Iji Fig. 27 (PI. I\'.). most of the jiosterior amniotic tube has
already disappeared or at least is unrecognizaljle. The ])roximal or
basal part of it is, however, veiy distinct. At this stage, the sero-
amniolic connection exists from the neck-region to the tij) of the
renmant of the ])osterior amniotic tube.

The maiuier in which tlie proamnion coiisisting oidy of the
epiblast and hy|)oblast is graduallv ri'])laced bv the amnion consisting
ol the e))i1jlast and mesoblast is exactlv as in Cleminys.

ON 'I'lIK Foi: TAL M EM IJüAXKS Ol' CllELONTA. !'•

II. Origin of the Allanlois.

lu'siilcs KitpH'rr who (U'i'ivcs tlu- allantois tVoiii iho lU'inviitoric
onnal. the one wlio has most can-l'iilly studied its origin in Kcjiiilia
is Strahl (Xos. 1 and 2). Acc(n'dinp; in this anth»n'. tlie allantois is
laid, in Fvaccrta, as a solid knol) at the post<'rior end of the emhrvo,
suhsequentlv hollows itself out, and only then comes to eonnniinicate
with the hind-gut hy an independcMitly formed allantoie stalk. Ft
then tiu'ns rciiind the tail end and comes to lie in lV<Mit of. anil lielow,
the latter.

After Strahl, llotl'niann (Xos. (! and 7) and I'ereiiyi (Xo. l(i)
have studied the oi"igin of the alhintois in lve])tiria. The \ lews
whieh Holfmaim expresses in his lirst papei- (Xo. (5) mainly support
Strahl's ohservations. while in his se<'oiid. paper (Xo. 7) lie seems to
have somewhat modified his idea. For the exact details in which his
later ideas difler from those of Perenyi(Xo. 1 ()), T must refer the reader
to the original papers themselves, as I have to confess my inahiliiv
lo grasp them ))recisely. Xotwifhstaniling Perenyi's sfateiiuMit
that thev dilfer, it appears to me that they are deserihing snl)stan(ially
the same process. Under the circumstances, I am unahle to sa\'
wdiether my results agree with the view of either or IxHh of these
aiitliors, although I think we have arrived at^ nearly the same
results, Hoffmann says that the origin of the allantois in Reptilia is
tlu'onghoiit the same as in lîirds (X'o. 7. ji. ISO). Such is the eon-
elusion r too have arrived at, after a eareful study oi' ("helonia.
In fact, ihis is so much so that (Jasser's figures (Xos. \'2 atid l."!) oi-
lîalfoiir's deseriplioii (Comp. Emhrvol. \'ol. If.) on the origin of
the allantois in liii-ds tnight lie hodil\ adopte(l to desia-iiic tiie same
process in ('lielonia.

As I lia\'e a more f'oinplete sei'iev of the 'frionvx emhryos

20

K. MITSUKURT.

illnstratiiiri- lliis ])r)i]it lliaii those, of ('Iciniiiys, I sliall hcoin with
the former species.

Figs. 60-63 (PI. VII.) aiKl Figs. S7 mid 87« (Fl. X.) give
successive stages in the development of the ullantois in Trionyx.

Fig. 60 is fr<:»m an embryo very simihir to the one represented
in Fig. 23 witli about seventeen mesoblastic somites. The sjjhmch-
noplenre has not vet l)een folded under to form the hind-gut. The
first trace of the allantois (All.) is, however, already visible as a
shallow notch in the posterior ]iart of the tail- lobe. In a surface-
view, this notcli appears as a shalh^w transverse slit as re))resented in
Fig. 20. Fi'om the first, tVic {)osterior wall of the allantois is lined
with a distinct epitlielium of the hy|iol)last. Its anterior wall is no
doubt also of the hypoblastic nature, but is her(^ fused Avith \\\v
indifferent cell-mass alx^ve it.

Figs. CA-G?^ (IM. Vll.):uid Figs. 87 and 87« (Fl. X.) s])C'A suf-
ficientlv foi- themselves :ind need tioI be miiiutelv exphn'ncd to those
who ai'e alreadv familial' wiih llic (•<»! I'espoudiiig shigcs iii j>irds.
]jv the gradual f )lding of the spianclmi^pleui'e on the ventral face,
the hind-gut is j)roduced, and on its ventral floor the allantois
becomes esta])lished as a vesicle at first wide open ahme (Figs. 62-63)
but with its gradual growth constricted at the neck (Fig. 87). Fig.
87« represents a cross-.se<'tioii of the alhuitoic region from an embryo
of the same stage as that represented in Fig. 87. It shows that the
cavity of the allantois is at this stage two-lobed.

Figs. 64-66 (1^1. VII.) are three successive stages in the develop-
ment <^f the allantois in Clcuniiys. Althougli these do not give as
complete a series as in Trionyx, they are }'et sufficient to show that
the process in Clemmys is in all essential res])ects similar to that
in Trionyx.

In none of my series of sections can I detect any trace of an

ox 'I'lIE FOETAL MI'rMl'.K'AXI'lS OF t'llELOXFA, 21

iii(lc|)('iulfiitlv i'ufiiK'd vc'sick' which aricfwiinls puts itself in ((tm-
niiinicatinii with ilic liiinl-f|'iit hv an iniii'|)cii(l('nl Iv locincil sfalk.
The fiii'nres üiven ab<)\e siitticiently wai-ranl us in oouchidin^' that in
Clielonia at least, the allantois arises as a d i verti e u I u iii
of the hinrl-o'iit and is from tlie first (;o7i t i ii no us witli
it.

III. Laler Stages of the Foetal Membranes.

hi tiie urecediuii" tuo sections \vc lollowcd sc[i;(r;ncl\- i he L;r<iwth
of I lie anuiion aiid (){' I he allantnis uji loa certain <tau('. It will he
more e(jn\ »nient to treat the later .stages of these meiidiranes
together. As tlie develo[)ment advances, they begin to differ in the
two genera (demmys and Trionvx, until, wlien completed, they
])resent important dillcreiices in their struct tu'cs. As those in
(demmys ])resent in mv cijiinion more primitive relations, I hegin
with tliat species.

(I. i 'h'Uimijs J(lj)<tH/nl.

As llie aliantois pushes itself out as a vesicle into the; extra-
eml)r\'onic C(elomi<' ca\itv. tlie allantoic hlood-Ncssels ai'e soon found
distrihiited in two groups (('ompare f'ig, 27 Tl. \\ .). One (the
right) s(M of arteries and veins is ]>laced in that part <d' the \esiele
fa<'ing antei"iorly while the otlici" (llie lel'l) set is placerl on the posterioi'
external aspect of tluî V(îsic]e. The mannei- of distrihiition of the
blood-vessels exerts a consid<'ra1)Ie influence on the fut lu'c shape of
the aliantois.

As the aliantois spi-eads itsell' over the emhrvo as well as over
the y<dk in the extra-emla-Noni*; Cd-lomic ca\it\', it assumes a peculiar
shape represented in iM"g. ()7 (IM. Virt.). 'l'he vesicle now flattened is
divided by two peculiar constrictions into two [)arts of une<pinl sizes.
The larger part is again subdivided into two lobes by the posterior set

22

K. Mfl'SUKlTi,'!

of blood-vessels. These two lohcs of flu- hii'u'i'i' |>:H't ni:i\- In- ciilUnl
respeetively tin* ri^-ht, and the left lobe, wliilc the smullci- linlf of
the allantoic vesicle may be called the niidillc lobe.

The 1 wo constrictions that divide the middle lobe from tlic larger
half of the allantoic vesicle are caused in two different ways. The
anterior constriction is ^ery easy to explain. It was mentioned
above that one set of the allantoic vessels runs on the anterior side
of the as yet small allantc^ic vesicle. Now, in the rapid growth of the
vesicle, the lines along which blood-vessels run cannot, on account of
their presence, keep up in their growth with the rest of the vesicle,
aiid ai'c iicccssaiMl V left behind until along these lines f hei-e are pi'odn(_'ed
groo\es at tlie bottom of whicli the blood-vessels rim. When the
allantoic vesicle is flattened, these grooves necessarily produce notches
Ol' bays ill the margin of the \esicle, more or less deep according to
tlie size of the bl<x)d- vessels. In the case of the anterior set of the
allantoic \essels, the gi'oove luis bec(^me so deep that the right lobe
and tlie middle lohe on the two sides of it have met again and become
llrndy a])])ressed with each other, so that praeticallv these blood-
vessels are sup))orted in tlieir course by a mesentery- like i'old of the
allantoic \esicle. This explains the origin of the anterior constric-
tion ol' the allantois. In a similai' way the notch tliat divides the
larger part of tlie allantois into the right and left lobes is prodiice(l
bv the ])osterior Ol' lel't set of lilood-vessels, althongh th<' iiot<h is
not as deep as in the anterior constriction and consists ol two or tliree
minor indentations.

The jiosterior constriction of tlie allantoic- \esicle given in fig,
B7 is also not very difficult to explain. There can be no d(»ubt tliat
it is due in the main to the fact that the vesiele finds itsell' iinalile
to spread freely over the eiubrvo on account of the sero-amniot ic
connection. The only thing it can do is to grow round tin; sero-

()-\ THE FOETAL MEMJJKA.NE« OF CUELOXJA. -io

;iiiiiii()tic <'()iiiiccTi(jii, thus piNjduciii^' ;i (lee]» incision in its oiitliiu'.
riie postcfini- ((»list I'ictitjii OWL'S its orii^'iii to tljis circuiiistaiKX'. niid
llms liL'twt'cii tlic mid.lle. and the left. IoIk' there is always interposed
tlie sei'o-aniiiiot ic coimeetion. There ;ire some details of" tin's poste-
ri(.)r (j(jnst rietion wlndi lam not altle to imderstand. The allantois
[irepares to meet the sero-aniniotie eonneetion. sometime before it
reaches in its lirowtli the latter struetnrc (/. c, before there is, so fai- as
I see, any mechanical necessity for a consti'iction) by folcbni^- itself
and pi'odncinii' a constriction. Thus in Tig. G 7 îhe apex of the
[)osterior const I'iction is some distance from the remnant of I lie
posterioi' tube <»f the amnion (which marks the ])()sterior end <>f the
sero-amniotic connection), and is nnu'ked by fohfs of the allantois
slioAvinu' tliemselves as wln'te streaks. The result of this is that in
later stau'es (Fi.i;'. VJ^) the jnjsterior ctjnst rictioii is. near its head,
divided into two liml)S : one contains the sci-o-anniiotic connection
and its termination, the remnant of the jjosteritir tube of the amnion,
and the othei' is simj)lv an incision in the margin of the allantoic
vesicle. In still later stages, the latter is much the deeper of the two
aïid becomes (juite <;ons])icuous (ligs. (ÎS and 71. Tl. \ III.). I am
unable to see an\' nei^essity for tlie existence of this incision. 1 can
not detect anv one large bloo<l-ve^seJ or set of blood-vessels, wliich
might cause it, as the anterior constriction of the allantois is caused
by the right set of allantoic vessels. It aj)pears t() be a congenitally
acquired character.

The nearl\' circular shape of the middle lobe is produced ])y iVie
fact tliat if is necessarily limited in its gi'uwth l.»_y the sero-amniotic
connection which obstructs its front. In fict it is the right and left
allantoic lobes that grow to cover the larger part of tlie yolk-sac.

The right and middle lobes are suj)plied mostly by \\n' I'iglit set
«•f bl(H)d-vessels. while the left IoIjc is sn[)plied by the left set.

24 k. MLTSUKUl.'!.

TIk' ullantui.s Ims not yet in this ^tnge entirely covered the
amnion and the eml^ryo from above so that tlie amnion witli its
sero-amniotie eonnection, and the anlei'ioi" dorsal part of the eiiilirvn
are visible beyond the margin of the middle allantoic lube. Tiie
amnion at these stages does not iit itself ti^btlv over tlie embrvo but
leaves a spacious amniotic cavity arcnind ilie embi-vo. ]']speci;ill\
there is a remarkable snout-like prolongation of the amnion extend-
ing in front of the head.

Fig. ()8 represents an embryo about forty days old. The
allantois has now spi-ead over a lai'ge part of the uj)per half <jf the
Volk, and this extension is due mostly to the right and left lobes and
iKJt to the middle lol)e. A j)eculiarly sharp demarcation between
the middle and left lobes of this stage is due to the fact that the
sero-amnioti<' c(»nnection is placed between them. The remnant of
the posleri<jr tube of the amnion appears as a white triangular ])alch
extending to the left at the head of the ])osterior incision. The long
white streak extending from the same point ol)li(jUely backward
over the back of the endjry»j i^ the sim[)le incision of the allantois
referred to above in Fig. 12 (PL TL). The allantoic vesicle being
of some thi<'kiie>>. the walls of tin incision which extend from the
iiiiM'f tii the outer limb of the allantois are of some depth, and
beim: pressed from abo\c are bent down and show as a wliite conical
streak of peculiar a p})eara nee. Note also the deep anterior constric-
tion with the right set of allantoic vesseU at its bottom.

ll may be remarked in passing that the position of the embr\o
on the Volk is not necessarily as in Fig. 68. fhe embryo is formed
at ai]y jjlace which liappens to be up[ierm!)st when the egg is
dej'(),xitcd. ll' an egg haj»pens to stand on it> end, the embryo Avill
(jccupv the eml oj' the oljlong yolk.

Fig. Gl' (^I'l. \ 111.) giA'cs a side \iew of a somewhat oldi'r

ox THE FOETAL MEMBliAXES OF CHELOXIA.

25

I'uiljryo. The aljunluis hus spread over the larger part of the yolk
ï>u that, ill the tigure, the latter show.s bare only at the posterior part
of the ventral .--ide. The figure shows the left lobe of the allantois,
the sero-ainiiiotic (-«Mjiiection with peculiar structures at its posterior
cud. and the left set of allantoic Aessels with the corresponding incision
in the margin of the alhnitoic vesicle.

ViiS. 7(1 gives a, \ entrai \iew of an cjj;^^' of aliout the same stage.
Tlie rlifee lo1)es of the allantois have n<.)W spread themselves over a
jiart of the lower half and are here \ ery conspicuous. The lol)e that
a])prar> to ilie ol)server's right is tlie left allantoic lol)e. Xext to it
is the middle lobe and finally at the o]>server's left is the i-ight
allantoif lobe. The incision between tlu' middle and the right lobes
is the anlei'ioi- constriction of Figs. ()7 and 68, and at its dorsal end
is found the I'iuiit set of* alhnitoic vessels. The incision between the
middle and the Icfi lohes passing o\er the head of the embryo
correspontls to ihc postcrioi- constricrion of the; earlier stages, and
has the scro-anmiotic connection placed in it. While in Figs. 67
and C)S the cnd)rvo is coiihned to the space l)elow the middle loljc,
in this hgiirc ihc head <d' the end;ryo appears below the left lobe
lo the left of the sero-amniotic connection. This is brought about
by the following circiuustances. In Fig. 67, the embryo lies on its
left side and, the sero-amnioti(.- connection l)eing a[)proximately over
its dorsal median line, we are viewing it. so to s[)eak, from the side.
As the embryo and with it the amnion gi'ow, the embryo comes
again to lie on its ventral surface, a> in the earliest stages, and the
seroamniotic connection ag;iin accomipanying the dorsal surface <^f the
eml)ryo is tiu'ned toward the obser\'er. 'Idle amnion is thus free to
grow to the left of it, under the left allantoic lobe. The amnion
being spacious, the eml)ryo is aT)le to mo\e within it, and the head
may now be seen to the left, or to tlie right, or directly under, the

26

K. MITSUKUlii.

sero-nmniolic connection, altliouuli the ])osition .sliown in Jnu". 71
appears to l)e tlie most normal.

The l)ho(3cl-vessels that pass thi'ono-h the umbilicus at tliese later
stages, are arranged as in Fig. 75 (PI. IX.). The most anterior is
the vitelline vein, then comes the \itelline artery, after it tlie allantoic
artery and last of all the allantoic vein. The last three di\ ide int(j two,
the right and left brandies, soon after their exit from the umbilicus.
The vitelline artery is distril)uted over the surface of the yolk, but
the vitelline vein is somewhat peculiar: it is much larger than the
vitelline artery and while it receives branches from the surface of
the yolk, the main liulk of it enters right into the substance of
the yolk. This no doubt makes the acquisition of nutriment from
the yolk much easier.

I may now proceed to describe the relations of the embryo, the
foetal membranes and the yolk shortly before hatching. (Figs. 71
and 71a PI. \U\. and Diag. VI. Pi. X.).

Tlie yolk sac (Fig. 71«) is now reduced considerably in
size and the three lobes of the allantois have entirely enclosed it.
These three lobes never fuse witli one another, but are permanently
sejiarate. The seams that sejiarate them are roughly speaking
tri-radiite, the center being at the anterior end of the yolk-sac slightly
to the left (to our riglit as we view it from the ventral surface) of
the median ventral line. The seam that extends transversely from
the center towards (he right (to the left of the observer) se]>arates the
middle (phi«.'e<l in front of it) fivtm the right allantoic lobe (jilaced
behind it) and corresponds to the anterior constriction in Pig. 67.
Hence, at its distal end. is found the rio-ht set of the allantoic arteries
and veins. The seam that runs back from the center nearly parallel
with the median ventral line separates the left lobe (placed to its left
or to the ol)server's right) iVoni llic right lobe, and corresponds to the

ON THE FOETAL MEMBRANES OF CHELONIA.

i>7

shallow notch prorluced by the posterior (the leff) set of blood-vessels
in Fig. 67, or to the inrision in Fig. 60. Hence, at its distnl end the
posterior or left set of allantoic vessels is found. The seam that
separates the middle fi'om the left allantoic lobe is difterent from the
other two, for here the two lobes of the allantois cannot come into
contact, being separated bv the sero-amniotic connection. Ft passes
over to the dorsal side of the embrvo (Fig. 71), and its dorsal end
has the triangidar i-emnant of the postei'ior tube of the amnion,
and the peculiar conical white streak caused by the simple incision
of the allantois. (Compare Figs. 68. PI. V[TT., and 12, PI. 11.).

There is one featnre in an e^»; tliiis advanced which deserves
special notice. The white of the egg which disa])peared vei-y early iVom
over the embryo continues to grow smaller and smaller in rpiantity.
But it persists up to a very late date, if it ever disappears entirely.
There is alwavs, even in very much advanced eggs, a
small mass of the white just at the point where the
three lobes of the allantois meet at the lower pole. This
mass seems t(^ have undergone some change in its chemical composition
for it is now much denser, slightly yellowish in color and sticky.
To receive this mass the mend^ranes are often shallowly depressed.
Into the center of this mass of the white a thick low ])rocess of the
membranes penetrates (shown in Fig. 71« on the left allantoic hdie,
just to the right of the sero-amniotic seam), so that when the
membranes are removed, the mass of the white with the central part
hollowed out appe;u-s like a bowl. The cells of the serous envelope
on the surface of this process are peculiarly modihed. They are more
columnar than in other parts (Fig. 21», PI. IV.). Their nuclei are
larger, irregular in shape, and stained deeper. In these cells are found
many large vacuoles which remain unstained. There can l»e n<^
doubt that these cells are absorbing albuminous particles from the

28 K. MITSUKUKI.

ninss of the wliitc. It seems to me that here wehnve in a
very primiti\e condition the strnctiu'C described by
Duval (No. 10) as the placenta in Birds.

'I'he amnion in tliose later stages seems to envelope the embryo
tolérai)! V closely, and its cavity is no longer spacious.

In hatching, the yolk-sac passes into the interior of the body
whei-e it lies for a long time — in fact for several months, for I found
it in young tortoises late in the spring of the year following that
in which they were hatched. I'he amnion is torn into shreds, but
the allnntnis seems to be spbt o])en by the anterior limbs of the
emerging embr\(> along the sero-amniotic scam — if not nlwavs, at
least in some cases, for I have specimens in which ihc adantois has
been cast away iii this manner and is uniiijinvd. The outer shell
whicli has Ijccomc \er\' brittle is easily In-okcn throngli and the
young tortoise emerges into the worhl.

We maA' now cxaiiiinc tlie mi<'roscopic striu-tures of these
mend^ranes. \\e left the scro-amniori«' connection in tlic condition
represented in Fig. 41) (I'l. \'L). Aftci- ihat stage, as tlu; distance
between the amnion and the serous envelope increases and th<' con-
nci'tion 1)ecomes accordingly elongated in its vertical extension, the
epiblast cells in it become flattened in the direction perpendicidar to
the plane of the eoinie<M ion. Fig. 7ß (PL IX.) represents a pai'i
of a section of the sei-o-aininotic' connection from the saine embi-yo as
that from which Fig. S(i (l*L X.) is taken, 'fiie cells show decidedly
the flattening referred to al)ove.

This flattening goes on more and moi'e. but I omit the inter-
mediate stages and proceed to tin- descri])ti(^n of the sero-amniotic
connection in the finished fœtal metnln-anes (Fig. 71, PI. \ II f.)
Figs. 78-80 (PI. TX.) are selected seclions iVom an embryo of the
same stage as thai rejn'csented in Figs. 71 and 71^/ (PI. \'fT[.). All

ON THE FOETAL MEMBRANES OF CHELONIA.

29

three are from the region of the remnant of the posterior tube of the
amnion, Fig'. 78 beinc; the most anterior, and in order to facilitate
the imderstandini>- of these sections, I have introduced a diagram of
this region in Fig. 77. In this diagram, the serous envelope is
represented as spread over all the structures ; the amnion is below it
and is indicated <wly by the line from which the sero-amniotic con-
nection arises. The remnant of the posterior tube of the amnion has
before this stage been modified into a solid eompressed string of
(•('lis and is shown l)y the hea\iest dark line whioli stretches from
tile amnion to the serous envelope. The sero-amniotic connection
is r('])resent('d shaded by parallel dotted lines. Throughont the
largest ])art of its length (froîii its anterior end to near its posterior
end) this structure lies in one plane and is sinijily a membrane stretch-
ing between the amnion and the serous envelope. Bnt on coming to
the r(^gion of the posterior tube (^f the amnion, it makes a sudden
turn of over 90° to the left and goes to the serous termination <>f the
posterior tu1)e, wliicjh latter structure it connects (m its way with the
serous envelope. It will of (jourse be undersiood that the sero-
amniotic connection from its anterior to its posterior end was
originally in one straigiit line and that its |)C(MiIiar bent termination
at this stage was Iii'ouglit about bv its a('(;oin|)an ving the posterioi*
anuiioti(.- tube in all its (.•liang(\s of r(îlîtti\(' position. It is this
peeidiar bent pai't of tlie s(M'o-amniotie (Connection wliich is seen as
the ti'iangidar white ])atch at the dorsnl end of t he sero-:unniotic
seam. (Figs. (18, 61), 71, IM. \'T[l.).

Fig. 78 is from the regi'^'i '*f ^'i^' simple sero-amniotic connection
(the line 1-1 in Fig. 77). in such a section, the sero-amniotic con-
nection is regally a striking structure, forming a broad and con-
spicuous connecti(W lietween the serous envelope and the amnion.
The e])il)last cells in it are now very flat and closely packed. riie

30

K-MITSUKTJRI.

îillantoic lobes become closely applied agjiiiist the mosoblii^t of tbe
connection but nre permanently sepnrate from on<'h other. The
epiblast of both the nmnion and the serons envelope consists of tAvo
layers of cells. The inner la\er of the former and the outer Inver of
the latter consist of very much flattened cells with hirge nuclei —
which, in the case of the serous envelope at least, nre much laru-erand
stained deeper than those of the second layei-. It is the cells of the
outer layer which become specially laro'e in the reo-ion of the placeiitn.
The second or underlying layer consists of cubical cells which in
some places may be present in more than one layer. It is this inner
second layer alone that forms the sero-;nnriiotic connection, the outer
taking no part in it. As reçfards the allantois, the outer limb is
generally much thicker than the iiiner litnb and has many more blood-
vessels distributed in it. The thickness of the allantoic walls is
crossed in all directions by slender spindle-shaped cells.

Fig-. 79 corresponds to the line 2-2 in Fig. 77 just through
the anofle of the bend w^hich the sero-ainniotic connection makes.
The sero-araniotic connection goes here on one side to the amnion
and on the other to the remnant of the posterior tube of the amnion,
which, being now reduced to a thick compressed and somewhat con-
voluted sti-ing of cells, shows in section as lobated cell-masses.

Fig. SO corresponds to the line 3-3 in Fig. 77. The sero-
amniotic connection is no longer continuous with the serous envelope
but goes to the remnant of the posteri<n' tube of the amnion. To the
rigfht of the sero-amniotic connection, the two lobes of the allantois
meet but are not fused. This is the section of the conical white
streak that stretches over tlie dorsal region of the embryo and (V)rres-
ponds to the simple incision of the allantois given in Fig. 12 (PI. Tl.).

In a few more sections, the sero-amniotic connection disappears.
The two lobes of the allantois, however, keep separate and meet in

ox THE FuETAii MEMBL'ANES (JE CHELuMA. 31

a sieaiii for many more sections, a>; the .simple incision of the alhmtois
extends considerably further postericjrly than the most posterior
point of the sero-amniotic connection. Finally, however, the
allantoic cavitv is continued across. As the incision is deeper in the
iimer liml) of the allantois than in the outer, the allantoic cavity first
becomes continuous near the external surface and then u'radually
extends tcjward the inner surface.

h. Trknujx Jaitonieu.s,

As in Clemmys. the allantoic blo(jd-\essels group themsehes
into two sets : the antei'ior (or the right) and the posterior (or the
left), while the allantois is still a small vesicle (Fig. 27 PI. l\.).
AVhen the allantois has advanced simiewhat in its de\'elopment, it
presents the shape represented in Fig. 72 (PI. VIII.). This cor-
responds to Fig. 67 of Clemmys but presents some important
differences. The allantois consists here of two lobes marked off
from each other bv two constrictions. One of these is just l)ehind
the eve and the other is directly opposite the first on the opposite
side of the ^•esicle. Pnlike Clemmvs, both these (/«jnstrictions are
pi'odiiccd in the same way. That is, the line along which each
set of blood-vessels jwisses from the inner to the outer liml) of the
allantoic Aesicle is left behind in its growth, and the parts on each
side of" the >anic hni' growing faster and meeting each other sotjn
pnjdiice a mesenterv-like f)ld slingiiig these blood-vessels. In
tjther uords, Ixjth the constrictions of Tri onyx are of the same
nature as the anterior constriction of Clemmys (Fig. 67). The
posterior constriction of Tri onyx is not well-marked in Clemmys:
it corresponds to the shallow notch caused by the posterior or left
set of blood-vessels. ( )n the contrary, that corresponding to the

32 11. MlTSUKUKi.

pD^tcrior (•(»ii.'^t ricli(jii of" (Icniiny.s — tu that caitsoil Ijy the prij.seiice
of the .seru-iiiJiuiutic (oiiiiectinii — is never [)roduced iii Ti'ioiiyx,
iilthough at the .spot whei'e it (ju;Li'ht to lie ])roduced, \\z., opp(jsite the
remuant of the posterior tul)e of the amnicjii, the alhiiit(jis is draw ii out
to a peculiarly ,siiape<l point as if it were trying;' to uo round the senj-
amniotic (•onneeti<ju. From these considérât ions, it follows that that
part of the Trionvx allaiitois in fi-oiit oï the aiilerioi- constriction
corresp(jiids to tiie ri<^"ht loljc of (I em m vs, the part fronî the anterior
(•onstrictitjn to the ptjint op])osite the I'emnaut of the posterior lid»e
of the amnion to the middle lobe, and the part lTo]n I lie same point
to the j)osteri<jr constriction to the left loljc. flic middle lohe faces,
as in (demmys, the sero-aniniotic conne<'tion.

Unlike ( ' 1 e ni m y s . h o w <ô\ e r . t h e if r o \\' t h of t li e
middle lohe pushes, so to speak, the sero-am n i oti c
connection before it, so that the amnioii comes gi'ad u a I -
Iv to assume the shape of a bag «jf which the sero- am-
niotic connecti<jn forms the puckered month. Keference
to Fias. 25 and 2(5 (1^1. lA.) will make these ])rocesses clear.

In Fi"'. 25, the sero-amniotic connection is still directly over
the dorsal side of the embryo and it is still straight. The allantois
is pressing on it.

In Fia'. 26, the arowth of the amnion has removed tlie jnain
j)ortion of it from the sero-amniotic connection. The Avhole amnion
has now assumed the shape of a l)ag hanging ])cndant by the sero-
amniotic connection. The allantois is still pressing on the sero-
amniotic connection, and its i)ressure, so to speak, has bent the hitherto
straiafht sero-amniotic connection like the letter \. Moreover the
a'eneral axis of the sero-amniotic connection which has hitherto been
parallel with the axis of the embryo is now at right angle with the
latter. The general appearance of the c^^g at this stage as seen from

ox THE FOETAL MEMBRANES OF CHELOXIA. OO

tlie \eijti-al sfide is gi\ eii in Fig. 7o (1*1. \'JII.). The allautoi.s ha.s
covered a larger j)arl of the yulk >a<'. lea\ iug ojiIn an oxal .s|»ace at
the lower [)n\r mieovered. Thi.s o\al .space i.s hounded anteriorly hv
the sero-aniniotic connection hent hke the letter \ , and from tliis is
seen stretching" tbrwai'd tlie anterior prolongation of the amnion
whicli unites the main portion of the latter with the sero-amniotic
connection (Comp. J)iag. VIL. IM. X.). The two constrictions or
meseuterydike folds of tlie allantcjis caused by the two sets of
blood-vessels are also seen distinctly in the tigure.

The final shape <jf the fœtal membranes is seen in Tig. 74 (1^1.
\ Jli. Comp. Diag. V'JL, PI. X.), and tliat part of the ventral surface
where the lobes of the allaiitois finally meet is represented in a more
enlarged scale in J: ig. 81 (1*1. IX.). The space left uncovered by the
allcintois in Tig. 78 is now mostly grown over by the growtli from the
posterior side. There is however a small space still left uncovered
by the allantois. It is triangular in shape; the apex of this triangle
is bounded anteriorly by the sero-amniotic connection, now com-
pressed to an irregidar horse-shoe shaped opaque streak. On two
sides of the triangle is the anterior allantoic lobe (which equals the
middle and left lobes): posteriorly, it is limited by the posterior
allantoic lobe (which e({uals the right lobe). From the lateral angles
of this triangle the mesentery-like fold of the allantois stretches on
each side to each set of the allantoic vessels. These correspond to
the two notches in F^ig. 72. The view fr<jm inside of this reofion
is given in FTg. Sla. In this there is a ridge across the middle of
the triangular area. This is the line of junction of the yolk-sac and
the amnion, as will become clear from the sections to be described
directly. It appears in the external view as an opaque line across
the horse-shoe shaped sero-amniotic connection (F'ig. 81). The
anterior [n'olongation of the amnion connecting the main body of

34 K. MITSUKUitl.

the latter with the .sero-anmiotic coiiiiectioii is now ,so dispropor-
tiunatelv small (-(Jtupared with the aiiiiiiou itself that it appears as a
small irivi;-iilai'ly triaiiuuhir white pateh (l"'i,u". 8L(f, Aiit. Proloiiii'.
Amii., and 1' i;L:. <Sl). Its oaA itv. n(3w ahiKJst ohliterated t)\' the aj)pres-
sion of its two walls, js still ('ontiniKjiis with the amniotie cavity
throiiiih n narrow slit (see Via;, (ila and Diaii'. A'll). It leads to the
sero-anmiotic connection. I'ig«. 82-85 (PI. IX.) are a series of sec-
tions of this region, the plane of sections being in the antero-posterior
direction. Fig. 82 is from the left side of the triangular unc(jvered
area, in Fig. 81« (or from tlie right side in Fig. 81). The two
allantoic lobes, the anterior and posterior, are only slightly apart from
each other. Attention is called specially to the section of a part of the
anterior prolongati(jn. Figs. 83 and 84 are from the triangular area
itself. Here the two allantoic lobes are wide a})art. The interspace
is tilled up mostly by the growth of the somatic mesoblast of the
serous envelope. The sero-anmiotic connection in Fig. 83 is of a
very complicated figure. This arises from two reasons. Jn the
first place, the section being taken in the antero-posterior direction,
it cuts ()ne lindj of the horse shoe shaped sero-amniotic (.-onnectioji
more or less longitudinally. In the sectjnd place, the sero-amniotic
connection is not in a simple straight line as in (Jlemmys, but
being compressed and rultled. ap])ears as of an irregular ])attern
in a section. In Fig. <S4, the sero-amniotic (•onnection is cut nnjre
directly across. To give an idea of the structiu'e of this region. I
have gi^ en a. part of Fig. 84 on a more enlarged scale in 1 ig. 84(/.
In Fig. So. wlii<-li was hi-oken by accident, the two allantoic lobes
have again a])j)roached eiicli other, the section being out of the
triangular area. The\" are, however, distinct and continue distinct
until the sections reach the allantoic vessels.

In Tri onyx, the renuiant of the posterior tube of the allantois

ON THE FOETAL MEMBRANES OF CHELONIA. 35

is not to be olearly distino-ui.slied. The ])]ace where it should be
present is drnwn ont to ;t point (Fio-. 26, PI. IV.): thnt seems to be
all the indication I'emninino- in later stao-es of rhe posterior rube (tf
the amnion.

The white remains to the last in Tricmvx ns in (MeniTnys.
!ind is found opposite the triana'idar area of the \entr;d |iole left un-
covered by the allnntois. This |)nrt is o'enerally more or less
depressed to receive the nr^iss which is sticky niid yellowish. The
outermost cells of the serons enxelope in this area imdero'o modificn-
tions similnr to those of the corresponding' spot in Clemmy s . Tlie\'
become taller and lar^'er and contain large vacuoles, their nuclei become
larger and irreo-nlar in shape and stained deeper (Fig. S4(f., Comp. Fig.
29, PI. ÎA'.). In Tri onyx howe\ er. there appears to be no process
that penetrates into the white as in Clem my s .

The yolk passes inside the embryo in hatching.

Of the completed fœtal membranes of Clemmy s and Tri onyx
above described, there can be no doubt that Clemmy s has retained
more primitive relations. The main ground for this (Conclusion is that,
starting from the same ipoint, (bfferent structures (above all, the sero-
amniotic connection) retain in Clemmy s tlieir original positicni and
arrangement, while in Tri onyx various structures are disturbed
from their first arrangement, the sero-amniotic connecticm being pushed
forward, l)ent and C(^m])ressed into a secondary shape. If the process
begun in Tri onyx were to go one step further, no spot would be
left uncovered by the alia uteris, the sero-amniotic coiniection might be
pressed out of existence, the two allantoic lohes might come in (^ontact
with each other, and then the condition liitherto accepted as occurring
in Birds would be the result.

3^! K. MITSUKURI.

General Considerations.

The notewi^rthy features in the history of the foetal meniliranes
of Chelonia as given above are : —

1. The presence of the Proamnion and the manner
in whicli it is replaced by the jiermanent Amnion.

2. The presence of a pecnliar tube stretchinis-
posteriorly from the posterior end of the Amnion
e o n n e c t i n g the c n v i t y of the latter with the ex-
terior— the Posterior Tube of the Amnion.

3. The permanence of the Sero- Amniotic Con nee -

t i O 71 .

4 . T h e d i tf e r e n c e s in the fa te of t h e S e r o - A m n i o t i c
Connection in Tri(^nyx and CI em my s ,

5. The presence (^f the rudimentary "Placentn."
Of these, the iirst point has been noticed in nearly all tlie

amniotn wliose development has lieen carefully studied within reeent
years. The only new feature is the fact thnt the dorsal part of the
proaniîuon consists at tirst solely of a solid sheet of epibhist
cells. The second, thii'd, and foiu-th points are. so far as
I am aware, broiiuht out for the first time in the coiu'se of the
present investigatioji. They certainly are vovy I'emarkable features,
and, so fai" as (»ur present knowledge goes, might be looked on as
distinctive of tlie ("heloniati fietal ]n«'inbi'anes. 1 think it, however,
highlv |)rol)aMe that if other gi-oiips of |{e|»riHa aii<l lîirds are
careful] V gone o\('i' again, maiiv struct ures wliicli are liighh sio-ui-
ficant in the light of the facts now <)btaine(l will lie found to have hitherto
escajied notice oi- been laid aside ;is iniim|>ortaiii . l^'or instance, in
the sections of Lacei'ta gi\(Mi bv se\(M"al antliors. the sero-aiiiniotic
connection is (listinetively lii^iu'ed v.xrw up to comparatively late

ON THE FOETAL MEMBRANES OF CHELONIA. 37

stages. Being possessed with the idea obtained fi-om the study of
Birds thnt it is soon to disappear, different writers have not thought
it worth theii- while to follow its history further. Nevertheless I ran
not but think that the sero-aniniotir- connection rmis a similar course
ill other groups of Re))tili;i as descril)e(| now for Ohelonia. I also
think that the posterior tube of the amnion is not such an unique
structure as it appears to V)e at present.

The fifth point, the pi-esence of the rudimentary placenta, is
certainly very interesting. If the depression into which the white
is received shoidd Ijeconie deeper, ajid the allantoic folds -should be
produced to enclose it, we shall have exactly the same structure as
"the placenta " described by Duval in Birds.

The Re])tilia, being the lowest group of the Amniota, are
of great importance in tlic (•<)niparative study of the foetal
membranes. What light dc^cs the history of the Chelouian fœtal
membranes as given aliove throw on the phvl<\f;"eny of those
membranes in the \'ei-febrat;i ? Witliout going into a o'enet'a.l disiiis-
sion of this difficidî prol)letu. I think T might offer here a few
suggestions which have |)f(-sented themselves to me in the course of
the present investigation. T strongly incline to the \ iew thnt the
amnion was originally descloped by mechanical cjuises. In Ohelonia,
when the head fold is |»ro(bicetl. there are two reasons why it should
sink into the yolk helow. fn the first place, the yolk is very Inro-c
and licpiid. especially just beneath the blastoderm, so that a slight
weight is enough to sink any struetiu'e into it. In the second place,
the white rapidly disappeai's from oNcr the blastoderm, which adheres
then firmly to the sh<'ll-niembrane : hence there is no space fo)*
Ihe head-fold to grow except towards l)elo\v. Without asserting
tluit these are the very same causes that produced the anterior

38 K. MITSUKTJRr.

fold of the amnion, I think it reasonahle to assume that it was
prodiieed by some such meehanical means. En this relati(^n, I think,
those inconstant adventitious folds ns o-iven in Fiu's. 1 and 2 (PI. T.)
are hio-hly significant. These iind(^ubtedly arise l)y the neio-hhorino-
parts sinkino- heh)w. We mit>ht suppose that in tlic earlier stufj-es
of development many such folds are produced, different in different
emhrvos aeeording to their individunl idinsynerar-ies, and the anterior
inid of the amnion may be looked upon merely as one of these.
Oidv the canse that ))rodnces it being present in all the embryos
:ind acting permanently and augmenting steiidily, finally gnve rise
to the structure whieh we call the ninnion, heredity (-)f course helping
a great deal.

The anterior fold of the amnion, when produced, consists only
of the hypoblast and epiblast, nnd is called the Troamnicm. AVe now
know that this is fonnd in all the groups of the Amniota, and I think
we ought to add the stage of the proamnion as of normal occurrence
n the development of the amnion. In Chelonia. the dorsal part
of the proamnion is for some time entirely epiblastic. Should this
be looked upon as a |)rimitive feature or as a secondary one? I am
inclined to adopt the former view for two reasons : —

(1). The inconstant adventitious folds are, as previously stated,
always pui-ely e])iblastic and exactly like the lateral folds of the
proamnion (Fig. 80a, PI. A'.): hence, it is reas<^nable t(^ conclude that
all such folds produced on the surface «^f the blastoderm are ;it first
always purelv epiblastic, and the solid epiblastic dorsal sheet of the
Proamnion produced by the coalescence of the lateral folds of two
sides have reason to be simply epiblastic.

(2). Tn Olemmys, whose deNt'lopmetit is certaitdy more pruni-
tive than that of Tri onyx, the solid dorsal sheet persists for a longer
time than in the latter genus, and thei-e is a ccmsiderahle interval of

ON THE FOETAL MEMBRANES OF CHELOXIA. o!^

time before the meso})lastic folds insinuate theniselves into the epi-
blastic sheel. I think, hijvvcver, thai ahliDUg'h these folds are solid
and \vith<.)ut any cavity, they ought t(.) he regarded as consisting of t\V(j
Jinibs, the inner and outer, which are hrinl\' ap|)ressed against each
otlier : otlierwise there is no reason \N'hy the .sero-anini(.)tic connection,
which ought to be regarded as the seam al«jng which the folds of the
two sides have met. should remain permanently and sej)arate the
mesoblastic idlds of the two sides to the end.

If the doi'sal part of tiie proanuiic.n consisted primarily of t he
e[)ibhist ahjne, why sliouM the meso]>lastic folds afterward insinuate
tliemsehes between tlie two limbs of that ])art, thus extending tlie
extra-eml)ryonic c<r']omic ca\it\ into that region? For the explana-
ti(jn ofthat process, I adopt Jxdfour's view. To give efücac\' to the
aliantois as a respiratory organ, it is desira])le that it should l)e
spread as extensixely as possibk^ chjse under the surface of the egij!; ;
hence the extra-em Ijrycjnic «-adomic cavity must ha\e spread pai'i-
jHissii willi the gradual growth of the aliantois. The extension of
tile folds of the somatic mesobiasf into the epiblastic fdds of the
[)roaninion is. I think, due ])rimai-i]y tn this caiise. That the
mesohJast s))reads itself at present long Ijetore the aliantois, is to be
ex])lained as a case of precocious dc^eh^j^nent.

rile sero-amiiioidc coiiiiection is, in m\' oitiiiion, decidedly a
pi'imitixe structure The manner in which t he aliantois sjU'eads itself
in (Memniys hy r»junding the sero-amniotic connection can also
be explaine(l only on ])hylogeiiet ic grounds. The manner in which
the allantoic blood-vessels -are slung <jn mesenter\ -like folds of the
aliantois, is, i think, also a primitive feature. The manner in
which the sero-amniotic c<jnnection in Trionyx is pushed forward,
bent and compressed, points out, I think, the way in which that
structure historically disappeared in higher forms. As i have stated

40 K. Mi'J'ISUKUlll.

;ibo\e, if tlic proce.-s.s liegiiii in Triuuyx i.s ("irried jn.sit une ste}»
litrther, the seru-ainniutic conne(.-ri<)n would ceu«e to exi.sl. Wluit i«
the cause which Ijrouulit about this di«a])[)earance? So far a« I can
«ee, the .sei'o-amniotic coiinectioii serxef- no |)ractical purpose in Ciem-
uiys and its presence is only to be accounted for phylogeneticaJly.
If such is the case, it would be undoul)tedly ecoiiomical to skip over
the roundabout iiiaiiner by which the aUantois spreads itselt in
Clem in \ s round the sero-anniiotic c(jnnecti(jn. Hence its dis-
appearance at last in higher forms. Whether the innnediate agent of its
forward shifting is the force exerted solely by the growing edge of the
aUantois i cannot tell. It is no doubt partly due to that, but in addi-
tion I oiler the following as a suggestion. In T r io n y x , the alhiiitoic
vessels come out synunetrically on each side; in Clemniys, the
symmetry is disturbed, the right set is found more anteriorly than
the left. As I have often remarked, CI em my s presents on the whole
more primitive relations, but I cannot regard this asymmetry of
the allantoic blood-vessels as a |)riiniti\ e (•(jndition : something being-
present in Clemmys has disturbed the <n-iginal symmetry and being-
absent in Tri onyx no longer intereferes with it and this something
I think is the presence of the sero-amniotic connection. May not
the tendency of the blood-vessels to assume a symmetrical arrange-
ment help to push the sero-anmiotic connection forward in Triony x?

ON THE FORTAL MEMBEANÇS OP CHELONIA. 41

List of Works referred to.

1. H. Strahl : — Ueber die Entwiclvlung des Canalis Myeloentericiis

u. d. Allantois der Eideclise. Arch. f. Anat. ii.

Physiol. 1881. Aiiat. Abth.
2. : — Beiträge zur Entwicklung von Lacerta agilis.

Ibid. 1882.
3. : — Beiträge zur Entwicklung der Reptilien. Ibid.

1883.
4. : — lieber Canalis neurentericus u. Allantois bei Lacerta

viridis. Ibid. 1883.
5. : — Ueber Entwickluno'svorträno'e am Yorderende des

Embryos von Locerta agilis. Ibid. 1884.
6. C. K. Hoffmann : — Beiträge zur Entwicklungsgeschichte der

Reptilien. Zeit, für Wiss. Zoöl. Bd. 40. 1884.
7. : — Weitere Untersuch uneven zur Entwicklunocssfe-

schichte der Reptilien. Morph. Jahrb. Bd. 11. 1886.
8. P]d. Ravn : — Ueber die mesodermfreie Stelle in der Keimscheibe

des Hiihnerembryos. Arch. f. Anat. u. Entwickl.

1886.
9. : — Bemerk, ueber die mesodermfreie Zone in der

Keimscheibe der Eidechse. Anat. Anz. IV. 1887

No. 5. , . ,

10. M.Duval: — Etudes histologiques et niurphologiques sur les

Annexes des Embryons d'Oiseau, Journ. de l'Anat.
et delà Physiol., XX. 1884.

11. E. Van Beneden et C. Julin : — Recherches sur la Formation des

Annexes fœtals chez les Mammifères (Lapin et
• Chéiroptères) Arch de Biol. V. 1884.

12. E. Gasser : — Beiträci^e zur Entwickluno^so^esch. der Allantois der

42 . K. MITSUKUKI.

Miiller'schen gänge u. des Afters. Frankfurt 1874.

13. : — Der Primitivstreifen bei Yogelembryonen (Hiihn u.

Gaus) Cas sel. 1879. . .

14. K. Mitsukuri and C. Ishikawa : — On the formation of the Germinal

Layers in Chelonia. Quart. Jour, of Micros. Sei.
1886. Also Jour, of Science College. Imp. Univ.
Tokyo, Japan. Vol. 1.

15. A. OstromiiotF: — Zur Entwickhmofsg'eschichte der Eidechsen

Zoöl. Anz. Nr. 292.

16. Jv. Perenyi : — Entwicklung des Amnion, VVolff'schen Gauges u.

der Allantois bei den Reptilien. Zoöl. Anz. Nr. 274.

17. A. A, W. Hubrecht : — The Placentation of Erinaceus Europeus,

with Remarks on the Phylogeny of the Placenta.
Quart. Jour. Micros. Sei. Vol. XXX. 1889.

18. J. A. Ryder : — The Origin of the Amnion. Amer. Natur. 1886.

19. A. Fleischmann : — Mittelblatt und Amnion der Katze. Erlan-

gen. 1887.

Explanation of Figures in Plates I — X.
List of Reference Letters.

* a. (Fig. 2) iucoustaut tidveutitious folds, a. I. f. anterior limiting fuvrow=
vordere Grenzfarche. All. Allantois. Amn. Amnion, b.v. blood-vessels. Coel.
coelom witliin tlie embryo. CoeV. extraembryonic coelom. vh. notochord. Ej>i.
Epiblast. H. f". head-fold. H///^ hypoblast. Lrt^/. J )wt. Lateral fold of Amnion.
Mes. mesoblast. ^V. R. Can. neureuteric canal. Post. Ta. Amii. Posterior tube of
Amnion. j)ro.v. />t. proximal part of posterior tube of Amnion. Proam. Proamnion.
Reimiatit. Post. Tu. Amn. Renmant of posterior tube of Amnion. Ser. Knv. Serous
envelope. Sero-Amti-Conn. Sero-Amniotic connection, v. v. a. anterior vitelline vein.
yk. yolk.

In colored figures of sections, the epiblast is always colored red, the mesoblast
blue, and the hypoblast yellow. In PI. IX. blue stands for the somatic mesoblast,
and green for the splanchnic mesoblast.

ox THE FOETAL MEMBKANES OP CHELONIA. 43

Plate. I.

Fig. 1. Dorsal view of a Clemmys embryo 2 days old. Zeiss
aa X 2. ^"^ '

Fig. la. Ventral vie\v of the same, aa x 2.

Fig. 2. Dorsal view of a Clemmys embryo 4 ^ days old, with 2-3
mesoblastic somites, aa x 2. rxxxsix.)

Fig. 2a. Ventral view of the same.

Fig. 3. Dorsal view of a Clemmys embryo 4 days old, with 6-7
mesoblastic somites. Extra-embryonic cœlomic cavities of
two sides distinctly seen almost touching each other ov(!i"
the median dorsal line of the embryo, aa x 2. ^"^^-^

Fig. 3a. Ventral view of the same, aay.2.

Fig. 4. Dorsal view of a Clemmys embryo 7 days old. aa x 2.

(xxxxi.)

lug. Ö. „ „ „ 8 „ „ aa x 2.

(XXXXIII.)

Fig. 6. „ „ „ 9 „ „ aa x 2.

(xxxix.)

Fig. 7. „ „ „ 4 ^ „ „ aa x 2.

(LIII.)

Plate. II.

Fig. 8. Clemmys embryo slightly older than Fig. 7. Enlarged.

(1.)

Fig. 9. Clemmys embryo 13 days old. aa x 2. (xxvm.)

44 K. MITSUKÜßl.

Fig. 10. Posterior tube of the Amnion higlily convoluted, from
a Clemmys embryo 14 days old. aa x 2. ^'''"'•^

Fkj. 11. Dorsal view of a Clemmys embryo, 10 days old, 6^^ mm.
long. Enlarged about 17 times. ^"'"'^

Fig. 11a. Ventral View of another embryo from the same deposit.
Enlarofed about 17 times.

Fig. 12. Semi-diagramatic view^ of the posterior constriction of the
AUantois in a Clemmys embryo 31 days old, seen from
outside the serous envelope, ca x 7. ^"'^'^

Fig. 13. Two Clemmys embryos 18 days old. Slightly enlarged.

(xxxxv.)

Fig. 14. Dorsal view of a Clemmys embryo whose amnion is open
toward the left. '^^^""•'

Fig. 15. Posterior tube of the Amnion disappearing. From a
Clemmys embryo 13 days old, 8 mm. long.

Plate. III.

Fig. 16. Dorsal view of a Trionyx embryo 34 days old. a« x 2.

(126.)

Fiq. 17. Dorsal view of a Trionyx embryo o.i It y old, 3 mm.

long, witli 5-6 mesobl. somites, aa x 2.

(128.)

Fig. 17a. Ventral view of the same.

Fig. 18. Dorsal view of a Trionyx embryo 74 days old, 3t\ mm.

long, with 7-8 mesoblastic somites, aa x 2.

(141.)

Fig. 19. Dorsal view of a Trionyx embryo 8^ days old, 4 mm.

long, aa x 2. ^'"'-^

Fig. 19a. Ventral view of the same, aa x 2.

Fig. 20. Ventral \'iew of the posterior part of a Trionyx embryo

ON THE FOETAL MEMBRANES OF CHELONIA. 45

8 dnys old, showing the beginning of the Allantois. A A x2.

(112.)

Fig. 21. Posterior tube of the Amnion in four Trionyx embryos
13 days old. Slightly enlarged. ("--^

Fig. 22. Posterior tube of the Amnion in two Trionyx embryos
16 days old. aa x 2. <i6^-)

Plate. IV.

Fig. 23. Dorsal view of a Trionyx embryo IOt days old. aax2.

(157.)

Fig. 24. Dorsal view of a Trionyx embryo 114 days old, 5^ mm.

long. Posterior Tube 2 + mm. long. *^"'

Fig. 25. Trionyx embryo 38 days old, seen from the side of the

yolk-sac which has however been removed. (^"^•'

Fig. 26. Trionyx embryo 53 days old. The yolk-sac removed and

the embryo seen from the ventral or yolk-sac side. x3.

079.)

Fig. 27. Trionyx embryo 10^ days old. ^^''^•>

Fig. 28. Same embryo seen from its dorsal aspect, with the serous

envelope lifted up, showing the sero-amniotic connection

and the remnant of the posterior tube of the Amnion.
Fig. 29. Cells of the serous envelope in the region of the " placenta "

in the Clemmys embryo represented in Figs. 71 and 7la.

DD X 4.

Plate. V.

Fig. 30-33. (5-8). Selected transverse sections from the Cleinmys
embryo represented in Figs. 2 and 2a. CC x 1.
Fig. 30. From the region of the lateral limbs of the Amnion.

46 K. MITSUKURI.

Fig. 31. From the region where the two lateral limbs have just

united.
Fig. 32. From the region where the head is partly sunk below

the level of the blastoderm.
Fig. 33. From the region of the head which is wholly sunk
below the level of the blastoderm.
Fig. 30a. Region of the left amniotic limb in Fig. 30. under a higher

power. BD x 4.
Fig. 31a. Left half of the amnion in Fig. 31. DD x 4,
Figs. 3-1-38. Selected transverse sections from the Clemmy?^ embryo
represented in Fig. 3 and 3rt. CG x 1.
Fig. 34. From the region of the lateral limbs of the Amnion.
Figs. 35-36. From the dorsal region.
Fig. 37. From the region of the heart.
Fig. 38. From the region of the liead.
Fig. 3ôa. Median dorsal part of the Amnion in Fig. 35. under a

higher power. DD x 4.
Fig. 36a. Median dorsal part of tlie Amnion in Fig. 36. under a

higher power. DD x 4.
Fig. 361). The same region a few sections in front of Fig. 36.

DD X 4.
Fig. 37a. Median dorsal part of the Amnion in Fig. 37. under a

higher power. DD x 4.
Fig. 39. Transverse section of the posterior tube of the Amnion
from the embryo given in Fig. 5. near its proximal end.
OCxl.
Fig. 40. Do. from about its middle. (70 x 1.

Fig. 41. Longitudinal Secti<ju, slightly out of the median line, of
a Cleuunys embryo from the same stage as that represented
in Fig. 3. BB x 2.

ox THE FOETAL MEMBRANES OF CHELONIA.

47

Fig. 41a, DiagTammatic longitudinal section of a Clemmys embryo
somewhat older than that o;iven in Fig. 41.

Plate. VI.

Figs. 42-47. Selected transverse sections from a Clemmys embryo

6 days old with 20 mesoblastic somites. CC x 1. '^^^"•>
Figs. 44a. Median dorsal part of the Amnion in Fig. 44. under a

higher power. D.Z) x 4.
Figs. 48-49. Selected transverse sections from a Clemmys embryo
9 days old. CC x 1.
Fig. 48. From the tail-region.
' Fig. 49. From the dorsal region.
Fig. 48a. Median dorsal part of the Amnion in Fig. 48. under a

higher power. BD x 4.
Fig. 49a. Median dorsal part of the Amnion in Fig. 49. under a

higher power. DD x 4.
Figs. 50-52. Selected transverse sections from the embryo repre-
sented in Fig. 17. CC X I.
Fig. 50. From the region of the lateral limbs of the Amnion.
Fig. 51. From the dorsal region of the Amnion.
Fig. 52. From the region when the head is sunk almost entirely
below^ the level of the blastodem.
Figs. 53-55. Selected transverse sections from the posterior tube of
the Amnion in the embryo represented in Fig. 24. Only
the epiblast and somatopleuric mesoblast are represented,
the hypoblast and splanchno})k'uric mesoblast being
omitted. DD x 4.
Fig. 53. Near the posterior opening of the tube.

48 K. MITSUKÜEI.

Figs. 54-55. At various distances in front of Fig. 53.
Fig. 56. Median dorsal part of the Amnion in a section from the

middle dorsal region of the Trionyx embr3o represented

in Fig. 24. DD x 4.
Fig. 57. Median dorsal part of the Amnion in a section from the

dorsal region of Trionyx embryo represented in Fig. 19.

DDx4.

Plate. VII.

Fig. 58. Longitudinal section of an embryo from the same stage as

that represented in Fig. 1. DD x 2.
Fig. 59. Transverse section of the embryo represented in Fig. 1.

DD X 2.
Fig. 60. Longitudinal section of the Trionyx embryo shown in

Fi^. 20. OCX 2.

Fig. 61. Longitudinal section of a Trionyx embryo lOj days old.

GG X 2. ^'"-^

Fig. 62. Longitudinal section of a Trionyx embryo 9 days old.

GG X 2. *"'-^

Fig. 63. Longitudinal section of a Trionyx embryo IH days old.

GG X 2. ^'''•'

Fig. 64. Longitudinal section of a Clemmys embryo 4 days old

with 16 mesoblastic somites. BB x 2. ^^^^-^

Fig. 65. Lorjgitudinal section of a Clemmys embryo 6 days old

with about 20 mesoblastic somites. BB x 2. ^^^^"^

Fig. 66. Longitudinal section of a Clemmys embryo 9 days old.

BB X 2. ^^""'^-^

ox THE FOETAL MEMBRANES OF CHELOXIA. 49

Plate. VIII.

Fig, 67. Surface view of a Clemmys embryo 28 days old. Seen
from outside the serous envelope. x 4^. ^''^^'•*

The upper transparent membrane is the serous envelope»
The lower opaque membrane with blood-vessels is the
yoîk-membraiie. Between these two membranes are
placed the enihnjo, the allantois &c. Different divisions
of the allantois are sufficiently explained in the text.
The white line close to and parallel with the median
dorsal line oi the embryo is the sero-amniotic connec-
tion. : traced posteriosly, it bends sharply to the left,
this short liml) being the remnant or proximal part of
the posterior tube of the amnion. Over the posterior
part of the embryo, is a clehcate, irregularly curved
white tube : this is the distal part of the posterior tube
of the amnion with its horse-shoe sha[)ed posterior
opening. It has no connection with the proximal part.

Fig. 68. Dorsal view of a Clemmys egg, with the embryo, the fetal
membranes, and the yolk-sac. About 40 days old. x 2.

(LXSII.)

Fig. 69. Side view of a Clemmys egg with the embryo, the fetal
membranes, and the yolk-sac. 51 days old. Nat. size.

(lxxiii.)

Fnj. 70. Ventral view of a Clemmys egg with the embryo, the foetal
membranes and the yolk-sac. 55 days old. Blood-vessels
on the yolk-sac omitted. Nat. size. a-ssv.>

B'ig, 71. Dorsal view of a Clemmys embryo, shortly before hatching
with the fœtal membranes. 45 days old. ^''^'^•*

50 K. MITSUKURI.

Fig. 71a. Ventral view of the same.

A low lobate process of the membranes situated close
to the left of the tri-radiate allantoic seams penetrates
into the mass of the white.

Fig. 72. Surface view of a Tri onyx embryo 15f days old. X 5^.

(1-7.)

This corresponds to Fig. 67. of Clemmys, and the ex-
planation of the latter is applicable to this. The
white line stretching from the neck of the embryo to
its posterior end is the sero-amniotic connection. Its
slight posterior expansion marks the remnant of the
posterior tube of the amnion.
Fig. 73. Embryo represented in Fig. 26. with the yolk-sac and the

foetal membranes. Blood-vessels on the yolk-sac omitted.

Slightly enlarged. ^^^^'^

Fig. 74. Ventral view of a Trionyx embryo 42 days old with the

yolk-sac and the foetal membranes. Slightly enlarged.

(ISl.)

Plate. IX.

Fig. 76. Blood-vessels that pass through the umbilicus.
Fig. 76. Part of a transverse section through the sero-amniotic
connection of a Clemmys embryo 13 days old. DD x 4.

(LIX.)

Fig. 77. Diagram of the region of the posterior tube of the Amnion.

Figs. 78-80. Selected transverse sections through the posterior part
of the sero-amniotic connection and the remnant of the
posterior tube of the Amnion in the Clemmys embryo
represented in Fig. 71. CC x 2.

ox THE FOETAL MEMBRANES OP CHELOXIA.

51

Fig. 78. Through the line 1-1 in Fig. 77.

Fig. 79. „ „ „ 2-2 in Fig. 77.

Fig. 80. „ „ „ 3-3 in Fig. 77.

Fig. 81. Region on the non-embryonic pole of the yolk-sac where
the allantoic lobes meet. From a Trionyx embryo similar
to Fig. 74. Seen from outside the serous envelope, x 3.

(182.)

Fig. 81a. The same region seen from inside.

Figs. 82-85. Selected sagittal sections through the region repre-
sented in Figs. 81 and Sla. aa x 2.
Fig. 82. is to the extreme left of Fig. 81a. and the sections
gradually proceed toward the right.
Fig. 84a. Region of the sero-amniotic connection in Fig. 84. more
highly magnified. DD x 2.

Plate. X,

Fig. 85. Transverse section from the head-region of the Clemmys

embryo represented in Fig. 11. aa x 2.
Fig. 86. Similar section from the head -region of a Clemmys embryo

13 days old. aa x 2. *""*

Fig. 87. Longitudinal section of a Trionyx embryo 16 days old

(the same embryo as that given m Fig. 2'1). CG x 1. ' '■'>
Fiq. 87a. Transverse section through the allantoic vesicle of an

embryo of the same stage. CC x 1.
Diags. I~VII. Give a summary of the development of the fœtal

membranes in Chelonia.
Diags. I-V. Applicable to both Clemmys and Trionyx.
Diag. I. Corresponds to Fig. I. (PI. I.) and to Fig.

^2 . K. MITSUKURI.

58 (PI. VIL). The head-fold of the embryo is sunk
below the level of the blastoderm and enveloping it is
the proamnion ns yet only slightly developed.

DIag. 11 and II'. Corres])ond to Fig. 2 (PI. I.). The
aioniotic hood prot'ceding backward has covered the
anterior half of the embryo. Its cephalic portion con-
sists of the hypoblast and epiblast ; its dorsal portion
of the epiblast alone. IT represents a cross-section of
the dorsal region. It shows clearly that the mesoblast
has as yet no share whatever in any part of the amnion
(or more properly proamnion).

Dwg. Ill and III'. Correspond to Fig. 3. (PI. I.).
The amniotic hood has extended nearly to the posterior
end of the embryo. The extra-embryonic cœlomic
cavities of two sides are nnited across in the head-region.
The mesoblastic f )lds have also insinuated themselves
into the hitherto solid epiblastic dorsal part of the
amnion (HT.) A partition — tJie sno-amniotic couiicc-
tion — in the median line, however keeps the cu'lomic
cavities of tw(^ sides separate in the dorsal region.

IJiag. IV and IV'. Represent the stage when the posterior
tube of the amnion is fully developed. The sero-
amniotic connection in a cross-section (IV'.) is now
closely invested on each side by the mesoblastic fold,
and is longer than in III'. The mesoblastic fold is
peeling the hypoblast otf the proamnion covering the
head. (IV.).

Diag. V. TAH but a small proximal part of the posterior tube
has now disappeared. The sero-amniotic connecticm is
more developed. The mesoblastic fold has now entirely

ON THE FOETAL MEMBRANES OF OHELONIA. 53

peeled the hypoblast off the proamnion, and the head
is now enclosed in the amniotic cap consisting of the
epiblast and mesohlast. Although these diagrams (III,
IV and V) show the encroachment of the mesoblast on
the proamnion as taking place from before backward,
it in reality takes place mostly from two sides. In
Diags. IV and V, the gradual development of the
allantois is shown.

Diag. VI. Shows the fœtal membranes of Clemmys as com-
pleted.

Diag. VII. Shows the foetal membranes of Trionyx as com-
pleted.

Jour. Sc. Coll. Vol. IV PI. /,

K. Nagahara del.

Jour Sc. Coll. Vol. IV PI. II.

Fig. '.I.

/•Vr,. ;,-.

F 1(1. 14.

Fi,,, i.;

K. Nuciulmni ,V K. Milsukiin dot

J

■i^''

K,-mimnl l'rnl ViiU Amn.

..^'"^

,a.""""-

/f"\

K MilHuliuil <V- li.^i-.julMni ,M.

Jour. Sc. Con. Vol. IV PI: III

Fig. 19

t
•A.

tJour. Sc. Coll. Vol. IV PI. IV

r

%z

AUuntui!

"-<-^.:

T

^K^^Mi

/)-. .V,to,A,„, A J{ X„,,„l„m, ,M.

Jour. Sc Coll. Vol. IV PI.V

VU! Il iiii'iWiiiiiiir*iii»' "

pia;

%.^

riß.SOa

% O

S:».

Ä7. 3« A

o

' '""■■" ^ 3 ^.

VÇ,® °-^

"^^^

,-, "G ^,

Vi^ *«

Jour. Sc. Coll. Vol. I Y PI. VI.

(à

C2>

3

cP-

di>

^

d'^

0

'^y

o *-■•'

6?

d?

0^0.^-

<^>iP

■5 ■'«.-oâ'""' '"l"cr,0^°ee

Jour. Sc. Coll. Vol. IV PI. VU.

Fig. er

Jour. Sc. Coll. Vol. IV. PI. VIII,

Fig. 72.

Jour. Sc. Coll Vol IV PI. IX

'ê

Jour. Sc. Coll Vol IV PI X.

i: V,h:.h„r, .r K .V„„.-,li„ra del.

On the Development of Araneina.

by

Kamakichi Kishinouye, Rigakushi.

Science College, Imperial University.

With Plates XI— XVI.

The following' observations où the development of Araneina were
made in the Zoological Laboratory of the Imperial University during-
the academic session of 1888-9. S<mie of the results I have arrived at
seem to be not without interest, l^efore going further I wish to
express my thanks to my teachers, Dr. K. Mitsukuri and Dr. I. Ijinia
for their kind and valuable advice during my work.

The materials used for the investigation were all collected by
myself during the summer of 1888 in the grounds of the university.
The genera that have been most carefidly studied are Lycosa and
Agalena, while Theridion, Epeira, Dolomedes, Pholcus were more or
less examined for comparison. A species of Lycosa which is very
abundant among grasses breeds constantly from the end of March to
the latter part of September, and carries about the cocoon so that we
are able to obtain its eggs in various later stages with great ease.
It, however, failed to breed in captivity, and for this reason, in the
study of earliest stages recourse wns had to the eggs of a species oî
Agalena which breeds very freely in captivity. The statements made
in the following pages refer to all the species examined unless otherwise
specified.

56

K. KISHINOUYE.

A few words about the methods of investigation may be of use.
Eggs of later stages were killed by heating in water to 70-80°C.,
while segmenting eggs were plunged directly in hot water. Heating
was stopped when the eggs became somewhat opaque and white.
They were then allowed to cool and transferred to 70% alcohol.
After 24 hours, they were examined one by one under a dissecting
microscope and those with unburst egg-membranes were perforated
with the point of a needle to facilitate the penetration of reagents.
They were then hardened in ascending grades of alcohol. I have
always found this method to be excellent for all spider eggs.

Staining was done with alcoholic cochineal, picrocarmine,
alcoholic carmine, or hasmatoxylin. Alcoholic cochineal and picro-
carmine have given best results. It is a remarkable fact that
j)araffin penetrates into eggs stained with picro-carmine more easily
than into those stained with any other reagent. Alcoholic cochineal
proved to be especially good for staining sections on the object glass.
Imbedding for section-cutting was done in paraffin.

Composition of the Freshly Laid Egg.

The egg has two investing membranes, the inner of which is the
vitelline membrane, and the outer the chorion. The external surface
of the latter is covered with a crust of minute spherical granules,
insoluble in alcohol. In a species of Epeira, these granules are com-
paratively large and closely encrust the surface of the eggs, in some
places in two or three layers, making the examination of the inside
almost impossible. They were easily removed by gentle rubbing with
the fingers. In species of other genera examined, the granules were
tolerably crowded in one layer, but being smaller than those of Epeira
did not seriously obstruct the view of the inside.

ON THE DEVELOPMENT OF ARÄ.NEINA. ,57

The conipositicjji of a freshly hiid egg has been tolerably ac-
curately described l)y previous writers, their opinions differing oidy
in some points of details. Tt may be conceived of as a scanty network
of protoplasm in the wide meshes of which yolk granules are im-
bedded. There is always more or less concentration of protoplasm
toward the centre which may be called the centroplasm. An ex-
tremely thin layer of protoplasm is fnind on the external surface of
the egg, directly inside the vitelline membrane and may be dis-
tmguished as the periplasm. The centroplasm and periplasm are no
doubt connected with each other by a scanty protoplasmic network,
although not always ajoparent in sections. The space between the
centroplasm and periplasm is almost entirely taken up by large yolk
granules which are arranged in characteristic radiate columns. In
each column the yolk granules are in several rows, one placed outside
another, and in each row there are generally two granules abreast.
The granules near the centroplasm are much smaller than those
placed more to the outside. In a freshly Jidd egg I was unable to
detect the germinal vesicle in any part. The first segmentation
nucleus appears in the centroplasm a few hours later. In Lycosa, the
so-called yolk-nucleus of tha usual appearance was distinctly seen in
the centroplasm. In Agalena, I could not find it.

The periplasm when seen from the surflxce presents the appear-
ance of being divided into irregular polygonal areas (PI. XI, tig.
1). The cause of this appearance has lieen a point of dispute,
Ludwig* even maintaining that there is no such. That the peri-
plasm is marked out into irregular polygonal arears, there can be jio
doubt. I agree with Locy** in assigning the cause of this marking
to a pressure whicli is exerted on the periplasm and presses it against

* Ludwig— IJ.'ber die Bildung d"s Blastodermes bei den Spinnen, Zeit, für wiss Zool
XXVL

** Looy— Observations on the Development of Agelena naevia, Bull. Mus. Couip. Zool. XII-

58 K. KISHINOUYË.

the peripheral end of the underlying yolk columns, thus causing the
former to receive the impression of the latter. The fact that in
freshly laid eggs the polygonal areas correspond with the underly-
ing groups of yolk granules favours tliis view. I must, however,
differ from Locy as to the cause of this pressure brought to bear on
the periplasm. Locy ascribes it to the contraction of the egg. This
can hardly be, for I could find in no case any trace of contraction, the
eggs being always very closely covered by the two membranes. I
think it much more probable that the polygonal markings are the
effect of the pressure to which the eggs are subjected as they pass
through the narrow oviduct. Locy states moreover that at an early
stage a number of faintly marked areas made their appearance at the
animal pole, ^vhile they could not be detected upon the opposite
hemisphere. I can not corroborate this statement, for I found the
polygonal marking covering the whole surface of the eggs from the
earliest period after l^eing laid. It should be stated that after a while
when segmentation begins, the yolk granules more or less shift tlieir
places ; hence we no longer find the coincidence of polygonal areas
with groups of yolk granules. The polygonal areas do not seem to
change their positions nor do they vary in number after they are
once formed.

From the Segmentation of the Ovum to the Formation
of the Germinal Layers.

According to Ludwig*, who gives a detailed description of the
segmentation of the ovum in Philodromus, the nucleus and the yolk
divide simultaneously first into two, then into four, eight, sixteen,
nnß. so on. Morin** who studied Theridion, IMiolcus. Drassus and

* Ludwig — loc. cit.
** Moïin — Zur Eutwicklungsgeschichte dey Spinnen, Biolog-. Centralbl. VI.

ON THE DEVELOPMENT OP AKANEINA. 59

Lycosa, «ta tes that there is no division of the yolk before there are
formed eight nuclei. In the species studied by myself, the yolk
columns are grouped into as many masses (yolk-pyramids or rosettes)
as there are nuclei, from the time when there are only two of the
latter.

In PI. XII, fig. 8, I have represented a section of an egg in
which there are two nuclei. It will be seen that the yolk is already
evidently divided into two masses or segments. In the lower seg-
ment, the nucleus is distinctly seen. In the upper, the nucleus does
not happen to be in the section, but there is seen the yolk-nucleus
(//. n). The latter does not divide and was often found even in eggs
of the 4 cell stage, always by the side of one of the segmentation
nuclei. The segmentation cavity (^seg. cav) is already present.
The yolk granules immediately adjoining the perinuclear protoplasm
are split up into small particles at whose expense the protoplasm
evidently seems to increase in bulk (PI. XII, figs. 8, 9, 10). This
process of assimilation is no doubt continued during the whole
process of segmentation.

From this stage on, as the nuclei divide, the yolk masses also
divide, assuming characteristic rosette or pyramidal shape (PL XI,
fig. 2). Strictly speaking, the segmentation is not total but syncytial,
as the periplasm remains undivided. Nor is it entirely regular, as
stages with 3, 11, 22, 34, 85 &c. nuclei were found. Nevertheless
the nuclei, after repeated division, are distributed fairly uniformly in
the egg.

As the process of segmentati(3n goes on, the segmentation cavity
which was already present in the 2 -cell stage gradually enlarges so that
iu stages represented in figs. 9 and 10 the centre of the egg is
occupied by a large cavity.

Side by side with their increase by division the nuclei together

60 K. KISHIXOUYE.

with theii' «ur rounding protoplasm gradiuilly travel toward the
periphery of the egg through the yolk pyramids (PI. XII, tigs.
8, 9, 10). When about 30 in number, they all reach the surface.
\Vhen they are almost at the surface, the continuity of the perinuclear
protoplasm with the periplasm by means of pseudopodia-like processes
can be demonstrated on surface views. Figs. 3, 4, PI. XL are tw<^
figures giving such views in which the radially arranged processes of the
perinuclear protoplasm (represented in the figures as dark lines) become
lost in the peri[)lasm whose polygonal markings are still visible.
Soon after such a stage the perinuclear protoplasm and the periplasm
are entirely mixed together forming a nucleated layer at the surface.
So far as my observations go, the nuclei emerge simultaneously all over
the surface of the egg — not, as Locy states, earlier at the animal pole
than at the opposite pole. When there are formed about a hundred
nuclei, this nucleated layer separates itself from the underlying yolk,
and then by the continual division of the nuclei the one-cell layered
blastoderm is established (PI. XIII, fig. 15). Coinciclently the poly-
gonal markings disappear and the egg recedes from the investing
membranes. Probably this is due to the swelling of the membranes
and not to the contraction of the egg.

Whether the yolk still contains nuclei or is entirely free from
nuclei when the blastoderm is established has been a matter of
dispute. In my own sections, I could not at this stage detect an}-
nucleus at all in the yolk, thus confirming the views of Morin in
opposition to lîalfour's.* Yolk granules are, however, still aggre-
o^ated into masses.

The change that comes next is of great importance. The cells of
the blastoderm when it is at first established are of uniform spherical
shapu throughout its extent. While; these cells gradually assume a

* Balfour — Notes ou the Development of the Araneina, Quart. Jouru. Micr. Sei. XX.

ON THE DEVELOPMENT OF ARANEINA. 61

flattened shape over the greatest part of the blastoderm, there is one
spot where the nuclei become conspicuously spherical and mnlti])ly
rapidly. The spot may be distinguished by reflected light as a
round whitish area (PL XI, fig. 5, j)rim. th). It is often a little
depressed at first ; but it soon becomes flat and eventually a little
elevated. Sections through this spot show a large accumulation of
blastodermic cells about seven cells deep (PL XII, fig. 11). I shall
call this thickening the pririiary thickening.

Shortly after this another thickening appears, close to the pri-
mary thickening, on the future median line (PL XI, fig. 6, PI. XII,
fig. 12, sec. tJi). This is also slightly elevated above the general
surface of the blastoderm (PL XII, fig. 13). I shall call it the
secondary thickening. The primary thickening now gradually ex-
tends itself in all directions and forms a whitisli disc-like area of the
blastoderm, the centre of which is thicker than the periphery (PL
XI, fig. 7). This white area is the first trace of the ventral plate.
The primary thickening as it spreads out surrounds and pushes away
the secondary thickening, so that the latter now lies at the margin of
the white area but is further from the centre of the primary thicken-
ing than before (PL XI, fig. 7).

There has been much confusion in regard to the nomenclature
of these two thickenings of the blastoderm. The secondary thicken-
ing corresponds to the primitive cumulus as described by Claparede,*
This appears at least very probable when we compare my fig. 7,
PL XI, with figs. 3 and 4, Planche I, of this author, Balfour was
of the opinion that the primitive cumulus becomes lost in the caudal
thickening. What is called the primitive cumulus by Locy is
undoubtedly the primary thickening above described, while his
" caudal thickening " is the secondary thickening. Morin admitted

* Claparede — Recherches sur I'Kvolutiou des Araignées, Naturk. Verhandel. I.

62

K. KISHINOUTE.

the existence of a blastodermic thickening giving rise to germinal
layers, but denied the identity of it with the primitive cumulus. He
says that the primitive cuniolus is formed after the formation of the
germinal layers and is composed of mesoderm cells. My observations
on Lycosa show that the secondary thickening, or the primitive
cumulus of Claparède, is formed after the formation of the primary
thickening and that both are formed before the distinction of ger-
minal layers is possible. lîoth are accumulations of indifferent cells,
not yet referable to any germinal layer (PI. XII, figs. 11-14). I can
not tell whether the position of the secondary thickening corresponds
to the anterior or to the posterior (.)f the future ventral plate. This
much is certain, that it entirely disappears at the time when the
germinal layers are established.

These two thickenings, the primary and secondary, are of a
great significance, as the germinal layers are established from them,
the primary thickening contributing the largest part in their forma-
tion. In a longitudinal section, these two thickenings are as in figs.
12 and 14, while in a cross section they appear as in fig. 11. From
these figures it is evident that they together form along the median
ventral line of the future embryo a ridge-like thickening which sticks
out into the cavity of the yolk. Cells from the top of this ridge
(the lowest part of the ridge in the figures) proliferate into th*e yolk
and become scattered without any definite arrangement through the
entire yolk. These are the endoderm cells. They become large
by takiijg nourishment from the yolk as they pass through it. The
cell-layer of the ridge nearest the external face of the egg becomes
established as the definite ectoderm. The cells of the ridge
which are left close under the ectoderm form the mesoderm
(PL XIII, fig. 17). They soon spread horizontally below the ecto-
derm. The mesoderm is at first in a single median mass on the

ON THE DEVELOPMENT OE ARANEINA. 63

ventral face and does not extend to the dorsum of the embryo which
is composed of the ectoderm only.

As to the nature of these two thickenings, the primary and
secondary, it is difficuU to state anything definite. The stage in which
the one-cell layered blastoderm is established on the surface of the
egg is to be looked upon as the blastosphere stage. When the ridge
appears in this Ijlastosphere along the line which becomes the median
ventral line of the future embryo and sends otf cells into the yolk
cavity, the whole process must be regarded as a modified form of
invagination and the ridge is to be looked upon as the b]asto[)ore.
Why there should arise two thickenings instead of one remains inex-
plicable to me. The primary thickening is without doubt the remnant
of the blastopore. Whether the secondary is to be looked upon as a
part of the same, I cannot decide.

From the Formation of the Germinal Layers to the
Reversion of the Embryo.

After the establishment of the ventral plate, its anterior part
becomes marked off as the cephalic, and its posterior part as the caudal
lobe, and the middle region between the two lobes is divided by trans-
verse rido'es into seo-ments. The least number of se^'ments observed
between the cephalic and caudal lobes was five. The foremost of
these corresponds to the segment which bears the pedipalpi and the
four following are the thoracic segments, each of which subsequently
produces a pair of ambulatory appendages. The segment which is to
bear the chelicerœ is soon after cut off from the cephalic lobe and the
abdominal segments are gradually cut off from the caudal lobe, the
process proceeding posteriorly, until there are formed eight abdominal
segments (Lycosa).

64 K. KISHINOUYE.

In this process of segmenting the mesoderm of the ventral plate
shares (PL XIII, fig. 16), and is divided into as many parts as there
are segments in the body (jf the embryo. Moreover it divides itself
into two longitudinal bands at the median line except at the cephalic
and caudal lobes. Thus there is formed in each segment a pair of
mesodermic plates. After a while, eacli of these paired mesodermic
plates produces a cavity — the cœlom — apparently by its splitting into
two layers (PI. XIII, fig. 18). The outer of the two layers is the
somatopleure, and the inner the splanchnopleure. The cœlom there-
fore consists at this time of a number of paired cavities (PI. XIA ,
fig. 22), which are separated from one another. Cœlomic cavities in
the cephalic and caudal lobes appear only later on.

Shortly after the formation of the cœlom, a pair of protuberances
appear on each segment. They are the first traces of the appendages
(PI. XIV, fig, 23, th. app). The order of their appearance cor-
responds to the order of appearance of the segments to which thej-
belong. The appendages are formed on segments of the chelicera^
and pedipalpi in all the thoracic and the second, third, fourth, and
fifth abdominal segments (PL XIII, fig. 20, PL XIV, fig. 22).
The cephalothoracic appendages are formed at the lateral ends of the
segments, while the abdominal appendages are formed nearer the
median line (PL XIII, fig. 20). Tlie abdominal appendages are little
round protuberances, and do not elongate as rapidly as other
appendages. The first abdominal segment bears no appendages, as
Schimkewitch* has correctly observed. This segment is gradually
aborted, and is not distinctly visible at tlie time of the reversion of
the embryo. The cœlomic cavities of each segment extend into
the appendages.

The foundations of the nervous system are laid soon after the

* Schimkewitcli — Etude sui- le Development des Araignées, Arch, de Biolog. VI.

ON THE DEVELOPMENT OP ARANEINA. 65

establishment of the ventral plate during this period. The ectoderm
of the cephalic lobe is very mach thickened as shewn in figs. 22 and
23. This process of thickening proceeds backwards as two longitudinal
bands, one on each side of the body, along the inner side of the attach-
ment of the appendages in the thoracic and abdominal segments, finally
meetino" each other in the caudal lobe. These two l^ands are the first
rudiments of the ventral nerve chain. Thus it is continuous from
the first with the cephalic thickening above mentioned which becomes
the brain, as in the case of scorpions observed by Kowalevsky and
Schulofin.* This is not in accordance with the view of some authors

a

who maintain that the brain and the ventral nerve cords are formed
independently of each other. The cells composing the ventral cords
aggregate in each segment and give rise to the ganglia.

The cephalic thickening of the ectoderm is now divided into two
semicircular lobes (PL XIII, figs. 20, 21). Near the front edge of
these lobes, there is formed on each side a semicircular groove (sem.
gr.). This paired groove which is cut ofi' from tlie ectoderm is the
chief origin of the brain. I'riice** compares it with the amniotic
f )ld of insects ; but the comparison is certainly not justifiable.
Kowalevsky and »Schulgin found that in Scorpions the ectodermic
invagination comparable to the amniotic fold of insects is distinct from
and formed earlier than the semicircular groove, which is no doubt
homologous with the similar groove of the spider, as it also gives
rise to the brain. Sections of the semicircular groove are represented
in fig. 23, PI. XIY.

Besides the semicircular grooves, there is a pair of small ecto-
dermic invaginations in the posterior part of the head near the
outer bor.ler (PI. XIII, fig. 20, PL XIV, fig. 23, lat. v). So far as I

* Kow.ilevsky and Sohulgin — Zur Entwicklungsgeschichte des Scorpions, Biolog, Cen-
tralbl. VI.

** Bruce — On Insects and Arachnids.

ee

K. KISHTNOUYE.

know these invaginations have been till now entirely overlooked.
Til fig. 19, PI. XXI, of Balfoar's work, I find one of these in-
vaginati(jns represented ; but he gives no informati(3n about it. It
is globular in form ; henceforward I shall call it the lateral vesicle.
The lateral vesicles, which are also gradually constricted off frc^m the
ectoderm, go to form a part of the brain (PL XV, figs. 44-46).

The stomodœum is formed as an ectodermic invnirination at the
anterior margin of the cephalic region (PI. XIII, fig. 20, PI. XIV,
figs. 24, 25). At this stage it is easy to see that all the appendages
are postoral in origin.

Late in this stage a number of large cells appear at the dorsal
part of the embryo. They are never found in the ventral plate.
They are very easily recognised by their large size and the peri-
pherally situated nuclei, their central porti<3n being filled with fat
(PI. XV, figs. 40, 41,/. c). Undoubtedly they are nourishing cells,
wandering everywhere, and some of them are changed into blood
corpuscles. They were called by Balfour the secondary mesoderm,
l)y Schimke witch the secondary endoderm, and by Locy the endo-
derm cells. These three authors ascribed the oriu'in of tliese fit cells
to the cells in the yolk, whereas according to Morin they are formed
in Pholcus from dispersed mesoderm cells originally composing the
so-called primitive cumulus,* and in Theridion which wants the
cumulus probably from cells of the mesodermal somites. Schimke-
witch, Locy, and Morin observed that these cells become blood
corpuscles. For my own part, I am inclined to agree with Balfour,
Schimkewitch, and Locy and to derive them from the endoderm.
For in the first place, they are found immediately above the yolk,
arid in some cases between yolk granules presenting the appearance as

* Morin states, what I have before referred to, that the primitive cumulus is formed after
tlio foruiation of germinal layers, and consists of mesoderm cells.

ON THE DEVELOPMENT OP ARANEINA. 67

if they have just emcrg'cd from the yolk. In the second place, their
nuclei ao;ree in their loroe size with those found in the yolk.

At the end of this stage the mesoderm in the caudal lobe is
faintly divided into two layers, between which an unpaired cavity
makes its appearance (PI. XIV, û^. 24). In the cephalic lobe also
the mesoderm is faintly divided into two layers on each side (PI.
XIV, fig-. 23), enclosing the rudiments of the cœlomic cavities. It
is still undivided in the median line. The cœlomic cavities in the
thorax secondarily fuse together into a single cavity. They remain,
however, quite distinct in the abdominal region.

The Period of the Reversion of the Embryo.

The stage in which the reversion of the embryo occurs is as
difficult to study as it is important, since many organs arise at the
same time. At the end of the last stage, the ventral plate had reached
the maximum limit of dorsal flexure, the cephalic and the anal lobes
almost toucl.dng each other (PI. XIV, figs. 24, 25). As Balfour
states, the reversion of the embryo is due to the expansion of the
dorsum ; and the expansion of the dorsum is due to the horizontal
increase of cells which compose that part. The head and the tail are
pushed away from one another further and further. As the dorsum
is very rapidly expanding and the cells are pressed for room, a groove
is produced immediately behind the tail lobe to increase the surface
of the dorsum, and the tail lobe then stands out as a conical process
(ri. XIV, figs. 26-29). The cœlomic cavities belonging to the
segment in front of the tail lobe being pressed from the dorsal side
by the increase of cells in the dorsum are compressed horizontally and
pushed into the conical tail process, enveloping the unpaired cœlomic
cavity of that process from the dorsal side. The caudal lobe stands

68

K. KISHTNOUYE.

out gradually more and more prominent, until the stage represented
in fig. 27 (surface view, fig. 21) is reached. After this, the tail
process gradually shortens (figs. 28, 29) until after a while there is
no tail projecting from the general body surface (fig. 32).

At about the same time with the increase of cells of the dorsum,
the two nerve cords begin to diverge from each other. They are
most widely separated from each other at the anterior part of the
abdomen and gradually approach each other anteriorly and posteriorly
until they meet in the cephalic and tail lobes (fig. 21). Their
divergence together with the expansion of the dorsum makes the
embryo assume the ventral flexure.

The cœlomic cavity of the caudal lobe now becomes gradually
conspicuous. This unpaired cavity is transformed into the so-called
stercoral pocket (Rectalblase, Kloake) of the adult spider. Hence
the stercoral pocket does not arise from the swelling of the internal
end of the proctodaßum, as has been supposed by other authors. This
organ is purely mesodermic in origin and nothing more than a
remnant of cœlomic cavities. This may be understood by examining
figs. 24-32, PI. XIV. From these figures it will be seen that the
proctodseum is formed in the caudal lobe later than the stercoral
pocket.

The fact that any part of the adult alimentary canal should be
derived from the cœlom seemed to me so remarkable that I have
repeatedly examined my series of sections and am convinced of the
correctness of the observation. I do not know how to interpret this
fact unless it be that the stercoral pocket is a part of the primitive
excretory system — a supposition which is strengthened by its peculiar
relation to the remaining part of the digestive tube (PI. XVf, fig.
55) and by the fact that the Malpighian tubes open into it.

At this period the mesodermic somites and the ganglia of the

ON THE DEVELOPMENT OF ARANEINA. 69

anal lobe and of the four appendage-bearing abdominal segments
have attained their utmost development. The first abdominal seg-
ment and those between the fifth and the last abdominal segments
are aborted.

The mesodermic somites which are produced at first in the
ventral plate now grow on dorsalwards and meet at the dorsal
median line (PL XV, figs. 40-43). They first meet at their dorsal
part, enclosing some of the large fat cells and their derivatives be-
tween them. The ventral part fuses later. Thus the dorsal circulatory
tube is formed, the wall of which is produced from the mesoderm,
while the blood corpuscles are produced from large fat cells (endo-
dermic in origin). I am inclined to believe that both the aorta and
the so-called heart are formed as stated above and not separately as
many authors believe. The fusion of the mesodermic somites to
form the dorsal vessel does not take place throughout the entire
length, as there are left paired lateral slits between each two consecu-
tive somites. The blood aerated at the lung-book returns to the
heart through these lateral slits. These slits shut and open as the
heart beats. They are found in the abdomen only.

In the basal part of the first abdominal appendage of each side,
there arises an ectodermic invagination whose opening faces away
from the median line. It is neither deep nor spacious but is a little
pocket-like invagination. This is the beginning of the lung-book.
The development of this organ, briefly stated, is as follows : Of the
wall of the invaginated pocket, that which faces the distal end of the
appendage is much thicker than the opposite wall, filling the interior
of the appendage. The cells composing it become after a while
arranged in parallel rows (figs. 34 and 47). Each two of these
parallel rows adhering together produce the lamellae of the lung-
book. The external epithelium of the appendage which cover these

^0 K. KISHIXOUYË.

■

lîimellaî becomes the operculum of the kmg-book after it is depressed
in height. Judging from figures (figs. LXXIX and LXXIX')
given in " On Insects and Arachnids," Bruce seems to have mistaken
the caudal prominence of the early period of this stage (see my figs.
24-28) as the operculum of the lung-book. According to him tlie
abdominal nppendage is invaginated to form the lung-book, but as
we have seen, it is not so. Locy has correctly described the forma-
tion of the lung-book lamella3. He says that the lungs arise from
infoldings ; but he is silent about the place where these infoldings
arise.

In the basal |)art of the second abdominal a[)[)endage on the
interior side, another ectodermic invagination is produced. It as-
sumes the shape of a deeply invaginated tube and remains in this
condition till after the time of hatching. The appendage itself is not
invaginated and becomes from this time gradually shorter.

It is very probable that the lung-books were derived from the
gills of some aquatic arthropodous animals such as Limulus ; for
the lunof-books are nothinsf more than the lamellar branchia3 of

O a

Limulus sunk beneath the body surface. The tubular trachea may
afterwards have been derived from the lunsr-books. The branchin 1
lamella} of Limulus are formed as outgrowths <jf the ectoderm at the
lower (posterior) surface of abdominal appendages, and those of spiders
are also produced really in the lower surface of the fii'st abdominal
appendage (in the dipneumonous spider). Hence I tliink that the
spider with two pairs of lung-books is the most primitive one, and
the one with one pair of lung-books and the other pair transformed
into the tubular tracheai is more primitive than the spider with only
one pair of lung-books. I cannot agree with the view of some
authors who maintain that the luno'-book is derived from a chister ol
tracheœ.

ON THE DEVELOPMENT OF ARANEINÂ. 7l

The third and fourth pairs of the ahdoiniual appendage are
modified into spinning niammillœ (PL XV, fig. 34). At the distal
end of each of tliese appendages a solid proliferation (sp. gl) of ecto-
derniic cells is for?ned. This becomes the spinning gland. Spiders
have generally three pairs of spinning mammillas ; two of which
are modified abdominal appendages, while the remaining one is added
very late, after the hatching of the embryo. The primitive spider
must have had only two pairs of spinning mammiilœ. Some tetrap-
neumonous spiders have only two pairs.

The two semicircular halves of the cephalic lobe, between which
there is at first a deep median notch (PI. XIII, fig. 20), now fuse
with each other at the median line above the stomoda^um, so that tlie
notch becomes much shallower (fig. 21). The grooves formed
along their anterior margin during the preceding stage separate from
the ectoderm beginning from their external end and sink down
beneath the body surface. They are cut off from the ectoderm latest
at the hindermost j)ai'ts of their inner limbs (PL XVI, fio-. 48).
The lumina in the two separated semicircular grooves come to com-
municate with each other at the anterior median part (PL XV, fio-.
45).

At the last point of separation there is left a shallow in\a<'ina-
tion or rather sac on the sui-face. The invagination is ])aired. The
openings of these sacs are directed towards the mouth of the embryi^,
and the invaginations are directed anteriorly. They are the first
traces of the posterior median eyes f^ee below) or the 'Hauptaugen'
of Bertkau* (PL XV, figs. 41-46, 48, P. M. E.). The anterior wall
of the sac is thicker than the posterior, the former being two t(j
several cells deep, the latter only one cell deep. The formation of

* Bertlvau— Beiträge zur Kenatniss der Siuuesoi'y;iue dur Spiuneu. Arch. f. Mik. Anat
XXVII.

72

K. KISHINOUYE.

the posterior median eyes in connection with the brain in .spiders is
quite analogous to the simihir process in scorpions as observed by
Kowalevsky and Schulgin. This interesting relation was not observ-
ed by Locy who studied the spider, or by Parker* who studied
the scorpion.

Hitherto these eyes were called the anterior median eyes ; but
morphologically speaking, this nomenclature is not correct. For all
the eyes of spiders are formed in reality in the Neutral jjhite, never in
the dorsum, and gain their apparently dorsal position in later stages
only by the bending upward of the ventral plate. Hence, in this last
position the eyes that composed the ])osteri(jr row in the ventral
position come to occupy the anterior position, while those that formed
the anterior row in the ventral position are thrust further backward
by the curving upw\ard of the ventral plate and thus become the
apparent posterior row. Hence those I called the posterior median
eyes are in the apparent anterior row of the adult.

The three remaining pairs of eyes are formed later than the
posterior median pair and in a different manner. Their first traces
are the local thickenings of the ectoderm of the cephalic region
Anterior lateral eyes (A. L. E.) ai)pear above the lateral vesicle (PI.
XV, fig. 46).

At this time the lateral vesicles are c<jnipletely cut off from the
general ectoderm (PI. XV, figs. 44, 46). Their walls are thick and
their lumen is cons[)iGuous. In development and position they ^'ery
much resemble the eyes of Peripatus.

The chelicera3 are now two-segmented. They have shifted their

position a little anteriorly and have approached toward the median

line (PI. XIII, fig. 21). Their ganglia are placed at the sides of

the stomodteum and form tlie commissural part between the supra-

* Parker— The Eyes in Scorpions, Bull. Mus. Comp. Zool. XIII.

ON THE DEVELOPMENT OF ARANEINA. 73

and infrfi-œsophageal ganglia. They are in contact with each other
at the anterior part. The basal joint of the pedipalpi is very broad,
the maxillary part being- easily distinguished. The ganglia of the
pedipalpi and of the succeeding four thoracic segments are well deve-
loped and are in close contact with each other, thus forming the
large snb-œsophageal ganglion. The ganglia belonging to the abdo-
minal segments are also well developed. . :

The stomodœum elongates itself obliquely upwards and is sur-
rounded externally by the well developed upper and lower lips (PI.
XIII, figs. 19-21 ; PI. XIV, figs. 24-26). The ectoderm forming
the wall of the stomodieal invagination is thick.

The ectoderm of the ventral part of the anal lobe is conspicuous-
ly thicker than that of the dorsal part, being continuous with the
two ventral l3ands. At the beginning of the reversion, it is
uniformly two or three cells deep (PL XIV, fig. 27); but when
the reversion is fairlv advanced, so that the elono-ated anal lobe beo'ins
to become short again, the cells in the middle part of it are elongated
and thei'e they are only one cell deep (fig. 28). At this part an in-
vagination takes place (fig. 29). Fnjm this stage the ectoderm of
the ventruni of the anal lobe, placed anterior to the invagination be-
comes two or three times thicker than the posterior part, and is dif-
ferentiated to form the anal ganglia (figs. 29-32, (r). The invagination
is the protodaeum. It is very shallow and small, aud its bottom is
in direct C(3ntact with the wall of the stercoral pocket. The wall of
the prootodasimi is thinner than that of the stomodsenm. It is
remarkable that the proctodîeum is not formed at the extreme hind
end of the ventral plate but somewhat in front of it directly behind
the- anal- ganglia, and that both the stomodœum and the proctodaemn
are produced at the two extremities of the nervous system simulta-
neously with the development of the latter near them. The portion

74 IC. KTSHINOÜTE.

of the ventrnm, posterior to the pi-octoda?iim, gradmilly thins oif,
and after the process of reversion is completed it can not he dis-
tinguished from the dorsum (PI. XIV, figs. 31, 31).

The posterior part of the mesenteron is formed hy an accumla-
tion of endoderm cells at the anterior ventral part of the stercoral
pocket. It is a wide open funnel-shaped tuhe, resting ahove the
mesoderm (fig. 32, Post, mrsent).

The stercoral pocket produces paired diverticuhi fr(^m its lateral
sides (fig. 33). At first, I was inclined to think that these diverti-
cula become the Malpighian tubes, as these tubes were formerly
th<mght to arise as a pair of outgrowths from the stercoral pocket.
But I found that these diverticula give rise to no definite structure
in the adult, and that the Malpighian tubes arise in a ditferent way,
as will be explained further on.

At this stage a very important organ is produced, which has
been almost entirely neglected by embryologists. I mean the coxal
gland, Avhich is f)rmed from an ectodermic invagination at the in-
ternal posterior base of the coxal joint of the first ambulatory
appendage (PI. XV, fig. 38, Co. gl). The invagination opens into
the cœlomic cavity (figs. 35, 36). Its development is traced farther
in the next stage.

After the formation of the circulatory system the cœlomic
cavities atrophy, except the one of the anal lobe forming the stercoral
pocket, and some part of the thoracic ones in connection with the coxal
gland. The so-called body cavity of the adult is not the remnant
of the cœlomic cavity ; but it is a secondarily produced blood-space.
T'he mesodermic cells which formed the wall of these cavities form
The covering of the nervous system, the alimentary canal and other
organs.

Some mesodermic cells at the base of the cephalothoracic appen-

ON THE DEVELOPMENT OF ARANEINA. 75

dages become rounded in outline (PI. XV. figs. 35, 36). They are
easily distinguished from the fat eells by their centrally located
nuclei, and from other cells by their well-defined spherical form and
slightly stninable protoplasm. They appear first in the chelicenv,
then in the pedipalpi, nnd so on gradually backwards. These cells
liave no relation whatever with the coxal gland uov with the poison
gland. Their function is unknown. It seems to me that Locy has
mistaken these cells at the base of the chelicera^ for the first rudi-
ments of the ])oison gland. He says that these cells are probably
derived from an infolding of the ectoderm.

From the End of the Reversion to the Hatching
of the Embryo.

This stage is characterized by the api)earance of a constriction
separatiiig the céphalothorax from the abdomen. The yolk in the
ventral part of the abdomen is absorbed, so that the abdominal
appendages of both sides appr("»ach each other at the median line.

The senncircular grooves of the cephalic lobes formed in the
preceeding stage are no longer grooves, nor semicircular in form.
Now they are completely constricted oif from the general ectoderm, and
are consequently tubes. Their inner Yunhs approach each other in the
median line and they form as a whole a T-shaped body (PI. XV, fig.
45). The lumina of the two tubes communicate with one another at
the anterior median part. They as well as the lumen of the lateral
vesicle begin to atrophy by the thickening of their walls and finally
disappear. At the same time the transverse bar of the T-shaped
mass becomes curved on each side to a peculiar shape shewn in
profile in fig. 45 a. This and the disappearance of the lumen change
the brain into a compactly packed mass, instead of having its various

76

K. KTSHINOYUE.

parts stanrling apart as heretofore. The transverse bar (fiof. 44, a) of
the T-shaped brain is separated from the median stem just behind
the point where the Inmina of the two sides communicate with each
other, while the median stem is in its turn transversely divided into
two segments (Fig. 44. b, e). Thus the spider's brain consists of
three segments, as l^ntten* claims. These three segments may be
called the transverse dorsal (Fig. 44, a), the anterior ver.tical (Fig. 44,
/>), and the posterior ventral section (Fig. 44, c). The lateral vesicles
arein the level ofthe third segment. From his description. Patten seems
to mean that in scorpions and spiders the three segments (^f the brain
are formed from three separate invaginations ; but I cannot cor-
rol)ornte this statement. Moreo^er he says that the anterior median
eyes (my posterior median eyes) belong to the second segment, while
the three remaining pairs belong to the third segment. Supposing
tlint his second segment is anterior to the third segment, 1 cannot
corroborate this statement either, as accoixling to my own ol)servati<^ns
all the eyes belong to the third segment. It seems to me impossible
tliat the posterior eyes should arise in a segment anterior to that in
whicli the anterior eyes are produced.

The opening of the sacs of the posterior median eyes becc^mes
gradually smaller and is finally closed (PI. XYI, fig. 49). The
anterior wall ofthe sac becomes enormously thick and obliterates its
lumen. The ectodermic cells which lie ii])on the sac elongate and
form the vitreous body (tigs. 49, 54, ivV). The anterior wall of the
sac forms the retinal part (fig. 49, PC). The retinal cells elongate
anteriorly. The anterior surface of the anterior wall of the posterior
median eyes, is morphologically the inner side of the ectoderm
though it faces externally. The lens is formed by a local thickening
of rhe cuticula, which is secreted from the epithelium at this stage

* Patten — Segmental Sense-organs of Arthropods, Journ. of Morph. II. ^

ox THE DEVELOPMENT 01" ARANEIXA. 77

(PJ. XVI, fig. 49. L). The nerve cLjes not enter the p<jstenor median
eyes even a few days attei' the hatching of the embryo. Probably
the nerve is sent out from tlie retina from the anterior (morphoJooi-
calJy inner) surface of it, as this is the case in the aduU. The deve-
lopment of the posterior median e^'es is comparatively slow. They
are homologous with the median eyes of scor|)ions, as the develop-
ment is quite the same.

The three remaining pairs of the eyes or ' Xebenaugen ' of
Bertkau* are formed later than the posterior metlian eyes ; but their
development is completed earlier. They arise from ring- like depres-
sions of the ectoderm (fig. 50). The walls surrounding these
depressions grow over them and finally meet (fig. 51). The spot
where the walls meet is one-cell layered. This spot gi-adually ex-
tends to a certain extent and forms the vitreous body which is
characterized by elongated cells (figs. 51, 54, vit). The growth of the
walls of the depressions is not uniform in every direction and there-
fore the point of closure may not correspond with the centre of de-
pression. Thus the ' Nebenaugen ' are also formed from ectodermic
sacs ; but these sacs are diiferent from the sacs of the posterior median
eyes. While it is the anterior wall of the sac that becomes the retina
in the posterior median eyes, it is in the case of the " Nebenaiigen'
the posterior wall of the sac, which, forming a central elevated })ortion
thicker than the anterior wall and siu'rounded by a ring-like depres-
sion, gives rise to the retina (figs. 50-54, R). Also retinal cells
elongate posteriorly instead of anteriurlij, as in the posterior median
eyes, and form nerve fibres (figs. 50, 51, A^). These nerve fibres are
subsequently connected with the fibrous portion of the brain. The
retinal portion is cut ofi:' from the general ectoderm at about tlie time

* Burtkau, lue. cit.

78 K. KISHINOUYE.

of hatching, mid at the same, time becomes concave (tig. 54), instead
of being convex as heretofore (ûg. 51).

In the ' Nebenaugen ' — but not in the posterior median eyes —
there are f(n-med transverse liars and a circumferential ling (^g'i.
51- 54, tai)) of chitinous natiu-e, posterior to the retinal cells and
secreted by these cells. These chitinous bodies (the tapetum) are
transparent and lightly yellowish by transmitted light and silvery
olitterinsf bv reflected liuht. Tlie lens is formed in a similar manner
as in the case of the posterior median eyes. The tapetum and the
lens are eqntdly secretion products of the ectoderm and both of them
are chitinous in nature, liut they are not homologous. The former
is produced at the proximal end of the ectoderm cells, while the latter
is formed at the distal end. : .

I know of only two authors who have studied the development
of the spider's eyes b)^ recent modes of investigation. They are
Locy and Schimkewitch. The results obtained by these authors are
not entirely satisfactory. Locy could not find the diHerence in the
mode of development between the 1 wo différent types of eye. He
says that the 'Nebenaugen' originate in substantially the same way
as the anterior median eyes (my posterior median eyes). Moreover
he states that the development of the eyes begins by a local thickening
of the hypodermis and a backward directed infolding which inverts
the thickened region. Schimkewitch says only that the retinal part
of the eyes originate from a pyriform eidargement from the brain,
upon which the ectoderm invaginates in the form of a ring.

Patten recently gave a short account of the spider's eyes in an
article entitled the " Segmental Sense-organ of Arthropods " in the
Journal of Morphology, Vol. IT. ; but his account differs from mine
in many jioints, as I have already mentioned. He says that there
are segmental sense-organs, homolog(jus with the eyes, at the base

ON THE DEVELOPMENT OF AKANEIXA. 7i)

of the legs. Unfortunately I could not find any trace of such an
organ, though I carefully searched after it.

The development of the pigment begins from the cephalic region
backwards, after the ditferentiation of the vitreous Ijody (fig. 51). In
the case of the ' Nebenaiigen ' the pigment is first priKluced in those
cells which form a kind of a cup around the retinal portion (figs.
51-53), and it seems to me most probable that these cells wander in
to the retinal portion, first jtmong the nerve fibres beneath the
ta[)etimi (fig. 53), then among those above thetapetum (fig. 54). In
the case of the posterior median eyes, however, the pigment is pro-
duced from the beginning in retinal cells, below the vitrecnis body.

As we have already seen, all the eyes of the spider are formed in
the ventral plate and near its anterior margin.

The concentration of the nervous system towards the céphalo-
thorax goes on further in this stage than in the previous stage. In
the thoracic region the two lateral oanplionic chains are united into
one and form the subœsophagcal ganglion. The inner portion of the
ganglion becomes finely fibrous. The abdominal ganglia gradually
atrophy and attach themselves to the posterior end of the subœso-
phageal ganglion. At this stage the whole nervous system is com-
pletely cut off from the ectoderm.

The stonioda'uni has devchjpcd very mucli. After chjiigating
itself obliquely upwards, it takes the horizontal backward direction and
reaches to about the segment of the fourth ambulatory a[)pendage. It
is lined with a cuticular covering which is continuous with the cuticula
<jf the general body surface. In the pliarynx, the cuticular lining is
tliick and transversely ridged. The ridges rim parallel with each other
and appear in the sagittal section like teeth, the pointed edge turning
dorsalwards. The wall of the stomodieum is very thick. The stomo-
dieum gives rise to the pharynx, the œsophagiis, and the st<jmach.

80

K. KISHINOUTE.

Early in this stage some endoderm cells ucciimiihite at the
posterior end of the stomach and form the anterior part of the me-
senteron. These cells are arranged as a funnel-shaped tiihe \vid(;
open posteriorly. The posterior funnel hîis united with tlic wall of
the stercoral pocket at its hind end (tig. 55). The anterior and the
posterior funnels of the mesenteron do not at this stage unite with
each other.

Locy says that on each side of the stomach are given off ca^ca,
which extend into the hases of the liml)s. lie adds that the cellular
elements comp(3sing the walls of these tubes are flattened ; hut he
gives no account concerning the time of their appearance Though I
have carefully examined embrycjs of all the stages, I could not And
such tubes.

The proctodajum is lined with a cuticular covering as the stomo-
daäum ; but the stercoral pocket has no such covering. This fact
confirms my observation that the stercoral [)ocket is not a portion of
the proctodœum. The communication between them is formed at
this stage. The communicating canal is very narrow. In the last
stage, the stercoral pocket was somewhat globular in shape (PI. XIV,
fig. 32), now it is elongated anteriorly and is oblong (PI. XVI.
fig. 55). Its lateral diverticula have disappeared.

I could not make out the development of the Malpighian tubes
satisfactorily ; but I am certain that they do not originate from the
ectoderm. Also it is certain that they are not outgrowths from the
stercoral pocket. It seems to me probal)le that they originate from
mesodermic cells belonging to the abdominal S(jmites in front of the
anal lobe. At this stage they are solid jjaired cords of cells (iig. 55,
' j\[alp. t) extending from the anterior end of tlie second abdominal
segment to the sides of the confluent point of the posterior mesenteron
with the stercoral pocket.

ox THE DEVELOPMENT OF ARANETXA. 81

The mesodermic cells of the coxal gland, which was formed in
the preceeding stage, are very much differentiaied from the ecto-
dermic cells of it. They are the glandulär cells, their size becoming-
large and their protoplasm granular and unstainal)le (PI. XV, tig.
37). The ectodermic cells form the duct.

At the distal end of the chelicera^ a solid growth inward of
ectodermi<* cells takes place. These cells are surrounded by meso-
d(H'mic cells. The distal half of the former becomes the s^landular
portion, and its proximal half the duct, of the poison gland, while the
mesodermic cells torm the muscular wall of the gland (hg. 39).

In this stage four paii'ed transverse septa are formed between the
four appendage-bearing segments of the abdomen by the sinking
of the mesoderm into the yolk. A median unpaired se|)tum, similarly
formed, also stretches forward fro]n the posterior end. These septa
are foi'med after the disappearance of the cœlomic cavités in the
al)domen. In fi,<r. 34, PI, XV., two anterior septa are represented.
The first pair of septa probably give rise to the generative organ,
and all or some of the others to the so-called liver.

After undergoing one or two moults, the embryo hatches. The
body of the embryo is covered with cuticular hairs. At the end of
the pedepalpi and the foui* ambulatory appendages, the claws are pro-
duced, and at the end of the chelicerœ the poison fangs, by thickenings
of the cuticula.

Summary.

(1) The polygonal areas are on the periplasm, and are pro-
bably formed when the eggs pass through the oviduct.

(2) In the process of segmentation the yolk and the nucleus
are divided at the same time. The segmentation is syncytial.

82

K. KISHINOUYE.

(3) The yolk nucleus is found in segmenting eggs on to the
four-cell stnoe.

(4) After the segmentation all the nuclei are found only at
tlic surface of the egg, and none of them remain in (he volk.

(5) The primary blastodermic' thickening may he considered
as a modified gastrean mouth, the formation of whir-h was obstructed
by the abundance of yolk.

(6) The secondary blastodermic thickening or 'primitive
cumulus ' of Claparede plays a secondary part in the formation of the
germi n al 1 ay er s .

(7) The brain and the ventral nerve cords are formed as a
continuous ectodermic thickening.

(8) All the appendages are postoral in origin.

(9) The first abdominal segment bears no appendages.

(10) The large fat cells are derived from the endoderm. They
ff^rni blood corpuscles.

(1 1) Ati invagination at the posterior l)ase of the first abdominrd
appendage gives rise to the lung-book. A similar invagination at the
base of the second gives I'ise to a tube — abortive trachea.

(12) The unpaired cœlomic cavity, belonging to the anal lobe,
changes to the so-called stercoral pocket. Prolwbly it is excretory
in function, not a part of the alimentary canal.

(13) The dorsal circulatory vessel is formed by the fusion of
the mesoblastic somites at the dorsal median line.

(14) The so-called body cavity of the adult animal is not the
descendant of the cœlomic cavity, but it is a secondarily formed

space.

(15) The brain is composed of the semicircular grooves and the
lateral vesicles cut off from the ectoderm. Later it is divided into
three segments.

ox 'I'PTE DEVELOPMENT OF ARANEINA. 88

(16) The development of the posterior median eyes is connected
with that of the hrain. Their development is quite different from
that of the other eyes ; but all the eyes are dermal in origin, not
neural. And the nerves of the eyes enter always from the inner
ends of the ectoderm cells.

(17) A pnii- of coxal glands opens at the base of the third
appendage. The glandulai- portion of it is formed from a portion of
the cœlom, while its duct is formed frc^m an ectodermic invagination.

(18) The alimentary canal of the spider is formed from the
ectoderm and the endoderm. The pharynx, the œsophagns, the
stomach, and the anus ai-e produced from the former, and the intes-
tine from the la t fer.

(19) Tlie Malpighian tubes are produced neither from the ecto-
derm nor from the stercoral pocket. They are mesodermic in origin.

<.

84

K. KTSHTNOUYE.

Explanation of Plates.

The figures are all exact representations of preparations, the
outlines, the nuclei, and other details being drawn faithfully by my-
self with the use of the camera lucida, and they are not diagramatic,
except in the case of a few figures expressly so stated.

List of References.

(t, first segment of brain.

((hd. app., abdominal a]i]iendao-o.

a. I., anal lobe.

A. L. E., anterior lateral eye.

ant. mesent., anterior portion of me

senteron.
fe, second segment of brain,

c, third »» >' >>
ceph. /., cephalic lobe.
eh., chelicerae.

ch. g., cheliceral gano-lion.
eo. gl., coxal gland.
cut., cuticula.

d, dorsal side.
dor., dorsum.
ect., ectoderm.
end., endoderm.

/. c, fat cell.
G, ganglion.

inv,, invagination of lung-book.
L, lens.
lat. v., lateral vesicle.

Malp. t., Malpio'hian tulto.

mes., mesoderm.

N, nerve.

pedip., pedipalpi.

P. M. E., ])osterior mediaii eye.

post, mesent., posterior portion of

mesent eron.
prim, th., primary thickening.
proct., proctodasum.
B, retina.

sec. th., secondary thickening-.
seg. cav., segmentation cavity.
sein, gr., semicircular groove.
sp. gl., spinning gland.
sterc. p., stercoral pocket.
stom., stomodaeum.
tci])., tapetum.

th. app., thoracic appendage.
V, ventral side.
vit., vitreous body.
y. n., yolk nucleus.

ox THE DEVELOPMENT OF AKANEINA. 85

Explanation of Figures.

Fiu'. 1. All iiiiseuiiieiited eyi»-, sliowiriü' the polvofoiîil nreas
above yolk graimle.s. (^Liictisa). 2 B (Zei>^s).

Fig. '2. A segnientatioii egg oï the four-cell stage, showiiio-
the rosette-like yolk pyramids, {hijcosa^. 2 B.

Fig. o. A segmentation <^'g^^, shewing the union of tlie poly-
gonal areas with tlie segmentation nuclei. (^Ltjcosa). 2 B.

F^ig. 4. The same as above, Ijut of a little later stage. This
shows that the yolk pyramids become very small and that the poly-
gonal areas do not correspond in p(jsition with the yolk granules.
(Lijcosa). 2 B.

Fig. 5. An tgg shewing the primary thickening of the blasto-
derm. (^Lycusa). 2 A..

Fig. 6. An eiig- ha^'in"; the secondary thickeniusf of the blasto-
derm, produced at the margin of the primary thickening. (Liicoaa).
2 A.

Fig. 7. An egg in which the primary thickening has extended
enormously, and the secondary thickening is at the margin of the
[)rimary one as before. (Lijcosa). 2 A.

Fig. 8. A section of an egg of the two-cell stage, shewing the
division of the yolk, and also y<j'k columns, the segmentation cavity,
and the yolk nucleus. (^Jjijcosii). 2 C.

F^ig. \). A section of an egg (jf the sixteen cell stage. {Lijcosli).
2C.

Fio-, 10. A section of a se^'mentation eg-u' in the stag^e of Fig.
3, containing twenty two nuclei. {Liicasa). 2 C.

Fig. 11, A portion «jf a section of an egg in the stage of Fig.
5. (Lijcosa). 2 C. .

SQ K. KISHINOUYË.

Fig. 12. A portion of a section of an egg in the stage of Fig.
6. {Lycosa). 2 C.

Fig. 13. A section of the secondary thickening. (Lncosa). 2D.

Fig. 14. A portion of an section of an egg, a little more ad-
vanced than the egg in the stage of Fig. 12. (Lijcosa). 2 C.

Fig. 15. A section of an egg after segmentation showing
the absence of the nucleus in the yolk and a number of small yolk
balls. (Liicosa). 2 B.

Fig. 16. A longitudinal section of an egg of the protozonite
.stage. {Agalrita). 2 B.

Fig. 17. A ))ortion of a cross section of an egg of the proto-
zonite stage. (^Agak'Nu). 2 1).

Fig. 18. A cross section of an egg showing the separation of
the mesoderjn into two lateral halves, the formation of the cœlomic
cavity, and the appearance of the appendage, in the thoracic region.
The mesoderm of the cephalic region is not yet divided. (Ai/alcfia).
2 B.

Fig. 19. The median longitudinal section of the embryo in
the reversion stage. ÇAgalenci). 2 B.

Fig. 20. A diagram of the ventral [)late (imagined as unrolled)
of an embryo in the stage of the maximum dorsal tiexare.

Fig. 21. A diagram of the ventral [»late (imagined as unrolled)
of an embryo in the stage of reversion.

Fiof. 22. A lonofitudinaJ section of an e^s^ in the staue of Fiy-.
20, showing the appendages and cœlomic cavities. (Agalenay 2 B.

Fio;. 28. A cross section of an eora- in the same stage as of the
previous figure, showing the semicircular groove, the lateral vesick',
the continuous mesoderm of the head, and the cœlomic cavities and
thoracic ganglia. (Agalena). 2 B.

Figs. 24, 2t5. Fortions of median longitudinal sections, showing

0\ THE DEVELOPMEN'T OF Ah'.ANh:iXA. ö(

clo.seiies.s of the ceplialif and anal lobes, and the formation of the
«toniodaïum (Agalcua'). 2 B. . ^ .

Fig. 2(). A portion of the median longitudinal section of an eg<^
in the reversi(jn stage, showing tlie expansion of the dorsum.
{A(jalena). 2 B.

Figs. 27-o2. Longitudinal sections of tiie anal lobe in succes-
sive stages, showing the formation of the proctod^eum and the change
of the cœlomic cavity of the anal lobe to the stercoral pocket.
{Aijakna). 2 D.

Fig. 33. A cross section of the anal lobe, sh(_)wing its unpaired
cœlomic cavity and gangli(jn, and its two lateral diverticula. (^Agalena).
2 D.

Fig. 34. A sagittal secticjn of the alxlornen of an embryo after
the reversion stage. Two anterior abdominal septa are represented.
{Agalenci). 2 C.

Figs. 35. 36. Sagittal sections of the coxal joint of the tirst
thoracic appendage, showing the communication of the cœlomic
cavity with the exterior by an ectodermic invagination. (^Agah'na).
2 D.

Fig. 37. The glandular portion and the outlet of the coxal
gland. (Agaleua). 2 D.

Fig. 38. A cross section of the céphalothorax, showing the
position of the coxal gland. (Agalena). 2 B.

Fig. 39. A cross section of the poison gland of an embryo, a
little before hatching. (A galena). 2 D.

Fio-s. 40-43. Portions of cross sections of the abdomen, show-
ing the formation of the dorsal circulatory organ. (^Agalcna). 2 D.

Fig. 44. A porti(jn of a frontal section of an embryo in the
reversion stage, showing the three segments of the brain. (Agalena).
2 B.

88 K. KISHINOUYE.

Fig. 45. A diagram of the brain and the cheliceral ganglia.

Fig. 45a. A diagram of the profile view of the brain and the
cheliceral ganglia of an embryo in the reversion stage.

Fief 45. A frontal section of the 1)rain of an embrvo \u tlie
reversion stage, showing the formation of the eye. (Agaleiiay 2 G.

Fig. 47. A porticm of a cross section of the alxlomeii in the
reversion stao-e, showino- the formation of tlie hino-l)ook lamella'.
{Agalena). 2 F.

Fig. 48. A sagittal section of the brain of an embryo in the
reversion stage, showing the formation of the posterior median eye.
(Agalena). 2 D.

Fig. 49. A sagittal section of the posterior median eye of a
hatched embryo. {Agalenci). 2 D.

Figs. 50, 51. Portions of frontal sections of the céphalothorax
in different stages of growth after the reversion of tlie embryo, show-
ing the development of the anterior eyes and the formati(jn of nerve
fibres, the tapetum, and the vitreons body. (Lycosa). Fig. 50. 2
F. Fig. 51. 2 D.

Fig. 52. An oblique frontal section of the anterior lateral eye
of an embryo about the time of hatching. (^Lijcosa). 2 F.

Fig. 53. A longitudinal section of the anterior median eye
about the time of hatching. (Lijcosa). 2 F.

Fig. 54. A frontal section of the anterior median eyes of a
hatched embryo. {Lycosa). 2 D.

Fig. 55. A sagittal section of the abdomen about the time of
hatching. (Lycosa). 2 B.

Jour. Sc. Coll. Vol. IV PI. XI.

Pr-im. th./

8

.^^^

Y.N.

^^0$'''

Jour. Sc, Coll. Vol. IV PI. XII.

'' jm^

mn

G "

^Seg. ca;'/

12

Prim.th.

Sec- th.

Sec- th.

Ù !^

lb

.^o e@@ 9®_

©©®'

^,

11

-Prim.th.

0 8 G ogva,eo

14^

-'S«©

»•©

Sec. th.

«I ©»©'

cr -

-Jl ^~

Jour. Sc. Coll. Vol. IV PI. XIII.

H^<rvç^fôV^^

f

r'-

17

^^

Ceph.

.••'i"

•::..x--<^^CQ /

%^-A

End.

Th. app-

V.iäls^'«^''''''«*^'*^''""*'*'»^ Coelom. •..,-.

Stom.

^n.

End.

,V — Sem.gr. ^1 ''\— ^

Jour, Sc. Coll, Vol, IV PI. XIV.

Ceoh.l

2 a

Otni. p-r.

1 Abd

Sterc.p.

Jour. Sc. Coif. Vol. IV PI. XV.

Sp.
4- Abd app.

Jour. Sc. Coll. Vol. IV PI. XVI.

Pjst. mesent. Malp. t.

Observations on Fresh-water Polyzoa.

(Pectinatella gelatinosa, nov. sp.)

by

A. Oka.

Iiuperia] University, Tokyo,
with Plates XVII— XX.

The pi'csent paper emborlies the results of my investigations on
a new species of Fresli-water Polyzoa that lives in a large pond in
the grounds of the Imperial University, Tokyo, and is published with
the hope that it may throw some light on certain points in the
structure and development of the order Phylactolaemata, which have
hith('rf(^ remained ol)scure in spite of many efforts of former investi-
gators. The researches were begun, in the spring of 1888, at the
suggestion of Prof, [, Ijima, and I am indebted both to him and to
Prof. K. Mitsukuri for useful advice. My thanks are also due to
Mr. S. Watase, now of the flohns Hopkins University, who, while
here some years ago, studied the same species, for kindly sending me
his drawings showing the formation of the statoblast.

Altliough the species which I have studied does not agree in some
points with the generic description of Pectinatella in Hyatt's Observa-
tions (5), it can belong to no other genus. The statements given
there were made when only one species, viz Pect, magnifica, was
known, and must certainly be modified to receive the new one. The

90 A. OKA.

diagnostic cliaracters of the present species to which I give the name
of Pect. geJatinosa, are as follows :

Colony oval, hyaline; branches of cœnœciuni di-
chotoinous ; no septa between the eel 1 s ; ec tocys t gel ati -
nous, fills ap the space; between the branches, forms a
common base for many colonies; invaginated fold
obsolete; alimentary canal straight when retracted;
tentacles 90-98; statohlast saddle- shaped , curved in
two axes; marginal spines minute, only seen under a
moderate power of microscope.

The colonies grow among aqueous plants and on the underside
of floating logs just below the surface of water, and seem to
flourish in direct sunshine as well as in shadow. They are found
together in a large number forming a luxuriant mass of gelatine,
sometimes two metres in length. The outline of each cohjny is
irregularly hexagonal on account of mutual pressure. The gelatinous
ectocyst of neighboring colonies coalesce, and form a common base
2-3 cm. thick.

This species furnishes very favorable materials for the student of
this group of animals, the transparency of its gelatinous ectocyst, the
unequalled large size of the polypide and the promptness with which
they evaginate, giving great faciUty for investigation.

The general a{)pearance of a group of colonies is represented in
natural size in fig. 1, PL XVII. The color of the cœnœcium and
tlie lopliopiiore is sligiitly yellowish, the œsophagus and the stomach
are brown, and the rectum usually contains dark grayish refuse
matter, otherwise of light brown color.

The largest colony that I have seen me.isured 7 cm. in diameter .
The polypides are most crowded and in fullest vigor along the
margin of a colony, aiid much less crowded in the middle portion^

OBSERVATIONS OX FßE.-jH-WAl'EK POLYZOA. 91

&ay about one in ei'.cli four «(^uure millimetres. The more centrally
situated polypides being older are the tirst to die, so that in old
colonies, th.', polypides are f(3und only on the outer part, leaving the
inner part bare and only marked with dark spots, the remains of dead
polypides. When agitated the polypides retract only for a short
time, and soon expand their tentacular crown again. Even in being
transferred from one vessel to another, some of the ])olypides of a
colony do not reti'act at all. In confinement, however, they seem
to become more timid and, once retracted, remain in that state for a
longer time than when free.

Each colony originates from a single individual that comes out
of the statoblast in the first weeks of July, becomes larger and larger
by successive budding, attains its full growth in October, and
C(mtinues to live until the end of December. Compared with a
species of Plumatella living in the same pond, the times of the first
appearance and of the total disappearance are each about two months
later. As I have not found this species anywhere else, I can say
nothing about its geographical distribution.

Methods of Investigation.

Before proceeding further, I may here give a brief account of
the methods of investigation employed. To kill the animal in
a fully expanded condition was in this case very easy, although it is
the princij)al difficidty met with in the preservation of all other
genera. When 70% ^dcohol is gradually [)Oured into a vessel contain-
ing the colonies, more than half the ])olypides die protruded. If we
use such stupefying reagents as chloral hydrate or ccjcain chlorhydrate,
every one of the polypides dies in a fully ex[)aiKled condition.

The colonies after being killed were put into alcohol to be hard-
ened. S(jme of them were fixed with a saturated solution of c<jrros-

92

A. OKA.

sive sublimate or a weak «olation (0.1%) of cliromic acid, previous to
hardening. For staining-, borax-carmin and picro-carmin were chiefly
used. In cutting sections, I imbedded sometimes a whole colony,
sometimes separate polypides, in celloidin and paraffin.

In studying the development of the polypide within the stato-
blast, I proceeded in the following way. First, a statoblast was put
into alcohol to harden its contents which in tlie fresh state consist
of a thick milky fluid. Then it was held between two pieces of elder
pith, and the edge was cut with a sharp razor so as to make an open-
ing in the chitinous shell. Next, it was stained and kept in alcohol
until it was to be cut. In cutting tlie statoblast, celloidin v/as in-
dispensable, for, the shell being too hard, it was impossible to get
good sections with paraffin only.

For examining fresh specimens, the only thing I had to do was
to put a colony (stupefied with Cocain) or a part of it on a slide,' and
cover it, putting a wire ring under the cover-glass to prevent over-
pressure. In this condition, the polypides had no power to retract,
and the ciliae were in vi^'orous motion.

To study their habits, I kept colonies alive in a glass vessel. I
kept also the statoblast in a vessel, in which a contrivance was made
to have water always flowing. At last the shells burst, and the little
polypides peeped out of the sutures, carrying about the shells like a
tiny bivalve. Each of them floats about for a very short time, and
then attaches itself by means of the gelatinous ectocyst to any object
it may meet with, and gives rise to a new colony.

A. Anatomy.

The branched membranaceous tube (cœnœcial endocyst) forming
the greater part of the mass of a colony, together with the gelatinous
covering (ectocyst) over it, constitutes the cœnœcium. The terminal

OBtJEHVATlOXS OX FRESH-WATEE POLYZOA. 93

portion of each branch is turned nearJy vertical to the plane of the
colony and is capped by another short tube (polypidal endoeyst\
through tlie pellucid wall of which is seen the alimentary canal con-
tained within. This terminal tube, with the tentaculate lophophore
at its free end, and several delicate organs in its cavity, is called the
polypide (fig. 2, PI. XYIL).

Besides this division of the colony into the cœnœcinm and
the polypides, we may divide it into a number of equal parts, each
consisting of a polypide and a portion of the cœnœcium. For the sake
of convenience I shall call such a part ^' pulijzouid" and the portion
of the cœnœcium belonging to it ^^ cystid.'' AVe thus consider a
colony as being made up of as many polyzoöids, all structurally alike,
as there are polypides.

In all genera with chitinous ectocyst. the ca'nœ^ciuni is dixided
by more or less de\eloped septa into a number c)f compartments or
cells, each destined to receive a polypide when the latter is retracted.
Such septa are not found in forms wdth gelatinous ectocyst, and the
cystidal cavities stand in open connection with one another.

When a polypide is retracted by the contraction of the muscles that
connect it with the bottom of the cystid, its tubular wall invaginates
and 1)6 G(3mes a sort of sheath for the tentacles, known as the tentacular
sheatli. In the process of evagination, the tentacular sheath begins
to reflect upon itself from the lower end. The evagination generally
stops when the lower end of the polypide is still within the
cystid. In other words, the evagination is incomplete, thus leaving
a permanent fold at the boundary between the cystidal and the
polypidal endocyst. In this genus, howevei-, the polypides are often
stretched out their whole lengtli, and then no such fold is to be seen.

The shape of the polyzoan colony is ditterent in dilferent genera
and species, but it is characteristic for each species. The manner of

94 A. OKA.

braucliiijo- of the cœuœciuni iu Vaci. iie]atirio.s;i is .shown in lii"-. 3.
PJ. XVII.. It is (lichotomouä with a sh(;/rt branch at each axiL
The branches are no bent that all the poJypides stand npright and
as the phnnous tentacles cover the whole surface of the coh)ny, their
regular symmetrical arrangement cannot be discerned without chjse
examination.

The general plan of the structure of a [xdypide and its relation
to ihe cystid are shown in tig. 4, I'l. XVII. The alimentary canal
is bent in the sliape of the letter V, and hangs freely in the perigastiic
cavity. The mouth guarded by a tongue-like epistome (Epist.) is
surrounded by a number of tentacles (Tent.) arranged along the
entire margin of a horse-shoe shaped Jophophore (Loj)Jt.). The anus
opens outside the tentacular area near the mouth, on that side of the
body on which the arms of the lophophore stretch out. A nervous
ganglion (X. Gang.) is seen on the anal side (jf the oesophagus. A
thin hollow tube, called funiculus, in which the statoblasts are
developed, joins the angle of the alimentary canal with the cystidal
wall. An ovarv (Orr.) is seen inside the tip of the cœnœcial branch.
The length of a polypide from the tip of the tentacles of the angle
of flexure of the alimentary canal is about 4 mm.

Although the term " individual " as applied to such foi'uis as
polyzoa is very diihcult to detine, yet homulogously with its nearest
relative, the Ijrachiopods, each polyzooid might be regarded as an
individual in the (ordinary sense of the word. Polyzoan individuals
show a close analog}' to " phytons " of plants.

The polypide and the cystid that constitute a pol\zoüid, are
respectively vegetative and reproductive in function. As will be
seen further on, all the functions for the preservation of the species
are performed by the latter, the funiculus being regarded as a part of
it, while the former serves to procure nourishment to the cystid.

OBSERVATIONS OX FRESH-WATEK POLYZOA. 95

All fresh-water Polyzoa are annuals, tlie vegetative and the reproduc-
tive portions undergoing entire decomposition every year, while in
marine forms, several generations of the vegetative portion, i.e., the
polypides, form and decompose themselves on the perennial cystid, like
leaves on thelmniches of a. tree. This singular phenomenon led many
naturalists (Allmnn nud others) to regard the polypide nnd the cystid
as two distinct individuals. In the present species also, the duration
of vitality of the two portions is by no means the same. The poly-
j)ides invarial)ly die after a certain period of existence, usually after
the formatioii of younger polyzooids of the fourth or the fifth order,
but the cystids remain until the colony itself disintegrates in winter.
In the central porticin of a large colony, therefore, we often see only
hare cystids, each with o dark grayish mass, the remains of the dead
polypide, banging in its cavitv, ;nid yet with statoblasts continuing
their development in tiie funiculus.

About the application of the terms " antei-ior,'' "posterior,"
"dorsal," "ventral," Ac, there is much diversity of opinion. For in-
stance Allman calls the free end of the polypide "anterior," and
the fixed end "posterior"; while Hyatt, following E. S. Morse,
calls the fixed end " anterior," and the free end " posterior."
Huxley homologizes Polyzoa with Tunicata, and names that side on
wliich the anus opens " neural," and the side opposite to it " haemal,"
although there exists no heart. Again, if we were to compare
this animal witli Phoronis, we should have to call the narrow spare
between the mouth and the anus "dorsal," and nil other parts
" ventral." In fact, every one mioht «five different sets of names in
orienting the animal, according to his conception ( f the homology
which exists between Polyzoa and other nnimals in which the an-
terior and posterior, or the dorsal and ventral ])oles are universally
recognized. In the following pages, I shall call the fixed end

96 A. OKA.

" linvor," and the free end " upper," the side on which the anas opens
"nnal,' and the side opposite the anus "oral."

The organs that constitute the Polyzoan hody may he classified
in the folio wi no- way.

A. Organs for the preservation of the polyzobids
or the colony.

1. Dermal System, consisting of the ectocyst and the endocyst.

2. Digestive System, consisting of the epistome, the oesophagus,

the stomach, and the intestine.

3. Tentacles.

4. Excretory Organs (?), consisting of two short ciliated tiihes.

5. Muscular System, consisting of five groups of muscles.

6. Nervous System, consistino- of a o-anoflion with two arms

f >r the lophophore.

B. Organs for the Preservation of the species.

7. Ovary and Testis.

8. Funiculus, in which the statoblasts are developed.

9. The part of the endocyst that prc^duces buds.

1. Dermal System.

The integinnent of Polyzoa consists of two layers, quite different
in their nature, the outer "ectocyst" and the inner "endocyst" (see
fig. 4). The latter is not everywhere covered by the f n'mer, but is
exposed on the polypides.

The ectocyst is gelatinous in this species. It fills up the space
between the branches of the cœnœcial endocyst, whereas in Pect,
magnifica, Leidy, there is no ectocyst between them. In this respect

OBSERVATIONS ON FRESH-WATER POLYZOA. y<

as well as in the erect position of polypides. this species comes nearer
the genus Lophopus. The oelatinons substance is formed Ly the
secretion of cells of the outer layer of the endocyst. Numerous cells,
some oval, others irregidar in their shape, are scattered in it (fig. 6,
PI. XVIII.). Their nucleus and nucleolus are distinctly visible.
These cells seem to have wandered out of the outer layer of the
endocyst, and may have helped in producing the gelatinous substance,
remindin-T us of the cells in the test of the Tunicates. 'i'he gelatinous
substance is adhesive and without taste ; it ser\ es apparently to
protect the coh^ny. On drying, it shrinks almost to nothing.

The endocyst consists of four layers (fig. 7, PI. XYIII.)

a. Outer cell layer (OiU. Icuj.).

h. Basement membrane (^Bas. vianhr.).

c. Muscular layer (L. mus. and Tr. nuis.)

d. Inner lining epithelium (Lin. cpith.).

All of these layers are not present everywhere, nor is each ol
them of the same structure throuHiout its distribution.

The cells of the outer layer, which represents the ectoderm, are
everywhere distinctly bounded, columnar on the cœnœcium, tiat and
h<)riz(jntally elongated on the polypide, except on the tentach^s and
the upper surface of the lophophore. In the former, they are culjical,
in the latter hexagonally prismatic, and distinctly ciliated in either
case. Many of the cells on the cœnœcium contain a vacuole (fig. 7,
vac.) filled with a very refractile fiuid. The numi)er of these
vacuolated cells increases as we ap[)roach the tip of cœnœcial branches,
where every cell shows a large vacuole, almost filling up the whole
cell (fig. 33, PI. XIX. Out. huj.).

In preserved specimens, the cells are more or less shrunk, often
leaving' spaces between them. The nuclei are oval, and have a dis-
tinct, well-staining nucleolus. The cells on the cœnœciiun are

98

A. OKA.

0.02-0.04 mm. liiuh. The nucleus measures about 0.007 x 0.00 i mm.

The basemeiit membraiïe situated directly Ijeneath the outer
celldayer is secreted either by this, or by the internal lining epithe-
lium, or by both. In the gretiter part of the cœnœcium where this
]nembrane is separated from the inner epithelium l)y the muscular
laver, it would be natnral to refer its origin to the outer cell-layer
alone, but where the muscular layer is deficient, it is difhcult to
decide. On the other hand, in the wall of the funiculus into which
this membrane and the inner epithelium, but not the outer cell-
hiver, are continued, it cannot but be the jn'oduct of the inner
epithelium only. Generally, the l)asement membrane and the mus-
cular Go;it are treated as one layer under the name of Tunica nuis-
cularis, but as they are in reality quite distinct from each other, it
will ]je better to regard them as two distinct layers. AVlien a colony
is treated with a weak solution of acetic acid, the basement membrane
separates from the rest of layers. It is thin, tough., transparent, and
homoii'eneous.

Next to the basement membrane conies the nuiscular layer, con-
sisting' of transverse and lomjitudinal fibres. The f3rmer run
external to the latter. They are not very densely set, so that in
a surface view they cross one another as in coarse linen. On the
main ]>art of the polypide, only the longitudinal fibres are present.
In such genus as Cristatella, the muscidar layer gives the colony the
power of slow locomotion, but what function it has to discharge in
fixed Pectinatella, I am not prepared to say. In the cœnœcium
where this layer is best developed, it is 0.005 mm. thick. It is not
found in that part of the endocyst wh.ere buds^ are formed, and is
also absent in the walls of the lophophore and the tentacles.

The internal epithelial Inyer lines the endocyst everywdiere. It
is thickest in the cœnœcium, especially at those points where budding

OBSERVATIONS OX FRESH-WATER POLYZOA. 99

takes |)]:ice, and is thinnest in the tentacles with nuclei scattered
widely apart (%s. 16 and 17, PI. XVIIL). The cells of this layer
are fused, hence cell boundaries cannot he distinguished. The
nucleus is oval, but I am unable to detect distinct nucleoli. The
size of the nucleus is nearly the same everywhere, and is about
0.00(S X 0.004 mm. This layer is furnished with short cilia, which
set the perigastic fluid in motion. Average thickness of the layei" in
the cœnœcium is 0.008 mm.

2. Digestive Sj^stem.

Minute algae and infusoria that pass by are caught in the whirl-
pool caused by the vibrating cilia of the tentacles, and sent into the
œsophagus. The e[)isti)me that guards the mouth is ftu'ni.shed with
special muscles which enable it to shut the oral aperture now and
then. Perhaps the entrance of non-nutritive matters is prevented by
this contrivance. The food, after staying ibr a short time in the
oesophagus, pushes open ihe funneldike valve (fig. 4, PI. XVI I.
fniuicl-like v.) that intercepts free conununication between the oeso-
phagus and the stomach, and enters the gastric cavity where it is
moved about by the peristaltic contraction of the wall of that organ.
After l)eing fully digested, the residue mainly composed of the cell-
wall of diatoms and other algae, passes through the pyloric valve little
by little, and accumulates in the intestine. Here, the refuse matter,
usually of a dark-grayish color, is cemented together into a mass by a
transparent gelatinous secretion of the intestinal wall. AYhen the
intestine is full, the contents are pushed out of the anus by the
agency of the muscles of that part. The form of the excremental
mass, characteristic of each genus, is the same in form as the lumen
of the intestine which in our species is an elongated oval tapering
toward the anus.

100 A. OKA.

There ni-e often certain amœboid cells to be found in the intestinal
cavity. The}^ stain very well, nnd are on that account vei'y con-
spicuous among a mass of unstained matter. Judging- fcom their
shape and size, it is very probable that they are parasitic Protozoa.

The process of digestion is carried on very rapidly. When fresh
colonies are brought from the ])ond and kept alive, all the polypides
dischnrge their dark intestinal contents in a few hours, rirndually, new
refuse niatters ])egin to accumulate in that organ, i)nt thev are always
i\ go;()d deal lighter in cohn'. These are again excreted iii the next
three or four hours. As the amount of food that these animals con-
sume is considerable, it was impossible to keep them alive more than
a week without furnishing them very often with water from the pond,
which C'.mtained minute organisms.

The layers that constitute the walls of the alimentary canal are
the same as those of the endocyst. In fact, they are direct continua-
tions of the latter only slightly modified to serve special piu-p(^ses.

The epistome is a tongue-like prolongation of the disc of the
lophophore on the anal side of the mouth. Its cavity (see fig. 8, PL
XVIII.) communicates with the general perigastric cavity by a com-
paratively narrow passage on the anal side of the cerel)ral ganglion.
The cells of the outer layer of its wall are similar in appearance to
those of tlie lophophore. They are prismatic, and the height increases
nearer the mouth. The oval nucleus with distinct nucleolus lies
near the base. The whole external surfiice is furnished with cilia.
This organ has no muscular layer in the wall, l)ut is furnished
with special muscular fibres which traverse its internal cavity.
These fibres are simply elongated cells with the uncleus at about the
middle of their length. They are separate and never form bundles.
The length of the epistome is about ^mm.

The oesophagus is that portion of the alimentary canal that lies

OBSERVATIONS ON FRESH-WATER POLYZOA. 101

hotween tlie month and the fiiunel-hke vnlve nt the cardia] opeiiino-
of the stoiiiaeh. Its upper aud lower sections are lined by epithe-
lia of quite different appearance. The cells of the nipper section
(fis:. 9, PI. XVIIl) have cilin, and their nncleiis lies near the base.
Yerworn says that the cells of this section do not come in to any
contact with one another throughout their whole length, being
separated l3y a narrow intervening space but I cannot find any
such space in Pect, gelatinosa, except such as is in all prob^ibility
due to the post-mortem contraction of cells. In the lower section,
the lining cells have no cilia, and the nuclei lie irregularly near the
middle (fig. 10, PI. XVIII). In the upper section, the free end of
cells is -flat ; in the lower, it is rounded. In both the nucleus has a
distinct nucleolus. The cells of the lower section do not stain Avell,
and seem to contain a secretive sul)stance, which may be comparable
with the saliva of higher animals.

The length of the œsophagus is about 1mm. and its diameter 0.3
mm. The lumen of the œsophagus when expanded is round in sec-
tion throughout its entire length, but in its upper section contraction
changes it into a stellate sliape. The inuscular layer is but scantily
developed in the œso'phageal wall. The outer covering is the con-
tinuation of the lining epithelium of the endocyst with which it
agrees in all respects.

The œsophagus in its downward course occupies an excentric
position in the tubular body of the polypide, and where the latter
is externally marked off from the lophophore l)y a slight constric-
tion it actually comes in contact with the body-wall on the oral side.
At this point, the lining epithelium of the polypidai wall is continuous
with the outer covering of the œsophagus, and fwrns a sort of
mesenterv (fig. 15, PI. X^'IIL). This mesentery extends horizontal-
ly on both side for a short distance, and prevents over-invagination

102 A. OKA.

of the body-wall when tlie polypide is retracted. Thus the ahmen -
tary canal is attached to the body- wall at four points, viz. the mouth,
the anus, the fnniculus nnd tlie above mentioned mesentery.

At the entrance of the stomach there is as already mentioned
a funnel shaped valve, with the free end pointing into the cavity of
the stomach (tig. 11, PI. XVIIL). It consists of a fnnnol-like pro-
longation of the basement memljrane, on the oesophageal side of
which are arranged the characteristic cells of the œsopliagus, and
on the gastric side, the pyramidal cells of the stomach. This valve,
whose length is about 0.2 mm., prevents the passage of food from the
stomach back into the œsophagus.

The stomach is a spacious saccular orgnn whose long axis is
bent in the sha[)e of V. bi'inging the pyloric opening nenr the cardiac.
It measures 2 mm. in length, and 0.6 mm. in breadth at the widest
part. The inequality of the length of the arms of Y brings the
cardiac opening about 0.5 mm. nearer the free end of the polypide
than the pyloric.

The inner layer of the stomach has two kinds of cells ; the long
club-like cells (fig. 13, I'l. XVIII. cL c.) and the short pyramidal
cells (pill'. ('.). As they are arranged in gr(3Ups forming alternate
longitudinal rows, the lumen of tlie stomnch is stellate in cross-
section. The number of the rows of each kind is generally twelve
or more (hg. 12, PI. XVIIL). In both, the nuclei lie at the base and
the nncK'oli are distinctly visible. The l<:>ng club-like cells do not
shiin well, while the short pyramidal cells freely take up the coloring-
matter. In the fresh state, the longer cells contain a yellowish brown
fluid and the shorter cells are of a light yellowish color, so tliat the
stomach appears longitudinally striped with yellow and brown
bands. As the alimentary canal has no distinct glandular appendage,
the brown fluid contained in the longer cells probably performs the

OBSERVATIONS ON FRESH-WATEK POLYZOA. lOo

function of the dig-estive fluid. Hence tliey huve been called hepatic
cells by Alluum. Tlie function of absorption seems to be pei-fornied
by the shorter cells. The length of the longer cells is various, the
longest measuring 0.06 mm., while the shorter [)yramidal cells mea-
sure approximately 0.02 mm. On the gastric side of the cardiac
valve, and at the blind end where the stomach is continuous with the
funiculus, the rows of the longer cells stop short, and only the short
pyramidal cells are })resent.

The muscular layer of the gastric wall, composed only of the
transverse fibres, is w^ell developed, especially below. At the thickest
part this layer is 0.007 mm. in thickness. At the l)lind end of the
stomach, however, tliere is no muscle, and here the inner cell-layer
comes in direct contact with the fluid contents of the funicular cavity
(fig. 32, PL XIX.). At this point, the wall is generally pushed
inward in the form of a shallow pit.

The outer epithelium does not diiter from the corresponding
la^er of the œsophagus and the endocyst.

The pyhjric valve is represented by a simple constriction of the
entire wall of tlie alimentary canal. Its opening is very narrow,
allowing the passage of only a small quantity of indigestible matter
at a time.

The intestime is a tul)ular organ tapering tow^ard the anus. It
is about 1.2 nnn. in length, and 0.3 mm. in width. The inner layer
is composed of only one kind of cells, which are much shorter but
somewhat broader tlian the longer cells of the stomach. The height
of these cells is about 0.025-0.03 mm. The nucleus is at the base and
the nucleolus is distinct (fig. 14, PI. XA'IIL). These cells do nut
stain well ; the gelatinous fluid they contain is probably the medium
by which the excrement is cemented into a C(jmpact mass. The mus-
cular layer of this part, in which onl}- ring fibres are present, is weakly

104 A. OKA.

developed except near the aiiu.-^, where it f(3nns a sort of si)hincter.
The anus when expanded is us wide as the widest part of tlie in-
testine, but when contracted it closes altogether. The outer cell-
Inyer is similar in all respect with that of other parts of the alhnentary
canal. At the point where the intestine is tightly pressed against
the oesophagus, the outer layer of the former ])asses directly into that
of the latter, bringing the cells of the inner layers of both organs in
C(3ntact.

3. Tentacles.

The tentacles are arranged in one continuous series along the
outer and the inner margin of a horse-shoe shaped lophophore, as
mentioned before. They are hollow cylindrical organs measuring
1 mm. in length, and 0.03 mm. in breadtli. They are t(3 be con-
sidered as prolongatioiis drawn out, as it were, from the endocyst.
In the living state, they are freely movable in every direction at the
will of the animal, but I have never seen them coil or contract.
Generally, they stand nearly parallel to one another in graceful
curves (fig. 2, PI. XVIL).

The cross section of a luphophoral arm (tig. 30, PI. XIX.) is
almost semicircular in outline, slightly convex above and rounded
below, measuring 0.3 mm. in breadth, and nearly as much in depth.
The ciliati(3n on the upper surface is distinctly visible on sections.

The cells of the outer layer of the tentacular wall have all the
essential characters of those of the endocyst. They rest on a fine base-
ment membrane and are furnished each with a long cilium (fig. 16,
PI. XVIII. (Jut. lay.), constantly vibrating in a certain fixed direc-
tion. The ciliation ofthat side of the tentacles turned away from the
mouth drives the water upward, while that on the opposite side tends
to drive it toward the mouth below. The iiiner layer of the tentacles

OBSERVATIONS OX PRESH-WATEli POLYZOA.

105

(fig. 16, PL XVIEI. Lin. ej)itli.) is very thin and has the nuch-i scat-
tered at great intervals. I was not aljle to detect any trace of cih'a on
the lining epithehiim, but the rapid motion 'of the [)erigastric "fliud,
a'oins" toward the tii) alono- one side and coming' back ak^n«' the other
in the narrow tentacular cavity, indicates their existence. The
lumen is a little more than 0.01 mm. in diameter.

The account, given by Yerworn, oi' the manner of junction of
tlie tentacles with the Io[)hophore and the tentacular membrane in
CrUtatella applies equally well to the species investigated by me.
In fig. 18, PL XVIII, I have endeavored to show diagrammatically
the relation of several parts at the bases of tentacles.

Externally io the row of tentacles there is a thin membrane,
the tentacular membrane, 0.3-0.4 mm. in Ijreadth, formed by a,
duplicature of the outer layer of the lophophoral wall along its outer
edge. It consists of a basement membrane covered on both sides by
a layer of flat cells, tlie direct continuation of the outer layer. The
basal portion of each tentacle is joined to the tentacular membrane
by another narrow triangular membrane.

Alternating with the bases of the tentacles, a series of duplicatures
(jn each side of the lophophoral cavity is produced in the inner layer,
so that if Tve were to cut across the arm and look into it, we should
see a series of vault-like arches. The tentacular cavity opens »nto that
of the lophoph(3re l^etween each two of such folds of the inner layer.
These folds descend almost to the floor of the lophophoral cavity, and
have been reckoned as part of the muscular system by Hyatt, under
the name of" brachial contractors," but I see no ground for regard-
ing them as such, since they consists simply of flat cells.

The bases of the tentacles are not in one plane. Those on the
anal side near the e[)istoine are the most elevated. Tlie numl)er of
the tentacles is generally even, but in some individuals there is a

106 A. OKA.

median tentacle on the anal tside, making the total num1)er «xld.

There can be uo d(juLt that the function of the tentacle« is three-
fold, äer\ing for respiration, for collecting food and for feeling.
Of these, however, the first seems to l)e their principal office,
when we consider the large extent of their surface exposed to
water, and the constant current kept np in the latter hy a special
contrivance, as well as the perigastric fluid that circulates within
their lumen. The Tentacles thus bear a close resemblance to the
fringed arms of ßrachiopods.

Circulation. The perigastric fluid contained in the general body-
cavitv may justly be regarded as representing the blood. Of its
nature and the mechanism of circulation, little was known before.
There are no special orgarjs, such as heart and blood vessels, and the
only means of driving the perigastric fluid is the supposed ciliation
on the lining epithelium of the general body-cavity. The nutritive
part of the food taken up by the alimentary canal is conveyed to all
parts of the body by this fluid. It is transparent, colorless, and has
no taste. Water seems to constitute the greater part of its con-
stituents.

The fluid contains, floating in it, numerous round cells, each
with a large vacuole almost filling up its body and filled with a
refi-actile fluid (fig. 20, PI. XYIII). The nucleus is pushed against
the wall by the vacuole. The study of the development of polypides
in the statoblast shows that these free cells are derived directly from
the iiianular mass that constitutes the main contents of the statoblast,
and in young stages they contain similar granules instead of the
vacuole. It is therefore plain that they are, at any rate, nutriment
carrying cells, which might be regarded as blood corpuscles.

OBSERVATIOXS ON FRESH-WATER POLYZOA. 107

Besides these, there are o'onerally present a g'reater or smaller number
of cells or fragments of cells of a c|nite différent appearance, which
have prc)hably detached themselves from some part of the body.

The floating elements were observed by previous investigators
(Allman, Hyatt, Verworn), but no great im])ortance was attributed
to what were brobably either parasitic organisms or detached cells.
Hyatt, for instance, observed "numerous organisms, many of which
probably parasitic, which float in the flnid, sometimes in such "a
number as to interfere with the examination of tlie internal structure."
It is prolmble that at least some of these " organisms '' were what I
regard as tiie blood corpuscles.

The direction of the blood currents as (feerved in the natural state
is show^i in fig. 21. On the anal side of the body cavity the fluid is
driven toward the free end of the polypide, evidently by ciliary
action, whicli however could never be actually Ijrought to view^ In
the lophophoral arms, tlie corpuscles travel ;dong the floor to their
ends, and either returii directly along the ceiling, or enter the
tentacles, in which they ascend on the side nearer the tip of the
loph(^phore, and descend along the opposite side. In the cavity of
the epistome, the fluid streams along the ceiling to its tip and com-
ing back ah:)ng the floor of that (^rgan, either enters the epistome
again, or goes to the tip of the lophophoral arm along its lower side.
On the oral side of the polypide, the fluid is always seen flowing
do ^vn ward.

Allman and Hyatt deny the presence of cilia on the external wall
of the alimentary canal, but Yerworn (saw them at the end of the
stomach in Gristatclla. My observations in living specimens of
Pectlnatella confirms tlie statements of tiie last author.

Both Allman and Hyatt observed that the cœnœcia of Lophopiis,
Cn^fatella, and Phnuatdla vitrea, readily emptied themselves of

I(l8 A. OKA.

their peri<^n><!fri(j fluid when taken out of the water. They assumed
that the fluid passed ont through pores in the endocyst, 1)nt they
searched in vain f(3r such communications.

It is certain tliat when a polyj)ide retracts, a portion of the fluid
contained must of necessity pass ont at some place, since the cœnœcial
wall does not expand heyond a very limited extent. Notwithstand-
ing my s[)ecial attention to this point, Feet, (jelatinosa also gave no
resnlt, and I should ])refer to go no further than to assnme the
presence of external openings in connection with the excretory organs.

4. Excretory Organ (P)

Joliet (6), in a paper entitled " Organe segmentaire des Bryo-
zoaires Endoproctes," gives a pretty full description of two short
funnel-shaped tubes in PedicelUna and Loxosoina, first noticed by
Ilatschek. In the division Ectoprocta, however, our knowledge on
lhis sul)ject is vei-y limited. As far as I know, the two ngures given
In' Fai're, ;uid the remarks by Uincks and Smitt, both of whom do
not go beyond conforming the observation of the first, constitute the
wliole ])i])liography on this subject. Tliey all noticed a ciliated pipe
that opens between the mouth and the anns in AUijonidiiuii and Mem-
hranipora, hotli of which are gymnoliematous. In regard to the order
Phylactolcematn. if we except the short account given by Yerworn,
illustrated with two semi-diasframmatic fig-ures. there exists no litera-
ture known to me. Yerworn left the terminations of that organ
nudetermined, confining his attention t«^ <^iily ^^hc middle portion
where it is most conspicuous. Braem touches on this subject in
his note in the Zoologischer Anzeiger, but he too could not deter-
mine how the tubes terminate. Such beins: the case, I have
investigated this organ with s[)ecial attention.

OBSERVATIONS ON FRESH-WATER POLYZOA. 109

There are two ciliiited tnhes just l^enentli the outer layer, on tlie
anal side of the body, l-jetween the aims and the bases of the median
tentacles of the inner row. The walls of these tubes are continuations
of the epithelial lining- of the invaginable portion of the endoeyst.
They open below into the body-cavity by fnnnel-shaped opening's.
They measure 0.15-0.19 mm. in length, though the portion where
the wall is entire is much shorter (fig. 26 bis. PI. XVIII). The
shape of the funnel-like openings may be compared most appropriately
with the obliquely cut end of a hollow tube.

The exact form of these tubes and their relation with other
organs will be best undestood by referring to figs. 21-26, PI. XVIII,
which show their cross sections with the neighboring parts at
various levels.

In a cross section passing through the middle part of the
tubes, we see them as two oval sections lying side by side (figs. 24
and 2-1 A). The ciliated epithelial wall consists of cells which are
cul)ical near the median plane of the polypide and flat on the opposite
side. Consequently both the nuclei and cilia are densely set in the
portion nearest the median plane of the polypide and scattered at
some distance from one another on the outer side. The tubes are
closely enveloped on the anal side by the outer layer of the invagina-
ble tube (Out. ki)/.), and on the oral side by the lining epithelium
(^Liii. cpit]!.) of the body-cavity. The diameter of the tubes measures
about 0.03 mm.

Tracirig these tubes downward, that part of the wall farther
removed from the median line soon disapjtears, i, e., the tubes
open into the body-cavity on that side (fig. 26). As the two tubes
deviate from each other below, a part of the perigastric s})ace appears
between them (fig. 26, rpistom. car.) This is the passage by which
the cavity of the epistome commu.nicates with the perigastric.

110 A.OKA.

The inodinn side of the wall ends abruptly on the anal side of
the o;anglioii ; below this point cross sections show only one conti-
nuous body-cavity. Thus, the body cavity is divided into three
brandies on the upper part of the polypide. The middle one (fig. 26,
(■r>i>ilom. cav.), passing along the anal side of the ganglion, extends into
the epistonie, while the lateral ones are prolonged into the lophophoral
cavity. The inner walls of these lateral branches pass gradually
into the ciliated tubes.

If we now trace the tubes upwards, they are found gradually to
approach each other, and their walls soon coalesce. A little higher
the cavities of both open into each other, and there is seen a single
flattened tube (figs. 23, Ncplir. & 23 A). The whole inner surface of
this part as well as that of the two deviating tubes below, show distinct
ciliation in sections, the cilia being always directed toward the
perigastric opening. If we trace this flat tube still further upwards, it
again becomes divided in most individuals into two, in some into three
tiil)es (tigs 22 t' & 22 A), each of which is continous with tentacu-
lar lumen. In this part, the ciliation is no longer visible, but
compared with the inner layer of the tentacles, there are more nuclei.
But, further upwards, the nuclei are fewer in number and the lining
epithelium presents similar appearance as that of ordinary tentacles
(figs. 21 ^' & 21 A). What can be the function of these ciliated tubes ?
The fact that they open int(^ the perigastric cavity by ciliated funnel-
shaped openings naturally reminds us of the segmental organs of
certain worms. And thus many observers have been induced to
regard the function of these tubes as being of an excretory nature. If
such is re:dl V th(î case, there should be some orifice by which they open
outwards, for the high degree of development they attain prove that
they are not useless remnants. This makes me venture to assume the
existence of minute apertures, at least on the two or three innermost

OBSERVATIONS OX FEßÖH- WATER TOLYZOA. HI

tentacles of tlie una! side, presumably at their tips, althonah I ;im
imaljle to prodnce any positive proof. The pores, if ever present,
must be of very minute size, indiscernible by ordinary methods in a
manner analogous to the pores at the tip of Actininu tentacles,

5. Muscular System,

The mnscnlar system consists of live groups of muscles. They
are :

a. Muscles of the funiculus.

h. Pariet<i-vaoinal muscles.

f. Retractor of the polypide.

(L Muscular layer of the alimentary canal.

e. Muscular layer of the endocysl.
T(^ these may be added the muscles of the e[)istonie.

Tlie first three are, as development shows, moditications of the
last two, which in turn may be regarded as only locally differentiated
forms of one and the same layer.

In the develo})ment of the polypide in the statoblast, the muscidar
fibres are formed from certain cells of the granular mass, and, in the
process of budding, from the lining epithelium. In either case, the
cells elongate, and become spindle shaped, with the nucleus at the
middle. They lengthen more and more and the nuclei become indis-
cernible, although these can often be made visible l;)y the aid of
acetic acid. Excepting some fibres of the parieto- vaginal muscles,
which remain in this state to the end, the muscular fibres are ex-
tremely thin, and do not sliow nuclei in tlieir interior. It seems that
these fine fibres arise by the longitudinal splitting of tlie original
muscle-cells, as is known to take place in many other animals.

The muscles are never striated. Even in the retractor of the

112 A. OKA.

polypide, which is obviously of grentest physiological iiiiport;iiice, the
fibres are smooth. In marine Polyzoa, however, I have observed
that the muscles of the avicularia and the vibracula are striated.

The muscular fibres belonging to the funiculus run longitu-
dinally on the inner surface of the basement membrane, on which the
cellular wall of the funiculus rests (fig. ol, PL XIX). They run
separately without forming bundles, and present the same appearance
as those of other parts. Their extreme fineness as well as their small
number agrees with the fact that the funiculus contracts, if ever,
in a very limited degree.

The muscles running between the cystidal wall and the bottom
of the invaginated fold (at the junction of the cystidal and the poly-
pidal endocyst) are called the Parieto- vaginal muscles (fig. -1, PI.
XVII, M"). Their fibres run either solitary or in bundles, forming
on an average 13-14 sets arranged somewhat radiall}'. Their points
of attachment to the cystidal Avail is irregular. These sets of
muscles cause the presence of the invaginated fold of the body-wall.
In i'6'c'f. (jdatinosa when the polypides fully expand, this fold,
which is otherwise distinctly present, disappears, the muscles relax-
in o' to their full extent.

The great retractors of the [)olypide consist of a pair of well
developed muscular bundles, right and left in the perigastric cavity
(fig. -1, PI. XVII, Jl/^^^). The fibres are modifications of the muscular
layer of the endocyst, extraordinarily developed to serve their special
purpose. The point of attachment of each bundle to the bottom (^f
the cystid is siiigle, but the upper portion is split into a large number
of smaller bundles which are inserted into the walls of the œsophagus
and the stomach at various places, but most numerously at the
upper part of the former. The bundle is ensheathed in a sort of fine
sarcolemma, which could distinctly be demonstrated at such places

OBSERVATIONS OX FERSH-WATER POLYZOA. 113

where the fibres were mechaiii<'a11y torn away leaviii^i!,- the shealh
uninjared.

In the nuisciilar layer oî tlie alimentary canal, (jiily transverse
fibres are well developed, and the longitudinal fibres, if ever there be
any, are very scanty. The layer becomes thicker as we approach the
blind end of the stomach. The musculature in (juestion [)erfjrms
peristaltic movements, periodically on the (esophageal and fairly
constantly on the gastric wall. The ])lii]d end of the latter is subject
to strono-er constrictii-ns in accui-dance with the thickened muscle-
layer cf this part. The peristaltic movement of the gastric wall helps
not only to move about the contents ofthat organ, but also to setid
the residue into the intestine. The muscular fibres of the intestinad
wall are especially well developed near the anal opening ; they serve
to discharge the excrements out of the l)ody and to keep the anus
tightly closed. At the point v\diere the blind end of the stomacli joins
the funiculus, there is no muscular layer (fig. oo, PI, XIX).

The muscular laver of the endocyst has already been treated
under the body-wall. The outer ring fibres are especially well
developed around the orifice oî the cœnœcial branches and f)rm a sort
of sphincter to close the opening produced when the polypides re-
tract. When the polyjâde is extended, the cœnœcial branch becomes
slender by the contraction of the ring fibres, but apparently it is not
by their agency that the polypides are pushed out, for tiiis pr(jcess
takes place even in a cœnœcial branch witii its wall cut open, so that
the fiuid contained can transmit no pressure upon the in\'aginated
polypide.

The muscles that move the epistome remain in a very primitive
state of development, coiisisting of hjosely distributed fibres which,
as already mentioned, are mere elongated cells with t lie nucleus at
the ufiddle. They tra\erse the cavity of the epistome, joining its

114 A. OKA.

underside with the ceiliu^^. As seen in cross sections, they are more
closely set nenr the edg-e and abnost entirely wanting in the central
part of the epistouie.

6. Nervous System.

This system has been described more or less fally in all works
on Polyzoa, but the accounts given are very difterent from my own
observations. Xearly all investigators describe the cerebral ganglion
ns a solid cellular mass. Nitsche, studying the process of gemma-
tion, states that the ganglion has at first a ventricle, which, however,
obliterates with the growth of the animal. Contrary to this state-
ment, Saetftigen (10) recently discovered that in Gristatdla and
Pliuiiatclla, the cavity of the ganglion persists throughout life, and
further that the ganglionic wall is not everywhere of the same
thickness, being at some parts as thin as the lining epithelium of the
body- walk I have observed that in rcctinatdla also the cavity
exists in the mature state ; it is so very large that at first sight it
might be mistaken in sections for a part of the body-cavity.

In tig. 28, PI. XrX, I have represented the form of the ganglion
in Pedinalella (jclatiiwsa. It may be compared with a spindle bent
somewhat in the fjrm of U, and fitted with its concavity to the anal
side of the œsophagus, in rather an oblique position with the arms
turned slightly upward. The end of each arm again makes a sharp
bend in the anal direction and is continuous with a large ner^•e trunk
which proceeds into each hjphophoral arm. The ganglion is in
direct contact with the inner cell-layer of the œsophagus, the outer
layer of that tube enveloping it on all other sides ; the ganglion is in
fact situated between the two layers of the œs(.)})hagus (fig. 29, PI.
XIX). The lophophoral nerve trunks are likewise located between

OBSERVATIOX OX FRESH-WATER POLYZOA. 115

the outer and flic iiinei- cell-layers of the bod 3^- wall ; they run,
namely, immediately beneath the outer layer of the lophophoral ceil-
ino-. covered below by the lining epithelium.

As mentioned above, the ({anodion is not a solid cell-mass as
has been described by nearly all investigators. On the contrary, it
contains a spacious ventricle, extending to the end of the arms, or
horns, as is diau^rammatically shown in fig. 30 a, h, c. PI. XIX. The
wall of the ventri(de is very thin and of an epithelial nature on all sides
except at the bottom somewhat on the anal side, ^vhere it is verv
thick, forming the ganglion sensu strict a, — a condition which reminds
us of the ïeleostian cerebrum.

This thick portion is distinctly bounded from the thin epithelial
pai-t of the wall, and is well seen in the fre.sh state as a somewhat reddish
mass, with a slight constriction in the median plane of the polypide.
It is this part that Hyatt took for the ganglion which he describes as
comp(^sed of two lateral masses united by a verv tliick commissure.
It is no wonder that he overlooked the thin epithelial portion, since
this is hardly recognizable in surface views. As can readily be
imaodned by combininçf the three sections i>dven in fio-. 30, passino-
through the brain in different directions, the thick portion is a
transversely elongate rounded mass, with a transverse slit-like de-
pression, looking orally and upward. The whole mass is not of the
same structure throughout, but shows a differentiation into peripheral
and central portions. In the former, the nuclei (of ganglion-cells)
are densely crowded, while in the latter we see a faintly stained
granular mass (Punktsubstanz) containing only a few or no nuclei.
Tlie cell outline to each nucleus is not to be seen.

The thin part of the wall of the ventricle differs in nothing from
ail ordinär}'' epithelium, being composed of a layer of flattened cells.
It is c«intiniious with the peri[)heral portion of the proper ganglionic

IIG A. OKA.

part. ITow the nerve fibres, if there he any, pass ont from the latter
into the nerve-trunhs, I have 1)een unal)le to ehicidate.

Tlic cross-section of the lophoplioral nerve trunk is kidney-
sh;ip('(h with the concavity tnrned above (fig. 31, PI. XIX. nrrr(').
In it tlie nuclei of nerve-cells are seen mnch crowded. Longitudinnl
sections show that the nerve-cells in cpiestion are spindle-shajx^i
(bipolar) with the nucleus at the middle, and closely packed tog-ether.
A few fil)rcs run amongst them ; these are probably to be regarded as
nerve-fibres. The trunks themselves are very thick and large
in companson wifh the mass of the central ganglion, and their
structnre gives the impression of an elongated ganglionic mass rather
than of a nerve. The trunk gives off on each side a branch into each
tentacle. Such a branch is of fibrous appearance and could be
traced onlv for a very short distance after its de])arture fi-oui the
trunk.

The presence of a circumcesophageal nervous commissure in
fresh-water Polyzoa is a matter of obscurit}^ having been accepted by
a few and denied by many. My observations on Pect, gdathinsa
convinced me of its absence.

T'he colonial nervous system present in many marine Polyzoa,
wliich keeps the action of the members of a colony in har-
mony, seems to be altogether wanting in this species, as is probably
the case in all other fresh-water Polyzoa. Special attention to
tliis point showed no trace of nervous connections between the
polypides in ])reparations of sectioned colonies. The fvct agrees with
the behavior of the p(^lypides in a living colony, in which only
directly disturbed polypides retract, while all the rest remain pro-
truded as if nothing had liappened.

OBSEP.VATION OX FRESH WATER, POLYZOA. 117

7. Ovary and Testis.

Pcct'niafdhi qchitinn^a has a distinct ovary, although it devolopes
only in verv rare instances. AV'hen present, it is situn,te<l inside the
cystid near its tip on the oi-al side. It is a solid clnh-shaped
ontoTowth of the internal lining epithelium, and nsnally contains ten
or more ova in different stages of development (fig. 33, PI. XIX),
the space between them being filled with connective tissue stroma.
Ripe eggs can fall into the perigastric cavity only by the rapture of
the ovarian wall. The largest ovarian (ninn measured 0.35 mm.
The length of the ovary is about 0.9 mm., and the breadth 0.5 mm.
Xo doubt can ever be entertained about the ovarian nature of the
b(jdy in question. Tiiat the funiculus has nothing to do with the
production of eggs has also been ascertained by Braem f)r Crisfafella
and PlumateUa.

As to the testis, my investigations gave no result. I searched fu-
it in hundreds of polypides, but in vain. I once saw son^.ething like
spermatozoa within the tip of a cystidal branch, l)ut I filled to make
it sure. At any rate, true sexual organs are very imperfectly deve-
h^ped in accordance with their secondary importance in the reproduc-
tion of this species.

8. Funiculus.

The funiculus is a hollow tubular organ, about 5-6 mm. long,
whieh Connects the blind end of the stomach with an o))posite })oint
of the cystidal wall. Its wall is composed of three layers, but the
innermost one, consisting of a few longitudinal muscular fibres,
hardly deserves to be called a layer (fig. 32, PI. XIX). The outer-
most layer is the continuation of the outer lining of the alimentary
canal or the lining epithelium of the endocyst, from either of which

118 A. OKA.

it rliffer.s in nothing. The cells of this layer rest on the outside of a
tiil)e of basement niem1)rane, which forms the middle layer, and are
rather thickly set, every cross section of the tube showing from
nine to twelve nuclei. Thus my observations on this organ are
identical with and only confirm Nitsche's. Verworn denies the
existence of muscular fibres in Cristatdla. In FectinatcUa they are
decidedly present, although few in number and isolated, so that tliey
are liable t<.) be overlooked if not specially searched for. The outer-
most layer is the only cell-layer in the wall of this organ. I cannot
l)ut assume that Kra?pelin had fallen into error in describing the
funiculus as made up of two cell -layers, the ecpiivalents of the outer
ln\er and the lining epithelium of the endocyst respectively.

The diameter of the lumen is about 0.02 mm., and the thickness
of the wall about half as much.

The narrow lumen of this tubular organ, whose wall must be
ivgarded as entirely mesodermal, is bounded at its upper end by the
inner cell-layer (entoderm) of the stomach, and at the lower end, by
the (Miter layer (ectoderm) of the endocyst (figs. 33 and 34, PI. XIX).
It is in this organ that the statoblasts are developed. That the
funiculus should not be reg^u'ded as tlie ovary, as was done l)v some
former investigators, is self-evident, at least in the present species as
well as in those in which a distinct ovary has been demonstrated
in quite another region of the.liody.

9. The part of the Endocyst that
produces buds.

Budding takes place at a, certain fixed position as Braem asserts,
namely, at a definite area on the oral side of the cystidal endocyst.
Here the endocyst is somewhat thicker than other parts of the same

OBSERVATION OX FKESH-WATEK I'OLYZOA.

119

wall, and the outer cell-layer and the lining epithelium are clciirly
distinguishable from each other, a.s at (jther places, although no
muscular layer intervenes between them. At this place, the cells uf
the outer layer are wanting in vacuoles and both layers stain more
deeply than anywhere else. The area is comparable to the growing
point in plants. How the buds arise, shall be treated under a s[)ecial
chapter later on.

B. Reproduction.

In fresh- water Polyzcja, reproduction may take place sexually or
asexually in three ditferent ways, as tabulated below :

[Reproduction by

Xature :
Function

iSTevv individual
originates from :

The number of
body-layers that
enter into the
formation of the
new individual :

1
Ovum
sexual

2 I 3

Statobhist BuddinjT

asexual

to form primary poly_
zooid irivinii' rise to
a new colony.

to form a nundjer
of new polyzooids,
thus increasing the
extent of a colony.

one
cell

one

many cells

two.

13y the first nuxle, an ovum should undergo segmentation, and
passing through a series of metamor[)hosis give rise to a new primary

120

A. OKA.

polyzoöid. This mode, however, heems t(j take place Nerv rare]y, if
ever, in the present species.

By the second mode, germs enchjsed in hard chilinous cases
(statoblasts) are produced in the funicuhir cavity of polyzoöids.
They are set fi-ee by the decay of the parent cohjny, and float on the
surface of water during" winter, sometimes packed in ice. Next sum-
mer a primary polyzoöid is developed in each, serving as a foundation
for a new colony. Thus, this and tlie first mode perform tlie same pur-
pose, in so far as both serve to establish new colonies, and the with-
drawal of the latter is supplanted l)y the great activity of the foriner.

By the third mode, a certain part of the endocyst adds, by
growth in a certain definite manner, new polyzcjüids to the primary
polyzoöid. This mode of reproduction increases the size and de-
termines the form of the colony.

In certain cases the colony may propagate itself by simple divi-
sion. For instance, Allman and Hyatt observed that in old colonies
of such genera as CV/6'f((f6'//(/, Lophopus iiuû FectiiiatcU't, all of which
have gelatinous ectocyst, the branches separate themselves from the
cœnœcial trunk by constriction. In Feet, (lelafinosa, however, I
have never met with the same phenomena. On the contrary, all the
("olunies collected by me shc^wed no sign of such fissiparity, all of
them being entire and of the form characteristic to this species. In
most of them, the shell-halves of the statoblast in which the primary
polyzoöid has developed were seen sticking to the underside some-
where about the centre.

With regard to the first mode of reproduction, I had no chance
of making observations any further than determining the j)resence of
ovaries in certain polyzoöids. The phenomena of reproduction by
the remaining two modes shall be treated, for sake of convenience,
under the followini»' four heads:

OBSEKVATIOXS OX PKESH-WATER POLYZOA. 121

1, StatobJiist,

2, Development of the Statobla.sl in the Funieulus,

3, Devel(.)pnieiit of the Polypide in the K^Slatohlast, und

4, ]j Lidding.

1. Statoblast.

The general strueture of this seed-like body, dilfei'iug in shape
and size in different species, is now well-known and the following
description refers specially to the statoblast of I'cct. gdatiiiusa. In
winter the dead colony is soon decomposed and the statoblasts con-
tained in it are set free. Dm'ing winter and spring months, they may
1je found on the surface of the water in large numbers, clinging to float-
ing logs, Ijamboo sticks or trunks of aquatic plants. They are of a
dark brownish color with a wide marginal zone of a lighter tinge.

Let us take one of them and examine it more minutely. Its
s]i:i[)e is, properly speaking, like a flat lens. The (jutline, as it lies
flat, is quadrate-oblong, about 1.5 x 1.3 mm., and about U.3 mm. in
thickness. I may here mention that this s[)ecies has the largest
statoblast among all known Phylactolcomatous Polyzoa. It presents
double curvature after the manner of a saddle (fig. 5, PI. XA^II). For
conveniences sake, we may call that side on which the longer axis is
convex as the "conv^ex surface," and the opposite side as the ''con-
cave surface," although tliese terms do not hold good with regard
to the shorter axis. On both sides, the whole surface is beautifully
marked into hexao'onal areas, more distinct in the marginal zone
than in the central portion. The extent of the central darker area
is various in different statoblasts, and it may also differ on different
sides of the same statoblast. Generally it ranges from 0.5 mm. to
0.6 mm. in diameter.

122 = A. OKA.

Closer examination shows that what appeared as a distinctly
reticuinted marginal area is a sort of broad rim around a cliitinous
body of compact nature. This rim consists of a nundoer of prismatic
caskets filled with gas, the diameter of the caskets increasing as we
approach the margin. The hollow caskets have their axis vertical to
the plane of tlie statoblast and are arranged in two horizontal lasers.
They serve as a buoy to float the central body, which is the most
important part of the statoblast.

If the free edge of this rim, or the annulus as it is called, be
examined under a strong power, we see a great number of minnte
hooks projecting from it (fig, 35, VI. XIX). They are found most
abundantly where the margin is somewhat angular. Soine of them
are complex, while others are simpler, but all are fijrmed by the
combination of simple hooks in various ways. They are mere out-
li'rowths of the ed^'e of the annulus, and have no direct connection
with the central body, as is the case in Cristatdla and Pect,
magnißcd. They are short and stout, and the tips are rounded.
They measure about 0.02-0.03 mm. in length and are too minute to
be of much functional importance. When the annulus splits hori-
zontally, as it does of itself when the polypide begins to develope within,
these, spine.5 are found onl}' on the margin of the concave side.

The curvature <jf the annulus and the presence of hooks on the
free edge seem to be worth carefid consideration. In :dl statoblasts,
tlie annulus serving as buoy p)erforms an eiiectual service in dis-
tributing the species as well as in enabling the establishment of
colonies near the surface of water. Where the annulus shows
curvature, the distributive power is evidently enhanced, ex-
posing its curved surface to the influence of the motion of water,
or, if dried up, of wind. As to the hooks I have no doubt thîit,
as Kraepelin has said, they serve as anchors to secure attachment for

OBSERVATIONS OX FJiESH-WATER POLYZOA.

123

the colony that is to grow. At the same time they must be looked
upon as assisting- (listril)ution to a great extent. B}^ their means,
foT instance, the statohhists have a chance of clinging to the feathers of
some water-birds or to floatmg logs or weeds and of being carried away
to distant localities. The strongly developed hooks on tlie statoblasts
of Cristntdla or of Pect, magnißca, in which the annulas is but
weaklv developed and cannot serve more than as a mere buoy, may
perhaps have in this respect a great importance. In the present
species, the extreme insignificance of the hooks as distributing organs is
probably sufficiently counterbalanced by the extensively developed an-
nuhis with its double curvature, so marked a feature of this as compared
with all other species. Braem regards the hooks as protective organs,
but as such they can have no great value in the cnse of Feci, fiel atinosa.

The annul us, as studied on sections (fig. 46, PI. XIX), is made
up as usual of two h(^rizontal strata, each consisting of a single
layer of upright hollow prisms which -remind us of cells in a honey-
comb. The central part shov^dng indistinct reticulation in sur-
face-yiews proves to be the exposed surface of a thick chitinous
capsule of spheroidal shape (centr. caps.^ This central capsule is
made up of two watch-glass like valves tightly apposed with rims,
the demarcation between them being visible as a faint line. The
wall is composed of two distinct layers of chitin, which maybe called
the outer and the inner stratum respectively. The outer is darker
in color, and by far the thinner of the two. This stratum is the
continuation of the chitinous wall of the annulas, and its exposed
surface is raised into l<nv ridges that form a network with hexagonal
meshes.

The thick inner stratum of the chitinous capsule looks bright
yellow on sections. Directly beneath this capsule, there is a mem-
branous envelope (r//ü. /^?^'//;/;y.) distinctly composed of flat hexagonal

124 A.OKA.

cells with centrally situated small nuclei. This cellnlar envelope
completely encloses a granular mass of protoplasm Qpyin. mass.) in
which are scattered minute nuclei. These nuclei measure only
0.001x0.003 mm. in average, and are thus several times smaller
than the nuclei of body tissues. They are very flat witli their plane
parallel to that of the statoblast.

The granular contents and the cellular envelope form the
essential part of the statohlast, while the chitinous cnpsnle, the an-
nul iis and the marginal spines nre all accessory organs for its
preservation and distribution.

2, Development of the Statoblast.

The knowledge of the origin of statoblast is certainly of vital
importance in determining the true nature of this gennnule-like
body, but in the rather scanty literature on this subject the state-
ments given are widely different from one another. As to my own
observations, I have seen in tlie lumen of the funiculus sometimes a
single cell and at other times a loose group of two or more cells,
representing the earliest st;!ges of development of statoblasts. They
are refund in outline, and each supplied with an oval nucleus, j^either
in size nor in general appearance do they perceptibly differ among
tliemselves, ov from those of neighboring tissues. This circumstance
deprives me of all grounds to share Verworn's view that the increase
of cells is (hie to continued division of an originally single cell. This
author sums up the earliest steps in the development of a statoblast
in the following words : An einer bestimmten Stelle des Funiculus
vermehren sich die Epithelzellen desselben zu einer kleinen Auf-
schwellung and drängen dadurch gegen das Lumen. Eine Zelle davon
tritt in das Lumen hinein and wird zur Eizelle, während die anderen

OBSERVATrOXS ON FRESH-WATER POLYZOA. 125

sich ZU einen Follikel formiren. Die Eizelle maclit einen reo-elimiss-
igen Fui-clinng8proces.s durch, dessen Resultate eine solide M(^ruhi
ist. Wie man sieht, wird also auch durch diesen F urchungs Vorgang
die Knospennatur der Statoblast widerlegt." Hence he concludes :
"Die Statohlasten sind als parthenogenetische Wintereier aufzufjis-
sen welciie sich im Gegensatze zu befruchteten Eiern am Funiculus
entwickeln." I did not find this view corroborated l)v facts.
Neither the thickening of the funicular epithelium nor aii "Eizelle,"
which to judge from his figures must have beeii several times lare'er
than any ordinary cell of the funiculus, couhi be found.

On the contrary, what I have seen in PectinateUa fidatinom
leads me to the conclusion that each statoblast originates from at
least eight cells of separate derivation. Where the}' come from is a
qiiestion which I cannot answer from direct observation. However,
that it receives no element from the entoderm is evident from the
fact that where there are many statoblasts in the same funiculus, the
older ones always lie nearer the stomach, completely shutting np the
]iassage. The question then reduces itself to whether the original cells
are derived from either the funicular wall (mesoblast) or the (mter layer
of the endocyst (ectoblast) at the point where it bounds the funicular
cavity below, or from both. As will be seen later on, the intrastato-
blastic development of a polyzooid essentially agrees with the process
of development liy budding, differing only in such points as are
necessitated by the mechanical conditions of each case. Wes hindd then
expect similar elements in the " anläge" of a statoblast as in a bud. that
is. both the funicular wall and the outer layer of the endocyst should
a priori give their contingents to form a statoblast. The correctness
of this assumption is proved by the observations of Braein (Zool. Anz.
1889.). According to this author, the primitive statoblast consists
of two kinds of cells, which are genetically different, one deriving

126 A. OKA.

itself from the fiiniciilas and the other from the ectoderm. It is
needles,^ to say that in the aho\e light, a statoblast cannot he
a?)y thing else than a specially modified form of hud, in other words,
a t)ortion of both layers of the endocyst protected against severe
climate by special contrivances for the preservation of the species.

But to return to the process of development, a certain nnmber
of cells, probably from tlie two sources referred above to, assemble in
the funicular lumen and arrange themselves into a group at first loose
and irregular. During this early stage, the funicular wall nowhere
shows thickening, contrary to Verworn's observations. Very soon
the group becomes compact and assumes a morula-like form. It can
now be safely asserted that new additions of cells no longer take
place, but that the morula henceforth increases in size l)y multjplic^a-
tion of its own cells. The mass bulges out the funicular wall as it
e id arges.

Arrived at a stage when the morula measures about 0.05 mm.
in diameter, a certain nnml)er of cells (8-12 as seen in equatorial
sections) on one side of it form a special group (fig. 38, PI. XIX),
at first very indistinctlv distinguishable from the rest of the cells.
Graduallv, a small cavity appears in the rentre of that 8[)herica!
group of cells which are steadily increasing, changing it into a
hollow, rather flattened sphere with distinct epithelial wall. This
hollow sphere is the " cystogene Hälfte" of German authors, so called
on account of its giving rise to the chitinous covering of the stato-
blast, and the remaining mass of cells constitutes the " Bildungs-
masse." x\ccording to my observations, these two portions are not
morphologically distinguishable from each, other at a very early stage,
but become secondarily distinct. This is also the view held by Kitsche
and Yerworn, while Braem saw them originate sharply separated
from the outset in Cristatella. According to the last-mentioned

OBSERVATIONS OX FKESH- WATER i'OLYZOA. 127

author the cystogenous sphere, which consists solely of cells
of ectodermiil origin, is the first to form and to this is added
the Bildungsniasse by proliferation of (mesodermal) cells of the funi-
cular wall. Provided that in either case the two [x^rtions are respec-
tively ectodermal and mesodermal products, it would be of but
secondary importance whether they are distinct from the beginning
or become outwardly indistinguishable for a time. More study of
this point is exceedingly desirable.

Further history of tlie development corresponds in the main
with what is already known. The cells of the two portions are con-
si'antly increasing in number and the entire mass in size. Meanwhile,
the cystogenous cells attain the character of columnar epithelium ; the
whole cystogenous sphere flattens, and s(-on takes the form of a shallow
watch-glass, the internal cavity disappearing (hg. 40, PI. XIX. ajst. c.)
We may speak of it as the cystogenous cu|). The C(jncavity of the
cu[) grows deeper, always closely clasping the mass of the remaining
cells, i.e., the " Bildungsmasse." The cells of the latter begin to
présenta granular a[)pearance by the deposition of refractile «pherules
in the protoplasm, comparable in nature to the deutoplasm of eggs
or of 3olk-cells in Plathelminthes. Braem could not convince him-
self of the truth of Nitsche's and Verworn's opinion that the granules
are direct products of the nuclei ; nor cou.ld I lind any sup[)ort to this
view. x\1jout this stage, the cells in cjuestion assume a spindle-;?hape,
the axis standing vertical to the cavity of the cystogenous cu)) (tig.
o'J, PI. XIX, gr. m.). This state was also noticed by Braem in Cris-
tatella. H<3wever, as the granulation advances, they become rounder
again, until each cell is represented by a globular mass oi' granules
with a nucleus at the centre (fig. -il, PI. XIX, gr. in.).

As the cystogenous cup grows in size, its rim begins to close
around the granular cell mass. This occurs after the latter has

128 A. OKA.

almost attained its maxiuuun size. In the nieaiitime, a tliiu sheet
of chitin is secreted ])e(\veen tlie two layers of the cystt)geiiOUs cup;
it is difficult to say whether it is the product of one or of both layers.
This chitinoLis slieet subsequently attains considerable tliickness.
We may speak (jf it as the chitinous cup, as it has that shape along
with the cystogenous cup. As the latter expands, its mouth narrows
and the whole body of the youiig statoblast somewhat flattens,
taking the form of a spheroid, the axis of whicli corre^|)onds
with tliat (jf the cystogenous cup. The two layers of the cystogenous
cup were at first of the same thickness, but now the outer begins to
thicken by the increase in heiglit of its cells while the inner undergoes
a contrary change. The cells of the latter begin io flatten tirst at
the bottom-portion of the cystogenous cup.

Along the equatorial hue of the spheroidal mass, the outer layer
of the cystogenous cup is thrown into a fold, whicli encircles the
young statoblast belt-like all around. The belt becomes more and
more extensive, and consists, as seen in sections, of two closely
opposed strata of cylindrical cells.

Meanwhile, a second chitinous layer is formed over the chitinous
Clip already present. Thus, the chitinous cup c<jmes to consist of two
layers ; the outer of which is by flir the thinner. Simultaneously and
directly continuous with this outer chitinous layer a thin plate of the
same nature is also deposited between the two e[)ithelial strata of the
belt. It may conveniently be designated the belt plate.

The elongated prismatic cells of the outer cystogenous layer,
secrete around their basal ends thin chitinous wall continuous with
the belt plate or the outer chitinous cup, on which they all sit. I'hey
thus bring forth hexagonal caskets open at one end, into which every
one of them abuts with their bases. But the wall of these cells does
not develope everywhere to the same extent. It keeps very low on

OBSERVATIOXS OX FRESH- WATER POLYZOA. 1^9

the exterior <jf tlic bottom of the cy.stogenou.s cup, aiid when the cup
closes into ÎI complete capsule, tis it does later, the same condition is
also seen on the opposite side, so that on a mature statoblast the polar
surfaces show only a network of very low ridges. However, on both
sides of the belt plate and of the capsular surface immediately adjoining
them, the chitinous wall of cells attains considerable height, but never
reaches the surface of the cell-layer. The open ends of chitinous tubes
thus formed are finally closed by the formation »jf what is called the
lid-plates. This process proceeds on the one hand centrifugally from
the outer layer of the chitinous cup, at a line which circumscribes the
reticulated polar area, and on the other in the opposite direction starting
from the marginal edge of the belt-plate, so that the tubes on the
midway are closed last. A glance on tigs. 48-J:6 will make the
matter clear at once. j\Ioreover, the lid-plates divide the prisniîitic
cells on either side of the belt-plate into an outer and an inner \)uv-
tion. The latter is completely enclosed in chitinous caskets, while
tiie former conjointly with the epithelium covering the polar area
invests the entire outer surface of the young statoblast. This invest-
ment is to be seen as long as the statoblast remains at the place of
its development, but decays when the latter is set free by the dissolu-
tion of mother-polypides. As the lid-plates are developing, the
marginal spines appear. Also at about this stage, the closure of the
mouth of the cystogenous cup takes place. It thus completely
encloses the granular cell-mass, followed by closure of the two layers
of the chitinous cu^), which then is turned into a perfect capsule.
After this, the two polar areas present no point of structural
difference.

The portion of the prismatic cells that are enclosed within the
chitinous wall soon undergoes decomp(jsition and gives place to a gas
filling up the caskets. Thus the formation of the swimming- belt is

130

A. OKA.

complete. Nitsche'.s statement that the cells evacuate the caskets
before their closure is probably an error.

As already said, the inner cystogenous layer thins out by the
flattening of its cells, and when the chitinous plates completely
inclose the granular mass it forms a thin epithelial covering to the
latter directly within the central capsule (figs. 42-46, PI. XIX, Enü.
m.'). The size of the nuclei becomes smaller as the height of cells
decreases, and reaches at last the dimensions given before wdien
the niatiu-e statoblast w-as described. The cells of this membrane are
distinctly bounded and hexagonal in sha})e.

Keturning to the stage represented in fig. 41, the granular
spheres, composing the mass contained within the cystogenous cup,
have each a centrally placed nucleus, and growing larger (fig, 42,
PI. XIX.) press upon one another so that they assume a polyhedral
form. P'lie}- remain distinctly bounded as long as the rim of the
chitinous cup remains open, but fuse together after the latter closes.
It is a singular fact, that in some statoblasts, either the granular
mass is produced in over-quantity, or the capsule formed is too small,
so that a portion of the mass is left outside the statoblast as the cap-
sule closes, afterwards disappearing.

The nuclei of the granular mass become smaller as the develop-
ment of the statoblast advances. Arrived at a stage represented in
tig. 42, PI. XIX, the nuclei almost lose their peculiar chromatin reac-
tion, and stain very faintly, so that in some preparations it is very
ditiicidt to detect them. This condition, however, lasts for a very
short interval, and in all the later stages the nuclei are again distinct-
ly visible. This [)eculiar behavior of the nuclei may have lead
Verworn to assume that the granules are the product of the splitting
of nuclei and that the latter as such are not found after the com-
plete development of the granules.

^ OBSERVATIONS OX FRESH-WATER POLYZOA. 131

The statoblast at the earliest stage of its (levelopniont is of
a milky- white color. The chitinoiis parts as they form tliemselves at
first present hght 3'ellow color, which, as the development advances,
darkens to the characteristic hue of the mature stato1)last.

On attaining a certain size, the statoblast bulges out the fimicidnr
wall chiefly on one side, with its plane always parallel to tlie axis of
the funicidas. When many statoblasts develope in the same fanicuhis,
tliey generally lie alternately disposed, by wliich means econoni}- of
space is effected. It is on that side of the statoblast with whicli it
joins the funicular tube that the cystogenous cup closes.

The number of statoblasts that develope in a single polyzooid is
usually five or six, in some cases as much as eight. Of these, the
uppermost one is the oldest and the lowest the latest formed, so tliat
at a certain period statoblasts in various stages of development in
serial order may be seen in the same funiculus. In those old poly-
pides that occupy the central part of a colony all the statoblasts usu-
ally attain maturity, while in the peripherally situated younger poly-
pides the latest formed statoblast is generally still in c|uite an early
stage of development at the time when the colony begins to dissolve
away. These immature statoblasts undoubtedly suffer common decom-
position with the mother-colony. As every polyzooid produces
statoblasts, their number in the entire colony is really very great.
Once I counted no less than 870 statoblasts in a very small colony
of about 1.5 cm. in diameter.

132 A. OKA.

3. Development of the Polyzoöid
in the Statoblast. *

As the mature statoljlast floats on the surface of water, the belt-
plate of the annuhis splits horizontally, so that tlie shell may now l)e
said as being composed of two valves. These however remain tight-
ly apposed during winter. On the arrival of warm temperature, they
separate from each other, l)ut holding the whitish contents between.
The two valves have then very much the a[)pearance of a pair of
cymbals. The separation takes place at a stage when no change is
yet percejitible in the contents ; hence I am inclined to ascribe its
cause to some external influence rather than to internal pressure.

The contents of the statoblast, i, e., the granular mass with its
enveloping epithelium, form a spheroidal mass. All along the outer
marö-in or the equator of the spheroid, where the separation of the
sliell-valves has brought it in direct contact with water, the enveloj)-
ing epithelium becomes thicker (figs. 48, Out. lay. and 48 A, PI. XX.),
owing to increase in height of cells, accompanied by great increase
in size of tlie nuclei, which are now as large as those of grown-up
polypides. The process of thickening thus begun at the equator
proceeds gradually toward the two poles of the spheroidal mass, so
that the membrane (hickens latest at these places.

Meanwhile, the cells at two opposite areas on the equator
become especially taller, so that the enveloping membrane acquires
a marked thickness at these places. The areas in question are oblong

* Aftoi- finishing the manuscript of this paper, I received No. 324 (1889) of the Zoolo-
gischer Anzeiger containing Braem's preliminary report entitled " Die Entwick]. d. Urj-ozoen-
colonie im keimenden Statobl." His statements differ in many fundamental points from mine.
There is sufKcient ground to assume that very considerable variation of development obtains
among different speci« s of Poljzoa.

OBSERVATIONS ON FRESH-WATER POLYZOA. 138

in shape, lying with tlicir long axes along the equator, ahhoiio-h no
sliarp hoiindary can be fixed. From an earlv stage, tliev show difFer-
iMicv.s in tlie apj)eai'an<3e of theii" cells and lake quite diiterent dii'ec-
tions in their future develo|)nient. The axis jc^ining the centres of
these areas corresponds, as Avill be seen later on, with the longitudinal
axis of the future polyzooid. With regard to the relation of this axis
with tlie longer or shorter axis of the statoblast, there seems to be no
constant rule, although in the majority of cases the former cor-
responded with the shorter axis of the statoblast.

In one of the two areas, the cells acquire distinctly cylindrical
form, and vacuoles are formed in some of them. In f ict. they soon take
the form and character of the cells of the outer layer of tlie endocyst.
They liegin to secrete gelatinous ectocyst of a sticky nature, by which
means the germinaling statoblast attaches itself to anything it may
meet with, be it the wall of an aquarium, floating wood, or shells of
other statcjblasts.

The either area gives rise to the polyzooid. Its cells are of less
height and vacuoles develope in them later than in the other area.
At alx^ut the middle point of the area, the cells nuilti[»ly, and a group
of them sinks into the granular mass below, forming a solid club-
shaped b(Mly, which a little later on becomes Ik^IIow by the retreat of
its cell« toward the periphery. We have now a hollow closed sac
bounded by an epithelial layer of cells and connected with the super-
ficial thickened area by means of a very short solid stalk (fig. 40, PI.
XX). Soon after, the latter also acquires a limien, and the cavity
(fig. 49, PI. XX. prim. /.) of the hitherto closed sac comes to com-
municate with the exterior. Some cells of the granular mass lose a
part of their gi-anules, and arrange themselves hito a sort of layer on
the outside of the sac (fig. 49. PI. XX. Lin. I'pifliy Tlie nuclei of
these cells Imtotiic lai'g(M' as the granules lessen in (juantitv. and

134 A. OKA.

appronrh those of , ordinavy cells in size and appearance. The outer
limit of this layer is by no means definite, gradually losing itself in
the granular mass.

As the sac elongates, it becomes constricted at the middle, divi-
ding into an outer and an inner chamber. The constriction be-
tween the two chambers is the future mouth, and the inner chamber
represents the future oesophagus and the stomach. The outer cham-
ber soon ac<|uires the form of a hollow cone, at the base of which the
mouth opens and which tapers towards the outer opening. At the
base of this conical chamber the epithelium is especially thickened and
eventually gives rise to the lophophore and the tentacles, the chamber
itself being the tentacular sheath. The investing layer derived from
the o'ranular cells (lin. epitli.') become more and more conspicuous,
and lines the entire outer surface of both chambers.

The lophophore is at first a semicircular ridge, clasjông the
mouth on that side which corresponds to the original bottom of ihe
cystogenous c ip (convex side). The ridge arises by the folding of
the wall, in which process both layers arc concerned. The ends oi"
this semicircular ridge are prolonged in the form of free finger-like
processes, the rudiments (^f lophophoral arms. The interior of the
h>phophoral rudiment is occupied by the granular mass as soon as it
is formed. The developing polypide lies on its anal side when the
statoblast is placed on its concave side.

Another constriction divides the lower chamber into the oeso-
phagus and stomach. The stomach begins to send a hollow process
upward to form the intestine (fig. 51, PL XX. Intest.).

The free edge of the lophophoral rudiment is divided into a series
of knobs, which are conspicuous nearer the median line, becoming
gradually smaller towards the tips of the arms. These knobs are the
origin of the outer row of tentacles. In the meanwhile, a second

OBSERVATIONS OX FKESH-WATEK i'OLYZOA. 135

ridge runniDg parallel with, but less extensive than, the first one,
developes on the anal side of the mouth. Its extremities soon meet
and fuse together with the limbs of the first-formed semicircular ridge.
Tentacles are formed on the new ridge in the same way as described
above ; the range of their row extending on either side to tips of
lophophoral processes. Thus the inner row of tentacles is established
on the lophophore.

The hollow process sent up by the stomach grows larger, and
finally its cavity opens into the u]>per clinmber or the tentacular
slieath, Avhich, when evaginated, forms the tubular body of the
polypide.

The account given above may suffice to show how the general
shape of a polypide is formed in the contents of a stntoblast. In the
meantime rudiments of many other organs, of which the brnin, the
muscles, and the funiculus are the most important, have begun their
development.

The cerebral "anolion arises as a pit-like invaoination of the inner

CO i. ~

layer of the oesophageal wall, which is continuous with tlie outer
layer of the body- wall. The process begins to take place at a stage
when the stomach sends up the process that afterwards becoines the
intestine, on the anal side of the oesophagus, just inside the mouth.
The invnjjfination is soon constricted off, turning" it into a closed sac,
wliich as it is being formed, carries with it the outer layer of the
oesophageal wall, so that the latter invests it externally, at the same
time connecting it with the oesophagus. The cavity of the sac per-
sists as a sort of ventricle. The lower portion of the wall of the ^ac
early begins to thicken, which process does not of course concern the
investing layer, and finally developes itself into that portion which
constitutes the main ganglionic mass (vide p. 115). The remaining
portion of the sac- wall, except at two points, becomes thinner and

136 A. OKA.

tliiiiiier a.s tlic entire j^anglion iiicieiise« iu size. The two exceplional
points just referretl to, are wliere the sae-wall produces a pair of solid
liorn-like processes, each of which gradually elongates towards tlie ti[)
of the lophophoral arms, passing between the two layers of their ceil-
ing. The position of the lophophoral nerve-trunks directly beneath
the outer layer led me at first to assume their origin from the latter,
in a way anrdogoiis to the development of the central nervous system
in \ertebrates. A careful study, howevei-, convinced me that such is
not the case.

At the time when the intestinal cavit)' becomes continuous with
the exterior at the anus, the whole body-cavity is still tilled up with
the ü'ranular mass. Some of the cells of the latter are seen to ditfe-
rentiate themselves from the rest, at two regions as seen in a median
sagittal section (fig. 51, PI. XX.), the one extendiug between the
involuted tentacular sheath and the cystidnl wall, and the other be-
tween the lower part of the oral side of the oesophagus and the part of
the cystidal wall opposite to it. At these places, the cells lose their
granules, elongate, and become spindle-shaped joining the two pointas
between which they lie. Their further development has been already
treated under the muscular system. The muscles that develope in
the above mentioned regions are the pariet(j- vaginal and the retractors
of the adult polypide respectively. The muscular layer of the endo-
cyst and the alimentary canal developes itself later, prob:djiy from the
cells of the lining e[>ithelium in a similar way.

Almost sinuiltaneously with the first appearance of nuiscles, the
cells of the granular mass lying between the blind end of the stomach
and the coenoecial wall opposite to it, lose a porticjn of their granides,
and aggregate into a solid rod, which is, in sections of stained speci-
mens, readily recognizable on account of the deeper coloring of its
cells in contrast with the surrounding faintly colored granules.

OBÖKUVATIOXS OX FUE«H-WATEU POLÏZOA.

137

Aftcrvvai'cls, what remains oî the granules in these eells is entirely
absorljed, and a linnen is formed inside the rod, converting it into a
tube, the rudiment of the funiculus. Thus, it will be noticed that
both the muscles and the funiculus are produced in sita from the
granular m;iss in the statoblast.

When the devehjpment (jf the polypide is C(jmplete, two buds are
idreudy j)resent on the oral side of the cystid;d wall, one on each side
of the median plane. These buds are first seen in the stiige when the
intestine is still blind. The manner of their development will be
treated under the buddini;'.

As noticed Ijefore, the granulnr cell mass compactly tills u|» the
entire body-cavity until after the formation of all the important organs
of the polypide. The cells then loosen themselves, as the conse-
quence of the decrease of granules, which are being constantly used up,
while the enhanced growth of the cystidal wall makes the body-cavity
more and more spacious. When the young pijlypide begins to
evaginate and expand their tentacular crown, naked conglomerates
of granules, each with a nucleus at the centre, are seen scattered in
the body- cavity. Mixed with these conglomerates, we see some <jthers
which have obtained a distinct wall, with the nucleus pressed against
it. In a somewhat later stage, the granules are no longer visible in
those cells with peripherally situated nuclei ; irjstead of them we see
a large vacuole in each cell, which has thus acquired the characters of
wdiat 1 have proposed to call blood-corpuscles.

It is perhaps worth nijticing th-.it the developing polypide carries
the shell halves on the anal and the oral side of its body, presenting
an appearance comparable to the condition of shells in Brachiopods.

138 _ A. OKA.

4. Budding.

Tliis mode of reproduction in Polyzoa has been studied by numer-
ous investigators, but their opinions ure more or less divided, especially
us to the origin from which the bud receiv^es its h^^poblastic elements,
and consequently, with regnrd to the relations of the germinal layers.
Most of them derive tlie hy[)oblast from the outer layer of the
enclocyst, while a few are inclined to believe that the bud receives it
from the gastric organ of the mother polypide.

According to Allman (1), who describes the process of budding
in Lop]iopm and Alcijonella, the outer la3'er of the end(jcyst gives rise to
all the lining cells of the alimentary canal, while the lining epithelium
of the mother polypide becomes also the lining epithelium of the Ijud.

Metschnikoif (7) gives an account of Ijudding in the embryo of
Alcijonella. He found that after continued segmentation of the egg,
the cells arrange themselves into a two layered hollow sphere, both
layers of which enter into the constitution of the bud, the outer
giving rise to the outer layer of the tentacles, the inner lining of the
alimentary canal and probably also to the nervous ganglion, and the
inner, to the lining epithelium and all the muscles.

Nitsche (<S) studied the process of gemmation in AlctjoneUa
fimgom and Cristatella mucedo. In both species, the wall of the
alimeiittu-y canal is formed from a part of the endocystic invagination
of the mother polyzoiiid. In other words, the lining layer of the
alimentary canal is deri\ed from the outer layer of the body-wall.
Both Metschnikoif and Nitsche regard the outer layer of the endocyst
as the ectoderm and the inner as the entoderm.

Hatschek's (4) acccjunt of l)udding in Cristatella is as follows.
A hollow sac lies directly beneath tlie outer layer, invested by

OBSERVATIONS ON FRESH-WATER POLYZOA. 139

the inner layer of the body-wall on its inner side, at the position in
which the buds are constantly developing. When a bnd is to be
produced, a portion of this sac is constricted off and gives rise to the
inner layer of the alimentary canal, while nil other parts of the young
polypide are formed by an invagination of the body wall. Thus,
the sac is being constantly constricted off, as long as new buds are
added to the coh)ny.

Reinhard (9) studied the first budding in the embryo of Aleyondla
fungom and Cristatella mucedo. The cells formed by the segmontn-
tion of the ovum produce a true gastrula by invagination. Tlic
blastopore, however, soon closes. The gastrula is comparable in all
respects to the type of some other animals, and, therefore, he regards
the inner layer as the entoderm. In the development of tlie bud, the
entodermal cells seem to push into a certain thickened portion of the
ectoderm, and form a part at least of the wall of the alimentary canal.

Salensky (11) also states that the outer layer of the zoœcium
(cystid) gives rise to the lophophore and to the internal cells of the
digestive tract, while the inner layer becomes the lining epithelium of
the new polypide. He believes that the entoderm of the alimentary
canal originates from the ectoderm of the zoœcium.

Haddon (3) who studied the gemmation of some marine Polyzoa,
came to the conclusion, on theoretical grounds rather than from
actual observation, that the alimentary canal is derived from the
entodermic tissue of mother polypides.

To the position and the order of budding, the previous
workers seem to have paid but little attention, except Bra^m who
dwells on the matter at some length. To this author we owe much
of the exact knowledge of the process of budding. As will directly
be seen, the process of budding takes place at certain definite poly-
zoöids and in a certain definite manner, thus determining the shape

140 A. OKA.

of the colony so characteristic for each species. What Er?em de-
scribes for Cristatella on this point does not apply in all its details to
the present species.

At the plnce where buds appear, there is no muscular layer, as
already observed by Nitsche, and the endocyst may here be repre-
sented as consisting- of only two layers, viz. the outer cell layer and
the inner lining epithelium. The latter, in direct contact with the
former, passively follows all the changes in form undergone l)y the
outer layer of the endocyst. So, it must be borne in mind, that
when in desci-il)ino- different stao^es of budding", the changes of the
outer layer (which is the inner layer of the bud as will be seen fur-
ther on) alone are mentioned, similar chang'es are repeated by the
lining epithelium (the outer layer of the bud).

At first, some cells of the outer layer push their way inward in
the form of a solid knol) covered by the lining epithelium (fig. 57. Ph
XX). At a certain observed stage in which the knob consisted
of eighteen cells, many more were on their way of entering.

A cavity ultimately appears in the centre of the knob (fig. 58,
PI. XX.) and the cells arrange themselves regularly around it in
epithelial order. The cavity soon comes to communicate with the
exterior by means of a cajial formed by the gradual retreat of cells at
that part (fig. 5Î), PI. XX.). The bud now represents a double- walled
sac whose inner and oiitei- layers are respectively eontinuations of the
outer and inner la3^ers of the endocyst. Thus it is plain, that tlie.
bud originates not by direct invagination of the two layers of the
endocyst, but by the formation of a closed sac which secondarily opens
outward.

As the bud grows in size, it inclines dowji wards and its oral side
is connected to the crenœcial wall along its wh(^le length by a me-
seiitary-like membrane whi<'h is the continuati<^n of the lining epi-

OBSERVATIONS ON FRESH-WATER POLYZOA.

141

thelinm. A glance at fig. 56, PL XX. will make this clear. The
middle portion of this mesentery-like membrane becomes thinner,
and is finally perforated as shown in fig. 57, PI. XX. The sac is
then joined to the cystidal endocyst at two points, viz. at its opening
and at the bottom. Rndiments of new buds are produced in the region
lying between these two points, which separate more and more from
each other.

The solid rod-like part of the lining epithelium whirh now joins
the bottom of the sac-shaped bad witli the cystidal wall, is the rudi-
ment of the funiculus. It gradually lengthens, and a lumen is
secondarily formed in it, turning it into a tubular organ. It grows
in size, and with the eai'ly appearance of scanty muscular fibres inside
its cavity the development of the funiculus is complete. Thus the
result of my observations on the formation of this organ seems to
agree essentially with tliat of Bra^m (2) wIkj describes the process in
the following w^n-ds : " In der Mediane erheben sich die Zellen des
äusseren Blattes in Gestalt einer an der Oralseite der Primärknos])e
herablaufenden Längsleiste, welche seitlich von den Fortsetzungen
der Magenfalten begrenzt erscheint. Indem sich die Zellen des
Knospenhalses dann nacli vorn umschlagen und an der Bildung des
Integuments betheiligen, löst diese Leiste sich von dem Mutter-
gewebe, welclies hintei- ihn zusammenfliesst, als selbstständige Strang
los und verdindet einen oral und median vor dem Primärknospe
gelegenen Pimct der Leil)eswand mit dem Grunde des Knospen-
Sackes.

A constricti<^n is formed in the middle part of the sac-like bud,
dividing it into two chambers. The constricted opening is the mouth
of the future ])«>ly|>id(', and the lower chamber developes into the
alimentarv canal. The upper (-hamber becomes somewhat conical in
shape tapering toward the orifice of the bud. At the basal disc of

142 A. OKA.

this chamber, where the mouth is situated, the cells of the inner wall
are prismatic while elsewhere they are flat.

We now recognize all the parts that we have seen at a certain
stao-e of intrastatoblastic development. The lophophore with its
tentacles, the nervous system and the intestine, all develope just in
the same way as described in the previous chapter. One im-
portant difference exists in this, that in the one case the lining
epithelium is produced from cells of granular mass, while in the other
it is the result of the increase in extent of the same layer of the
mother cystid. It will be noticed from above statements, that the
entire inner layer of the alimentary tract is derived from the solid
knob sunk in from the outer layer of the endoeyst. The hollow
process (the intestine) sent up from the stomach meets with and
opens into a pit sent in from above outside the tentacular area, on the
side turned toward the centre of the colony. The lophophornl arms
of every individual always project toward the anal side of the poly-
pide ; consequently they are nil directed toward the centre of the

colony.

While new polypides are thus being developed, their cystids are
also growing in size, and some cells of its lining epithelium gradually
o-ive rise to the muscular layer. At tirst, when the young polypide
is still represented by a simple sac, the portion of the mother cœnœ-
cium arouud its orifice is only slightly elevated above the rest of the
wall, but as the growth of the polypide advances, it becomes more and
more prominent, growing in such a manner as to form at last a cell
for the young polypide.

The retractor muscles of the polypide begin t(^ appear when the
bud is still a simple sac, shortly after the formation of the rudimentary
funiculus. At the point of junction of the rudimentary polypide and
the cœnœcium, some cells of the lining epithelium becomes difteren-

OBSERVATIONS ON FKESH-WATEK I'OLYZOA.

143

tinted from the re.st by assuming a spindle-sliape. These cells gradual-
ly separate from their mother-layer and form two loose bundles
which join the cœnœcium with the middle portions of the now two-
chambered bud. The parieto-vaginal muscles also originate in a
similar way, but at a considerably later stage, when the lophopliore
already shows a certain number of knob-like tentacles at its median
portion. Thus, in the process of budding, both the funiculus and
tlie muscles are de\ eloped as ditferentiations of the lining epithelium.
The young polypide. as it first evaginates, is a very pretty little
animal with less than thirty tentacles. The more medianly situated
tentacles are best developed, while they are yet knob-like nearer the
tip of the Icjpliophoral arms, where new tentacles are being added by
degrees.

The buds arise ou the marginal cœnœcial branch alone, on the side
facing away from the centre of the colony, i.e., on the oral side when
we take the [)olypide into consideration. I'hey always develope in
pairs, one on each side of the median plane. Hence the dichotomy
of the cœnœcium, with a polypide-beariiig branchlet at each axil.
The colony as a whole is consequently fan-shaped at first. With
continued budding, it grows toward the periphery, its radius leng-
thening in arithmetical, and the marginal line in geometrical ratio.
The two extremities of the latter soon touch each other in a complete
circle ; after this the growth of the colony throws its marginal line as
well as its hitherto fiatly expanded surface into folds, which make
the regular arrangement of polypides unrecognizable at a glance.

The upper series of diagrams in fig. 62. 1*1. XX, show early
stages in the development of a colony, each circle indicating an
individual. These figures represent for sake of simplicity each in-
dividual as giving off only two buds at a time, and each of these buds
again performing gemmation after some time. In reality, however,

144 A. OKA.

such is iKjt tlie ca«e. On the contrary, we usually see in an actively
budding individual at the margin of a colony, not (jnly buds of the
iirst order but also tlKJse of second and third order already formed,
lîuds of the first order are present, as already stated, in a single
pair, while those of the second occur in two pairs, and the next
order, the most rudimentary, in four pairs. AVhen the buds of the
first order have grown sufiiciently to be regarded as new individuals,
those of the second and the third order occupy the grade of the first
and the second order, while those of the third order arise anew. A
comparis(3n of the lower series of diagrams in fig. 62. with the upper
series will help to make the matter clear. The blackened spots in
the lower diagrams slnnv the gemmiparous porti«jn of the endocyst.
This spot might appropriately be compared with the growing point
of plants. With the growth of the colony, it advances centrifugal ly,
splitting dichotomously at regular intervals. In this way, the colony
oTows as long: as the condition is favorable.

It need scarcely be pointed (jut that the development of the first
polyzoiUd in a statoblast essentially agrees in process and conditicjn
with that of hiter polyzooids l)y means of budding. In fact the first
polvzoöid is similarlv budded oft" from the statoblastic contents, the
whtjle of which is to be seen in the light of a piimary cystid derived
(jf and containing all the essential elements of cystids of the previous
year. Whereas in marine forms the cystids winter as such, those of
fresh-water forms persist only in the form of statoblasts to germinate
in the following year as do the perennial cystids of the former. In
the budding of fresh- water Polyzoa, a cystid and a polypide are
formed simultaneously and an intrastatoblastic primary cystid is to

OBSERVATIONS ON FRESH-WATER POLTZOA. 145

be considered as a particular sort of bnd in which the formation of a
polypide remains latent until the next year.

With regard to relations ef germinal layers in a primary C3'stid,
all the granular cells of the " Bildungsmasse " might with propriety
be called the mesoblast on grounds of their geneyis and of their future
history. For the same reasons, the enveloping epithelium might be
looked upon as the ectoblast except at the growing point, i.e. where
the buds are formed. At this point the cells are still in undifferen-
tiated embryonal condition comparable to cells of a blastula which
differentiates into Ectoblast and Entoblast for the first time at its
invagination. As the colony grows, the growing point of the pri-
mary cystid is split and transmitted into each succeeding bud, very
much like the growing point of a plant ; in other words all the
growing points seen in marginal polyzooids of a polyzoan colony have
started. I believe Bra?m is of the same opinion. Considering, on the
contrary, the outer layer of the ectocyst at the growing point as
strictly epiblastic, the conclusions, to which jSTitsche, Joliet, Salensky,
• &c. were led, that no hypoblast enters into the bud and that it is
formed as a secondary product of the epiblast, are certainly unavoid-
able. But such a conclusion does not accord, as was pointed out by
Haddon, with the generally accepted nature of budding in the animal
kingdom. In my opinion the budding in Polyzoa is only so far ex-
ceptional as the Epiblast and hypoblast take part in an undifferentiated
embryonal condition.

Hn A OKA.

Works referred to.

1. G. J. Allman. — A Monograph of the Freshwater Polyzoa. 1856.

2. Fr. Brœni. — Untersuchungen über die Ijryozoen des süssen

Wassers. Zoo]. Anz. XI. 1888.

3. A. C. Haddon. — On budding in Polyzoa, Quart. Journ. Mic.

Sc. XXITI. 188P).

4. ]j, Hatschek. — Embryonalentwick. u. Knospung d. PedicelHna

echinata. Zeit. f. w. Zool. XXIX. 1877.

5. A. Hyatt. — Observations on Polyzoa, Suborder Phylactola^niata.

1865.

6. L. Joliet. — Organe segmentaire des Bryozoaires endoproctes.

Arch, de Z(^ol. exjM'r. et gén. VIII. 1880.

1 O

7. E. Metschnikotf.— P>ull. de l'Acad. de St. Pétersbourg. XX.

1871.

8. H. Nitsche. — Beiträge zur kenntniss der Bryozœn. Zeit. f. w.

Zool. XXV. Suppl. 1875.

9. W. Reinhard. — Zur kenntins der Siisswasser-Bryozoen. Zool.

Anz. IIT. 1880.

10. A. Sa?fftigeii. — Das Nervensystem der Phylactol.iînien siisswassei'

Bryozoeu. Zool. Anz. XI. 1888.

11. M. Salensky. — Etudes sur les Bryozoaires endoproctes. Ann.

des Sc. Nat. 6 sér. Zool. V. 1877.

12. M. Verworn. — Beitrage zur Kenntniss der Süsswasser Broyzoen,

Zeit. f. w. Zool. XLVI. 1888.

OBöEltVA'l'lONS OX FKESU-WATEK PoLYZOA. l^'^

Explanation of Plates.
Plate XVII.

t'ig. , 1. A «mall group of colonies', nat. size.
Fig. 2. A polypide. x 10.
Fig. S. vShape of the cœnœcial endocyst.

Fig. :/. Diagrammatic representation of a [)olypide and a portion
of the cœnœcial endocyst.

Tent. Tentacles. Epist. Epistome.

N. Gang-. ISTervons g-ano^lion.

Oesoph. Oesophagus.

Tnvafif. tube. Invai^inable tiil^e.

Over. Ovary. Stato. Statoblast.

M. I. Muscles of the funiculus.

M. II. Parieto-vaginal nuiscles.

M. III. Retractor of the polypide.

M. lY. Muscles of the gastric wall.

M. V. Muscular layer of the endocyst.

M. VI. ^luscular fibres of the epistome.

Nephr. Nephridia. Loph. Lophophore.

Tent, membr. Tentacular membrane.
Fig. 5. Statoblast. a. Front view. h. View in profile.

Plate XVIII.

Fig. 6'. Cells in the ectocyst. F x 2.*
Fig. 7. Section of the endocyst. F x 2.
Out. lay. Outer layer.

* Zeiss' powers.

148 A. OKA.

Bas. iiiembr. Basement membrane.
Tr. mu.s. Transverse muscular fibres.
L. mus. Longitudinal muscular fibres.
Lin. epitli. Lining epitbelium.
A'^ac. Vacuole.

.F((j. 8. Longitudinal section of the epistome, with the ganglion
and the excretory organs. B x 4.
Gang. cav. Ganglion cavity.

Fig. 0. Cells of the upper half of the œsophagus. F x 2.

Fig. 10. Cells of the lower half of the œsophagus. F x 2.

Fig. 11. Section of the cardiac valve. B x 4.

Fig. 12. Cross section of stomach.

Fig. 13. Cells of the inner layer of the gastric wall. F x ::^.

pyr. c. pyramidal cells.

cl. c. club-shaped cells.
Fig. U. Cells (jf the rectum. F x 2.
Fig. 15. a. Diagram showing the extent of the mesentery.
h. Section of the mesentery. D x 2.

Fig. 16. Cross section of a tentacle. F x 2.

Fig. 17. Longitudinal section of the tentacle. F x 2.

Fig. 18. Diagram showing the base of tentacles.

Fig. 19. Diagram showing the direction of the currents of the

perigastric fiuid.
Fig. 20. Cells floating in the perigastric fluid. F x 2.

Figs. 21, 22, 23, 24, 25, 26. Sections at various levels of the upper
portion of a polypide. B x 4.

Figs. 21 A, 22 A, 23 A, 24A. Sections of the excretory organs. F x 2.

Fig. 26 his. Entire form of the excretory organs.

OBSERVATION'S OX FliKSH-VVA'l'ËK FoLYZOA. l'lî>

Plate XIX.

Fig. 27. Nervous ganglion.

Fig. 28. Saggittal section of the gatiglion. E x 2.

Fig. 29, a, h, c. Diagrammatic Sections of the ganglion, showing
the extent of the ganglion caA'ity.
<T, sagittal, b, horizontal, c, frontal, sections.

Fig. 30. Cross section of a lophophoral arm. D x 2.

Fig. 31. Cross section of the fimiciiliis. F x 2.

Fig. 32. Longitudinal section of the upper extremity of the funi-
culus. D x2.

Fig. 33. Longitudinal section of the lower extremity of the funi-
culus. D X 2.

Fig. 31. Section of Ovary. F x 2.

Fig. 35. Marginal spines of statoblast. F x 2. •

Fig. 36. The Envelopingcell-layer of thestatoblastic content. F x 2.

Figs. 37-15. Various «tages in the development of the sttitoblast.
37-39. Fx2. 40. I) x 2. 41-45. B x 4.
Fun. Funicular wall. Cyst. c. Cystogenous cells. Gr. m.
Granular cell-mass. Caps, chitinous capsule. Env. m.
Envelojung cellular membrane.

Fig. 16. Section of a mature statoblast. B X 4.

Fig. 16 A. A portion of the statoblastic content. F x -j.

Plate XX.

Figs. 17-52. ^'arious stages in the development of Polypide in the
statoblast. B x 4.
Prim. 1. Primitive lumen.
Bl. c. Floating cells.

150 A. oitx

tigs. 47 A, J^8A. Portions of the .stutobJnstic content in the stages
corresponding to Figs. 47 and 48. ¥ x 2.

Fig. 52. FJoating cells. F x ii.

Figs. 53-60. Stages in Budding. 53, F x '1. 54-56, D x !>.
57-60, Bx 4.

Fig. öl. Diagrams showing tlie manner of budding. The lloman
numerals slnnv the order of the individuals.

'«^'^'•Vii

Jour. Sc. Coll. Vol. IV PI. XV H.

'^^^ 'WiiM

'i^i'Ä^/ -.'

%

% ^^H

^%|!itii^i*^^^*''' ^■'

%

MVI-
Epiät.-

UiK

^^

-oph.
■ephr.

— Intestine

Ten t. m em br

Inva?t.Fold

-rt'ç. /.

^ '/

Ä?.f

Jour. Sc. Cell. Vol. IV PI. XVIII.

ZropSnphmtii.e^viij

Jour. Sc. Coll. Vol. IV PI, XIX.

I' )V

ftff. 36.

V -rJS-l^

Jour. Sc. Coll. Vol. IV PI. XX.

— Outlay

Û """

Fig. 60

%' \ ~

'tj i'iy jtto- pMik- pr"

On Diplozoon nipponicum, n. sp.^^

by

Seitaro Goto, Rigakushi,

Post-graduate Student in Zoology, Imp. Univ.

With Plates XXT— XXriI.

Since Diplozoon paradoxum was first discovered and described by
V. jS[ordmann,"Mt has been made the object of special investigations
by many eminent naturahsts. But our knowledge of the anatomy
and especially the histology of this interesting genus, hitherto with
but a single species, is, notwithstanding the publications of Paulson,
Zeller, and others, by no means as complete as could be desired. I
have, therefore, undertaken, at the suggestion of Prof. Ijima, to
subject it to a renewed investigation. I at first believed that the
Japanese species was identical with the European ; but as I went on
with my work, many points came to view, that made me doubt this
identity ; and a close comparison with some preparations of the Euro-
pean species taken from Leuciscus rutilus, and brought back fr(3m
Germany by Prof. Ijima, has led me to erect it into a new species,
for which I propose the name of

Diplozoon nipponicum.

Before proceeding any farther, I must here discharge the
pleasant duty of acknowledging my deepest obligations to

1) This paper was originally presented as a graduating dissertation.

2) Nordmaan— Mikrographische Beiträge. I. Heft, 1832. p. 56.

152 s. GOTO.

Professor Ijima, already named, not only for constant supervision of
my work, but also for lending me his books and preparations pertain-
ing to the subject at hand. He has also handed over to me his un-
finished manuscript, in which the anatomy and external features of
many ectoparasitic Trematodes have been made out to a great extent
— a circumstance f)r which I here express my warmest thanks.

Dipl. nifponicum is very common on the gill of Carassius vulgaris.
Its differential characters as compared with Dipl. paradoxum are 1) the
smallness of the posterior suckers, 2) the greater length of the
posterior half of the body, 3) the shortness of the " connecting canal "
between the intestine and the oviduct, 4) the presence of a pair of
glands at the entrance of the mouth, and 5) the fact that the intestin,
does not present lateral branches in the posterior portion of the body.

"Comment la reunion des Vers, a-t-elle lieu? sont-ils réunis
comme les frères Siamois, ou bien sont-ils ci'oisés comme les deux
jambes d'un X?" Hy the investigations of v. Siebold^' and Zeller,^^ it
has been established beyond doubt that the double animal results by the
union of two Diporpac in the form of a cross — a fact which had already
been anticipated 1)v Dujardin^^ their discoverer. The manner in
which the two individuals are united, and the details thereof have
already been made out by Zeller, who has also discussed the various
opinions of his predecessors, and corrected their errors. I shall
however, add a few remarks on some points not noticed by him,
some of which are perhaps peculiar to the new species. For examining
the external features of the worm, as well as for other purposes, it is
best to kill it with boihng sublimate, in a, watch-glass in which just
sufficient water has been placed to cover its body. The worm which

1) V. Siebold — Ueber die Conjugation des Dipl. paradoxum. Ztschr. f. wiss. Zool. Bd. III.
1851. p. 62.

2) Zeller — Untersuch, ii. d. Entwicklung d. Dipl. paradoxum. Ztschr. f. wiss. Zool. Bd.
XXII. 1872. p. 168.

3) Dujardin — Histoire naturelle des helminthes. 1845. p. 316.

ox DIPLOZOON XIPPONICUM, N*. SP. 153

has been killed in this way, preserves a natural {xjsition corresponding
to its condition of rest, and can be examined when convenietit.

Each individual, if considered separately, is elongated and lan-
ceolate in form, with a deep notch on one side a little posterior to the
middle of its winde length, by means of which it is united with the
other individual ; so that we may hereafter speak of the anterior and
posterior halves of the body. The anterit)r lialf is widest near the
place of union, and becomes narrower anteriorly, where it ends with
a rounded outline, and where the mouth is situated on the ventral
side. In cross-section it presents an oval outline, which gradually
becomes more circular as we proceed anteriorly. If the worm has
died in a contracted state, the siu'face of the body is thrown into
numerous transverse folds ; otherwise the surface is entirely smo<jth,
except where little conical elevations, hereafter to be described, exist.
The posterior half may briefly be described as an elliptic cylinder in
the anterior portion, which, posteriorly, passes irregularly into a rect-
angular prism. [t is also much slenderer than the anterior half.
Seen in profile, the surface of the posterior portion is always, in
specimens killed with hot sublimate, thrown into a number of strong
folds, due no doubt to the powerful development of the diagonal
muscular libres in this region ; so that here the margin is deeply
crenate or even zig-zag (PI. XXI, Fig. I). A cross-section through
one of the posterior zig-zag folds presents nearly a rectangular out-
line. The folds suddenly come to a close at a short distance before
the beginning of the posterior suckers. In this portion the crcjss-
section presents a flattened ellipse ; and this part is, on surface view,
distinctly marked off from the sucker-bearing portion which directly
follows it, and still more so from the strongly folded more atiterior
portion, by a sudden change of level (Fig. 1). The lateral margins
of this sucker-bearing portion are suddenly thickened on the ventral

154 s. GOTO.

side, so that a cross-section through it is somewhat cbo- shaped.
Under the pressure of a cover-glass, this portion asstunes a somewhat
oval form — a circnmstance which probably induced v. Nordmann^^
and Paulsom' to indicate it as an oval "Scheibe." Van Beneden^-* further
speaks of" un pédicule " by which the " deux organes ", which carry
the suckers, are attached to the body ; but I have not observed any
such structure in my species, and it is probably due to a deformation
cause<l by pressure and the extreme mobility of this part. The pos-
terior margin of the body shows either a nearly straight line
or, more commonly, a slight concavity (Fig. 2).^^ Toward this
concavity, the body again becomes thicker, the thickening begin-
ning this time in the median line, and thence spreading toward the
sides, as indicated in Fig. 2, where the shading is made as if this
portion were seen from the ventral side, on a longitudinal section,
therefore, which does not pass through the hiteral suckers, the bcjdy
is seen to present posteriorly a club-like thickening, and end suddenly
as if it were cut off. Van Beneden speaks of an "excavation plus ou
moins profonde," by which he no dcndjt means the hollow, just spoken
of, between the suckers.

The peculiar sudden bend (Knickung) towards the ventral side,
which the body of the worm suff'ers at the place of crossing, has
already been noticed by Zeller. There is, however, another feature not
observed by him. If the worm, ijamely, is viewed from the p(^sterior
end, or if sections of this part are cut, it will easily be noticed that
the bodies of the two individuals do not stand exactly opposite each

1) Xorduiaun — 1. c. p. 60.

2) Paulson — Zur Anatomie v. Dipl. paradoxum. Mém. d. I'acad. St. Petersbour>^. VIT.
Sér. T. IV. 1862. p. 4. I have not been able to gain, access to this work, and am indebted for its
account to Prof. Ijiuia's notes.

3) P. J. V. Beneden — Mémoire sur les vers intestinaux, p. 41.

4) The European species shows a decided convexity.

ON ÜIPLOZOON NIPPOXICUM, N. SP. 155

otlier, but that one is always a little either to the left or to the rio-ht
side of the other, according as the worms are united by the corres-
pondiiiiT sides of their bodies. This feature is usually less noticeable
in the anterior hah es, l)ut it can easily he brought to view by bring-
ing them close to each other. It is caused, no doubt, by the flict
that the bodies of the two individuals are closely united onkj at the
point of crossing ; as may be seen, if one phices two pie(^es of straw
against each other in the form of a «-ross, and presses l^heni together
between two hngers at the point of rrossing. Beside this imperfect
apposition of the corresponding halves ()f the two individuals, must
also be noticed the twist, to which each is subjected at the place of
crossing, in consequence of the fict that one grasps with its ventral
sucker the dorsal papilla of the other. To this twist, though very
small in degree owing to the presence of the notch already men-
tioned, must be attributed the common occurrence that, when the
^vorni is killed under the pressure of a cover-glass, tlie anterior and
posterior halves (jf tlie same individual present to view opposite sides
of the b()dy — the anterior half presenting the dorsal if the other half
presents the ventral side, and vice versa. I'he two individuals are
united with each other by their sides, so that here a deep indentation
arises — the notch already spoken of. Here the epidermis is absent,
and the muscular layers of the two individuals are directly applied
to each other. Zeller describes a direct connection between the vas
deferens and " Laurer's canal" of the two individuals ; but a careful ex-
amination has ctuivinced me that this view is erroneous. I lind Laurer's
canal to open distinctly into the intestine, and the vas deferens of one in-
di\ idual into the yolk-duct of the other, as will be {)roved later on. In
this connection, it may be mentioned that the same writer thinks the
Diporpa incapable of "eine noch weiter gehende Entwicklung ohne dass
zuvor die Copulation mit einer zweiten Diporpa zu Stande gekommen

i5é s. GOTO.

wäre."'* But last «iimmer I met with two Diporpae which were
ah'eady producing eg'g«, but which were not united. They were
attached to the «ame gill very near to each other. They were quite a«
large as any average Dipluzuuu, and measured about 6 mm. in length
in a completely outstretched condition.-* They were provided with
four pairs of posterior suckers, but there was no trace either of the
ventral sucker or of the dorsal papilla. In place of the ventral sucker,
the longitudinal muscular layer was very strongly developed in the
corresponding part ; and the body shewed a sudden increase of
breadth just anterior to the anterior end of the ovary, looking as if
this part were bandaged. I have used the utmost care in detaching
the worms from the gill, inasmuch as I carefully scraped off the
gill- slime with a s[)atula, avoiding as nuich as possible any direct
contact with the worms. The Dq^orpae in question were observed
to be quite independent of each other from the moment they were
detached from the gill ; nor have I been able to detect any mechanical
injury, or the notch by wliich they might have been united to each
other ; so that the chance of their being detached Diphzoon is, I
believe, almost entirely excluded. Such a case of isolation is of
course exceptional ; but it shews that the Diporpa can, under certain
conditions, become mature without uniting with another Diporpa.

It would have been interesting and instructive could I have
determined where, in this abnormal case, the vas deferens opened.
But unfortunately, owing to my inexperience then, I killed both the
Diporpae under the pressure of a cover-glass and prepared them for
oTobs mountinu" ; and when I afterward cut one of them into sections,

1) Zeller— 1. c. p. 17G.

2) The size of the common Dipori^a varies according to its stage of development. Dujardin
o-ives it as 0.26 — 0.56 mm. in length and 0.18 — 0.35 mm. in breadth (1. c. p. 317). A specimen
of the Diporpa of Dipl. parapoxum lent me by Prof. Ijiuia and possessing three pairs of suckers
measured about 0.6 mm. iu length. That of Dipl. nipponicum vi the same stage is of about the
same size.

ON DIPLOZOON NIPPONICUM, N. SP. 157

I could no longer trace the course of such a delicate canal as the vas

deferens.

Remark : — In Prof. Ijima's manuscript I find the following
passage which I have his permission to puhlish.

" Ich will mir endlich noch eine Bemerkung über die von
Heller^^ beschriebene Monstrosität erlauben. Dieser Forscher lässt,
obschon ihm das Verhältniss des CopiiJafio lateralis decussata
(Siebold) nicht fremd blieb, sein interessantes Exemplar sich
dadurch erklären, dass die Verwachsung der beiden Diporpen sich
über die ganze vordere Kck-jjerhälfte ausgedehnt bätte. Paidson,
der sieb übrigens mit die. Ansicht Leuckart's fheilt, dass die
Diporpen einfach mit Bauchfläche zusammenhängen, hebt die
Unmöglichkeit des Zustandekommens jener Monstrosität durch
Cojndatio lateralis decussata hervor, und nimmt an. es handele sich
um eine Missbildung per defectum eines Diplozoons, bei welchem
sich einer der Vorderleiber gar nicht entwickelt hätte. Dabei
kam er sein- nah an die richtige Inter[)retation der Heller' sehen
Monstrosität, die meiner Ueberzeugung nach, nichts anderes sein
kann, als eine Diporpa, nicht Diplozoon, mit in doppelter Anzahl
angelegtem Schwanzende, also eine Misslnidung per adjectwn.
Dies darf man nicht W luidei* nehmen, denn wir wissen zahlreiche
Fälle ähnlicher Missbildungen unter den l'lanarien. Ich keniie
selbst einen Fall von ganz jungen, eben ausgeschlüpftem Dendro-
cœlum lacteum, mit zwei hinteren Hälften, deren je eine einen Mund
und einen Pharynx besitzt."

I shall now proceed to the consideration of the various parts.

1) Heller — Merkwürdiger Fall vorderer Verwachsung an Dipl. paradoxum. Sitzungsber.
d. k. Akad. d. Wiss. Wien. 1857. p. 109.

158 :^. GOJO.

I. The Epidermis.

The nature of the integument of the Trematodes has been
variously represented by various authors. This subject I hope to
discuss more fully in a later work which shall treat of our ecto-
parasitic Trematodes in general. Zeller'^ tells us that if no occa-
sion is offered the embryos to attach themselves to the gill, " schon
nach Verfluss von 5 Stunden (^ after the embryos have left the e^^g)
einzelne der Wimperzellen reissen sich los, bald mehrere und schliess-
lich alle, ilinmiern aber auch abgetrennt noch eine Zeit lang fort."
The embryos finally die. It is not clear from his statements
whether this throwing off of the " Wimperzellen " is a normal process
or not. In Polijstomum he merely says that they " schrumpfen," but
does not describe their exact fate. In the case of Distomum, how-
ever, it has been proved by Schwarze^^ and Biehringer^^ that the
so-called " cuticula '" consists originally of cells which undergo one
by one a peculiar transformation, and which do not at any time possess
the typical ej)ithelial arrangement. After the first rough manuscript
of these pages had been finished, I received the article of Braun''* in
" Centrbl. für Bakteriologie u. Parasitenkande," in which the writer
brino-s forward some strong and interesting evidences as to the
epidermal nature of the integument. In view of these facts estab-
lished by the preceding investigators, I believe I may regard the
inteo-ument of the monogenetic Trematodes as a modified epidermis, —
the more so from the consideration that it has a distinct cuticle and

1) Zeller— 1. c. p. 173.

2) Schwarze — Die posi^embryouale Entwiekhmo- <lei- Treinntoden. Ztschr. f. w. Zool. Bd.
XLII. 188G. p. 49.

3) Bielu'inoer — Beitrüge zur Anatomie n. Entwickl.-gesciiichte d. Trematoden. Arljeiten
a. d. zool.-zoot. Inst, in Würzbnrg. Bd. VIT. 18S5. p. 4.

4) Braun, Max — Einige Bemerkungen ü. d. Körperbedeckung ectoparasiti scher Tremato-
çlen. Centralbl, f. Bakteriol. u. Parasitenkunde. Bd. VII. 1890. p. 594 (Nr. 19.)

ON DtPLOZOON NIPPON ICUM, N. SP. 159

sits on a basement membrane An embryolog-ical stncl}', however,
of the transformations which the original epidermis undergoes is. as
Braun maintains, very desirable.

The integument of Dipl, nippouicum is comj)Osed of two layers,
the cuticula and the underlying matrix. The cuticuLi, when
examined in a living worm, is a very thin, structureless, refractive
membrane. In sections of hardened specimens it appears as an
insignificant line bonnding the subcuticular (= epidermal) layer
against the external world. It is very well seen in a living specimen
wliich has been nllowed to macerate in water for some time under
the cover-glass. Numerous watery blisters then form in the epider-
mis, and separate the cuticula from the underlying layer. The
former cnn then be examined as a separate structure. Transverse
canals have been described in the cuticula of many Trematodes, but
I have not observed any in the new species. The cuticula is reflected
inward for some distance into the mouth.

Directly luider the cuticula lies the epidermal layer, a uniform,
granular matrix in which no nuclei are to be observed. I believe
I have observed indistinct dark Hues traversing the breadth of this
Liyer but not quite reaching the cuticula. The epidermal layer, like
the cuticula, is continued into the cavity of the mouth, and the sticky
glands hereafter to be described (p. 166) are but local modifications
of it. Wierzejski'^ describes the "Haut" of Calicotyle Kroyeri as
consisting of " einer feinen Cuticularschicht mit den darunter liegenden
kleinen, runden Matrixzellen "; but judging from his figure, I beheve
he has mistaken the imclei of the connective tissue for his " Matrix-
zellen." The epidermal layer rests on a basement membrane, which
eagerly takes up coloring matter, and is very conspicuous in cross-

1) VVierzejski— Zur Keantaiss des Baues von Calicotyle Kroyeri. Ztschr. f. wiss. Zool.
Bd. XXIX. 1877. p. 552.

160 s. GOTO.

sections as a dark line with indistinct l/orders, sepai'ating the epider-
mal from the muscular layer. It is also much thicker than the
cuticula. The total thickness of the integument, with the cuticula
and hasement membrane taken together, is about 0.004 mm.

-, It has already been mentioned that little conical elevations exist
here and there on the surface of the body. These are more abundant
on the ventral than on the dorsal side, and are entirely absent in the
posterior half of the body. They are simple elevations of the epider-
mis with an almost homogeneous mass of connective tissue under it.
Here the nuiscular layers do not touch the basement membrane, but
p:!ss straight on ; so that these elevations are somewhat subject to
changes of form. I have represented one of them in section in
Fig. 7 (PI. XXII). As will be seen, they are pointed at the end.
A very similar structure has been described in Sphijranura Osleri,^^
where it seems to act as a sense organ. But although I directed my
special attention to the point, and applied the highest magnifying
power at my disposal (Zeiss Imm. L.), I could not discover any
canal opening at the apex, or any hair-like projection, or any fibrils
such as have been observed in the above-mentioned species to supply
these conical bodies.

II. The Muscular System.

The muscular svstem is const inued bv the muscular wall of the
body, the dorso-ventral muscles, and the muscles pertaining to the
various organs.

The muscular layer of the body consists of three layers. These
are, counted from outside inwards, the circular, the diagonal, and the
longitudinal muscles. The circular fibres run everywhere immediately

1) R. Wriglit and Macallum — Sphyrauura Osleri : a Coutribuiiou to American Hclmin-
thology- Journ. of Morph. Vol. I. 1887. p. 9.

ON DIPLOZOON NIPPOXICUM, N. SP. 161

beneath the basement membrane. They run isolated without form-
ing bundles. This layer is most strongly developed at the anterior
extremity of the body in the region of the anterior suckers, and im-
mediately anterior to them, especially on the ventral side (Fig. 9),
where its thickness amounts at some places to 0.01 mm. In the
posterior part of the body it is very weakly developed, and in
the region of the posterior suckers the fibres ai*e very difficult to
detect.

Closely applied to the circular layer of muscles run fhe second
or diagonal fibre« (Fig. 10). In the anterior half of the body these,
like the transverse fibres, run isolated without forming bundles ;
and those coming from opposite sides of the body cross each other
at an angle of nearly 120°. In the posterior half of the body, this
layer is strongly developed in the region of the folds already men-
tioned, where the fibres run in flat bundles and close to one another.
According to Taschenberg,^-" the diagonal fibres are situated inner-
most in Tristomum ; but in all the species of ectoparasitic Trematodes
I have hitherto examined, viz., in Microcotyle, Axine, Octohothrmm,-^
Dactylogyi'us, and a new genus not yet named, the diagonal fibres
are situated between the transverse and longitudinal muscular layers.
Lorenz, ^^ who includes the transverse and diagonal muscles under
one head, also places the "zarteren Fasern " (by which he means the
two sets of muscles just mentioned) outward ; and an examination of
the sections of Tr. molae, kindly placed at uiy disposal b}' Prof. Ijiuia,
has convinced me of the error of Taschenberg, occasioned perhaps by
the circumstance that in Tristomum the longitudinal fibres describe a

1) Taschenberi;- — Beiträge zur ICenntniss ectoporasit iseher mariner Trematoden. Halle,
1879. p. 11.

2) In a species of this genus wliich I have examined, there are in addition isolated longi-
tudinal fibres between the diagonal and circular uiuscles.

3) Lorenz — Ueber die Organisation der Gattungen Axine u. Microcotyle. Arbeit, a. d.
zool.-zoot. Inst. d. Univ. Wien etc. Bd. I. 1878. 3. Heft.

162

s. GOTO.

curve in the lateral portions, corresponding to the circular or oval
outline of the worm. The same writer did not observe the diag-onal
fibres in OnchocoUjle appendiculata and Pseudocotyle squatinae ^'; but
since they are present in all the species 1 have examined, they were
probably overlooked by him.

Internal to the diagonal muscle, and separated from it by a
greater or less amount of connective tissue, run the longitudinal
muscular fibres in parallel bundles of greater or less strength.
They are more strongly developed on the ventral than on the dorsal
side of the body, as is usual in most Trematodes, and cause a slight
curve of the body on the ventral side when the worm is killed with
hot sublimate. The fibres that constitute the bundles are but loosely
united together by connective tissue, and form by no means such
compact muscular bundles as we see in some other Trematodes.
Tliey appear in cross-section as dots, separated from one another by
a greater or less amount of connective tissue between. Some of the
fibres of a bundle often diverge from their previous course, and enter
into the formation of a neio^hborino^ bundle. Most of the lono;itudinal
fibres combine toward the posterior part of the body to form a certain
number of strong bundles, which proceed posteriorly, and are inserted
one to each sucker on the median chitinous piece of the posterior
wall (Fig. 5).

The dorso-ventral muscles (dvm in Figs. 11, 13, 16. !24)
are well developed. Each muscle generally breaks up into a few
branches dorsally and ventrally before being inserted into the base-
ment membrane. They traverse the brain, vitelline body, and other
internal organs. In longitudinal (sagittal) sections of a specimen, in
which the vitelline body has not yet well developed, the dorso-
ventral muscles are seen to be placed at pretty regular intervals. In

I) Tascheaberg — Weitere Beitrüge zur Keautaiss ectopar. mar. Treinatoclen. Halle, 1879.

ON DIPLOZOON XIPPONICUM, X. SP. 163

specimens with fully develojjed vitelline body, these muscles are
obscured to a g^reat extent.

III. The Organs of Attachment.

The organs of attachment are constituted, posteriorly by the
four pairs of suckers already mentioned and a pair of hooks, and
anteriorly by a pair of suckers and sticky glands. Each posterior
sucker (Figs. 3, 4, 5) may Ijrieüy be described as a short-ovate, flat
bag with its wide mouth directed ventrally, its walls very thick, and
the line of its greatest breadth directed transversel}^ to the long axis
of the body ; so that we may speak of the anterior (aw), posterior
(pw), and lateral walls. The first two walls are very thick, and are
directly continuous with each other at the bottom of the bag
(Figs. 4, 5). The lateral walls are very thin, and seem to consist of
a cuticula-like refractive membrane only. The entire structure is sup-
ported by a framework of chitinous rods, which are by no means
so numerous or complicated as Nordmann has represented them.
They are five in number : a U-shaped median piece (pm), a pair of
curved pieces (ppa), (resemliling in form certain fishing-hooks), to
support the anterior wall, and a pair of simihu- pieces (ppp), with a
large process ([)p) at the base, to support the p(rsterior wall. Having
thus given a general idea, I shall now proceed to the explanation of
the three figures already referred to, by which I hope to make clear
the structure of the suckers. Fig. 3 represents the chitinous rods as
very commonly seen in a specimen observed under the pressure of a
cover-glass, with the mouth of the sucker directed below in the figiu'e
and the rods belonging to the anterior wall shaded more deeply.
Fig. 4 represents a section made in the direction indicated by the
line ab in Fig. 3, whereby it is to be remarked that the median

164 s. GOTO.

piece has been cat nearer the fundus of the sucker. This section
shews the thickness of the anterior and posterior walls, as well as their
fibrous structure. The prismatic fibres, of which these walls are com-
posed, are strongly refractive, and are scarcely colored by haem-
atoxylin. They seem, therefore, not to be of the same nature as
the muscles of the body ; these being well stained by the same
coloring fluid. In fact, they seem to be not contractile but
elastic fibres. The supporting rods are all of them hollow, with,
the inner surface, however, not quite smooth, but with irregular
projections, which sometimes unite with those of the opposite side,
and form septa-like partitions (Fig. 3). The paired rods are
some\vhat triangular in section, and are imbedded in their respec-
tive walls along the margins. The rods of the posterior wall are
articulated at their bases each with another piece (pp), which is im-
bedded in the substance of the wall, and imparts greater strength to
it — a flict well in accordance with the circumstance already mentioned
that the main bundle of muscle is attached to this wall. Fig. 5 repre-
sents a section made in the direction indicated by xy in Fig. 3, i. e.,
in an antero-posterior direction. In this section, the direct continuity of
the anterior and posterior walls is clearly seen ; the U-shaped median
piece has been cut but in part, as also the extremities of the paired
pieces at the entrance of the sucker. The median piece exhibits, in
the posterior wall, a deep cut, where the main bundle of muscle is
attached for controlling the varied movements of the sucker. Beside
this bundle, weaker (jnes are attached to the paired pieces. The
fibrous substance of the wall is Ijounded by a cuticula both against
the external world and the surrounding inesenchyma. The support-
ins; rods are v^er\- easily broken into fi-aa'ments when the animal is
subjected to too much [,>ressure ; and tliis takes place pretty regularly
in the manner represented by Nordmann, who, however, describes the

ON DIPLOZOON NIPPOXrCUM, N. SP. 165

fragmentary pieces ii>* " Bügel, zahnfönnige Vorspriingf. Ki}ipen,
u. s. w." All the |)«)steri<^r suckers are of the same build ; but they
vary somewhat in size, the last pair being always smaller than the
anterior ones,^^ and the first pair very often smaller than the following
two pairs. Measurements on five individuals gave the average breadth
of the suckers as 0.093 mm.''

Besides the suckers just described, there is, on the dorsal side, a
pair of solid chitinous pieces (Fig. 6). Each piece consists of two
parts. The basal portion, to which a small bundle of muscle is
attached, is straight, and acts as a handle. To this is articulated a
hook-like piece, whose end alone sticks out fr(mi the surface of the
cuticula ; the handle as well as the other part of the hook lying in
the integument. The straight handle-like portion and the liook con-
stitute a single piece, and not two pieces as v. l^eneden'^' thinks. The
total average length of the piece is 0.072 mm.

The anterior suckers are either round or egg-shaped, according
to the different states of contraction, and are situated right and left
at the entrance of the mouth. Like the posterior suckers, the walls
(Fig. 9) are composed of prismatic fibres placed at right angles to
the investing membrane, which lines the whole internal cavity, and
bounds the wall from the surrfMuiding mesenchyma. In cross-section,
the sucker is generally circular in outline. Each is provided with a
numlier of special muscles for the control of its uKnements in suction.
These muscles I have represented in Fig. 12, where there will be seen
three bundles coming from the dorsal side, two of which are attached
to the anterior border of the sucker, and the remaining one to the

1) This faet must not be taken as proving that the hindermost pair is formed last.

'J,i A corresponding- measurement on tlie European species «f about the same size gave the
average result as 0.144 mm. fort lie sucker, and 0.084 mm. for the tot;il length of the handle
and hook.

3) P. J. V. Beneden— 1. c. p. 42.

166 s. GOTO.

posterior ventral border. A biuidle, which soon divides itself into
two smaller bundles, proceeds from the ventral side, and is attached,
one of the branches to the same point as the posterior dorsal bundle,
the other branch a little more ventrally and anteriorly. Two weaker
bundles start, in addition to the above, from the upper and lower lips
of the mouth, and are attached to the corresponding borders of the
sucker. I have observed some of the fibres of these various bimdles
directly continued through the substance of the wall, and inserted on
the cuticula that lines the cavity of the sucker. By the combined
action of these muscles, the worm can exercise a strong suction on the
gill of the host, and. extract its blood.

Besides these suckers, there is a pair of glands at the entrance of
the mouth, just anterior to the suckers, which seem to be pecnHar to
Dipl. nipponicum. They can be seen well in a living specimen under
the cover glass, or in preparations of the entire worm, as a round,
paired body. One of them is seen in section in Fig. 8, which shews
it to be a gland formed bv the invagination and local modiiicarion of
the epidermis. It has generally a reniform cavity, which opens into
the mouth by a canal, just anteriorly and close to where the sucker
opens into the mouth. The epidermis is continued into the canal for
a certain distance, and then changes its character, becoming firmer
and refractive like the cuticula. The cavity of the gland is destitute
of any distinct epithelium, but is generally filled with a granular mass,
which stains very well. This mass is densest near the wall, and gra-
dunlly becomes thinner towards the centre, where there is generally
an empty space. I have often observed the exit canal filled with
a deeply stained granular mass, very similar in appearance to the con-
tents of the sticky glands of DacNilogi/riis and other allied forms, and
which is doubtless the sticky secretion of the gland. Next the gran-
ular content is a basement membrane. The wall is exceedingly

ox DIPLOZOOX NIPPOXrCUM, X. !SP. 167

thick and niu.sriihir. The luiisciiJar übres are mostly arranged meri-
dionally, i. e. if we suppose the ventral and dorsal pole of the gland
to correspond to the two poles of the earth, the muscular libres are
arranged nearly in the plane of the meridians. Fibres also come
from the dorsal side of the animal, and enter the wall. ]>et\veen the
muscular fibres, I have sometimes observed nuclei, which are to all in-
tents and purpose exactly similar to those of th^ general mesenchyma
of the body, and probably belong to it.

IV. The lYIesenchyma.

Ut" the mesenchyuiatous connective tissue of the Trematodes,
Leuckart^' distinguishes two forms. In the Hrst torm, the mesen-
chyma consists of a "fast homogene helle und feinkcirnige Substanz
mit zahlreich eingelagerten kleinen Kernen '" ; in tlie second ioviu oi
the mesenchyma, \ve see "' Zellen von mehr oder minder ansehnlicher
• rrösse, die îuit einer meist wasserhellen Masse gefüllt sind ' and
generally of a polyhedral form, with a tibroiis net- work between.
Taschenberg"' regards the mesenchyma " als ein Bindegewebe, wel-
ches zu einem Maschenweike entwickelt ist, in welchem die urs})riing-
lichen Bildunuszeilen iheilsnoch \'orhanden sind, tlieils aljei' nur an
dein Proto[)lasnia mit darin eingelagerten Kernen sich erkennen
lassen." All these forms of the mesenchyma, however different they
may seem U) be with one another, can. in m\ (ipini(jn, l)e deri\'ed
from the dilferentiation in different- directions of a single primiti\e
fjrm. fhe strong resemblance of the mesenchyma of the Trematodes
to the chorda dorsalis of the \ ertebrates lias already been observed by
i.euckart ; and I 1)elieve the former is formed jn.st in the same manner
as the latter. But first the mesenchyma oî Dlplozoon.

]) Leuckart — Die Parasiten des Meuschen. It. Auflaj^e. I. Bd. 3. Liefg. p. 13 et seq.
3) Tasehenberg — 1. c. p. 13.

168

s. GOTO.

In this tissue are imbedded all the organs hereafter to be des-
cribed, as also some of the organs already described. Owing to the
presence of the vitelline body, the mesenchyma in the anterior half
of the body is mainly confined to the peripheral portion, but is also
present in a scanty quantity between the lobes of the vitelline body
and the cells of which they are composed. When one takes it up for
study, he finds great perplexity and difficulty in making out the true
nature of the elements that compose it, until he compares it with the
mesenchyma of other allied species. In Diplozoon, it consists of a
fibrous substance, in which are seen nuclei each with a distinct mem-
brane of its own. These nuclei always enclose one or more deeply stained
nucleoli. The nuclei are of various size and shape. In the anterior
portion and generally in the anterior half of the body, they are gener-
ally of a comparatively small size (Figs. 7, 8, 9, 25) ; in the posterior
half of the body, however, they are generally of a larger size (some-
times having the diameter of about 0.01 mm.) and have a circular or
oval outline (Fig. 13). In the vicinity of the internal organs, where
the connective tissue is generally more or less compressed, the nuclei
are smaller and often fusiform in shape. Around the pharynx, the
fibres form a fine close net- work (Fig. H).

Beside these elements, we see here and there, scattered apparently
without any regularity in the parenchyma, large vesicular bodies of a
circular or oval outline (Fig. 13), with a large conspicuous nucleus in
the centre surrounded by a mass of granular protoplasm, which on
close inspection betrays a fibrous structure, and which gradually
thins out peripherally, and leaves an empty space between it and the
wall. These vesicular bodies are sometimes drawn out towards one
end, and are very abundant in the posterior half of the body, posterior
to the testis. In the region situated between the ovary and the testis,
the mesenchyma consists of distinct cells with a granular, generally

ON DTPLOZOON XIPPONICUM, X. SP. 160

well-stained protoplasm, of a polyhedral form, and leavino- irregular
intercellular spaces between (Fig. 14).

In Axme, the mesenchyma is distinctly seen to consist of laro-e,
vesicular cells, each with a nucleus generally in the centre, but some-
times attached to the wall, and filled with a hyaline fluid containing
numerous almost uncolored granules. The nucleus as in Diplozoon,
has a distinct membrane, and encloses a deeply stained nucleolus, but
is considerably smaller. Beside these cells, there are, as Lorenz^' has
already observed, in the neighbourhood of the vagina, cells whose con-
tents take up the staining finid very eagerly and appear like ganglion
cells. In MicroGOtyle, the mesenchyma presents somewhat different
aspects in ditferent parts of the body — a statement that holds good
to a greater or less extent in all other allied forms. Around and
outside the vitelline body, the mesenchyma presents an appearance very
similar to that of Diplozoon. Nearer the median line, it consists of
large cells with tlie nuclei in tlie centre, from which protoplasmic fibres
radiate to the wall, whose cavity is filled with a clear fluid without any
granule. Along the median line, finally, the mesenchyma consists of
cells with a granular somewhat fibrous protoplasm which deeply stains
with haematoxylin.-* Here in Microcotijle, I believe, are manifested
the transitional steps tlirough which the mesenchymatous connective
tissue such as that of Diplozoon has been differentiated from the
primitive parenchyma cells. These primitive cells are, I believe, very
nearly represented by the cells of the median portion of Microcotijle.
The next step onward toward the differentiation of connective tissue
is, according to my view, represented by such a form of mesenchyma
as that of Axine, or that portion of the same in Microcotylc situated just
inside the vitelline body — composed of cells of a vesicular appearance

1) Loreaz — 1. c. p. 7.

2) In appearance, these cells are very similar to the yolk-cells of Diplozoon during the
wiutor season. Vide. Fig. 20, PI. XXTl.

170

s. GOTO.

and filled witli a hyaline fluid. A step further onward in the same
direction would result in the formation of abundant fibres, and the
boiindai-ies of tlie oriji'inal cells would be partly absorbed and entirely
obliterated ; so that we should then have a ground-mass of irregular
fibrous substance, with miclei scattered therein — in fact just such a
form of mesenchyma as we really see in most parts of the b(xly of
Diplozoon. The lai-ge, round, vesicular ))odies above mentioned (Fig.
13) arc in fact the remnant cells of the original parenchyma, and the
poi'tion, already referred to. situated between the ovary and the testis,
seems to ha\(' tuulei'gonc ])ut little transf >rmati(^n. and to have pre-
served tlic original cclhdar structure. According to the view here
stated, the so-called j)seudocoel of tlie Trematodes would he not spaces
formed by the departing of the cells from one another leaving inter-
celhilar spaces between them, l)ut spaces which were before triil v infra-
cellular. I do not, indeed, entirely deny the presence of truly rnfer-
cellular spaces, but these ;u-e, 1 believe, comparatively insignificant.

Similarly, of the two t\pical forms of Leuckart, the first results
apparently l)y a simple oblitei-ation of the boundaries of the original
cells. The second f >rm can be (lci'i\ed by a process similar to that
Avhich we have seen to ha\-e taken phice in Microcoli/le, in which some
of the cells (the larger vesicular ones) have maintained their cellular
natiu-e moj'e com])letelv, while others have been moi'e or less com-
pletely transformed into conne(;tive tissue, and j)ressed in, forming the
" Maschengewebe," between the former ; as already |)roved embry-
ologically by Schwarze. ^^ The two forms above mentioned, are con-
nected by numerous intermediate forms, and an actual transition
between them has been observed in some species."^

1) Schwarze — 1. c. p. 59.

2) Looss — Beiträge zvir Kenntniss der Trematoden. Ztschr. f. \v. Zool. Bd. XLI. 1885. p.
432.

ON DTPLOZOON NIPPONICUM, X. SP.

171

Beside the various elements hitherto described, there nre, in the
neighborhood of the brain nnd pliarynx, lari^'e cells of a roundish or
polyg'oiial outline, easily distinguishable from the surrounding
elements of the parenchyma (Fig. 25). They are of a gigantir- size,
and in some sections they seemed as if they were drawn out into
fibres in more than one direction. They have conspicuous \esic(dar
nuclei enclosing each a (U^eply stained nucleolus, which again gt-ncr-
allv encloses a vacuole. They are entire! v destitut<' of cell-walls, and
have a finely gramdar protoplasm. Their very appearance suggests
their nervous nature. i)ut more than that, oareftd examinations have
convinced me tliat these large cells are very constant in their position
and number. Thev are f~)und. namely, laterally and behind ihe
pharynx, and ran be seen in living specimens under the cover-glass,
especiallv well after the water has evaporated to a cei'tain extent. As
will be seen from the figure, they are situated symmetrically, right and
left, on ))oth sides of the pharynx. Ijesides the four cells on each side
and a median ventral one. drawn in the figure, I ha\'e counted
anotlier pair and a median unpaired one more posteriorly. T'here are
also similar <'ells. which are scattered aj)parently without symmetry,
around the brain, but always outside it in the mesen(-hyma. Two of
these are shewji in Fig. 17.

Considering the form and appearance of these (^ells, the constancy
and symmetry of their position and luunber (at least in the more
anterior <Mies), together Avith the circiunstance that there are no
nervous cells m the brain or nerves, I am sti-ongjy inclined to ntti-i-
bute nervous functions to these gigantic cells ; but I have not been
able to trace anv direct connection with the nervous system. I have
tried methyl-violet and c()Ghineal stain. By the latter, they are but
slightly colored, and neither of these stains affords any better clue
into their exact natiu'e. They seem to l)e ditt'ereni from the remnant

172 s. GOTO.

cells of the parenchyma ah*eady described (Fi,o-. 13); hut I must leave
the exact nature and function of these cells undetermined.

Y. The Digestive System.

The digestive system consists of the mouth (Fisr. 2, mo), the
prephnrynx (p[>h), the pharynx (ph), the oesophagus (oe). and the
intestine (int).

The mouth is a funnel-shaped opening situated on the ventral
side of the anterior extremity of the body, at the entrance of which
are placed the glands and suckers already described. Its cavity is
lined hv the continuation of the cuticula of the general surface of the
bodv. The fundus of the fiumel leads directly into an expanded
cavity, the prepharynx, into which the anterior half of the pharynx
protrudes. This latter is an ellipsoidal body which has a narrow
tubular cavity passing through the centre, and whose major axis is
directed antero-posteriorly. In cross-section (Fig. 11) it is circular.
The internal tubular cavity is lined by a comparatively thick struc-
tureless juembrane. The thick wall is composed of muscular fibres
arranged in regular grou|)S, and of connective tissue, in which nuclei,
very similar to those of the general mesenchyma, are to be observed.
Most internally, and separated from the structureless membrane lining
the internal cavity by a sort of basement membrane, is a thin layer of
circular fibres (mci). Most externally, and directly internal to the
cuticidn-like membrane that envelopes the whole pharynx and sepa-
rates it froia the surrounding mesenchyma, is another layer of circidar
fibres, about double as thick as the first. Besides these, there are radi-
al fil)res extending between the internal basement membrane and the
external cnticula of the pharyngeal wall. These radial fibres are
weakly developed, and do not riui in bundles, as they have been observ-

ox DIPLOZOON NIPPON ICUM, N. SP. 173

eel to do in some other Trematodes. Between these fibres is found
a muss of connective tissue with conspicuous nuclei. These nuclei
are doubtless the remnants of the cells that produced the muscular
fibres and the connective tissue of the pharynx. Strong dorso- ventral
muscular bundles (Fig. 11, dvm) are closely applied to the wall of
tlie |;harynx, and no doubt assist in its action. The total thickness
of the pharyngeal wall, the internal membrane inclusive, is about
0.02 mm.

The cavity of the pharynx leads directly into the oesophagus, a
simjjle, slender, tubular portion, which is directly continued into the
meilian trunk of the intestine. This median trunk sends out in the
anterior half of the body, right and left, lateral branches, which rami-
fy dichotomously once or twice. Some of these lateral branches are
distinctly paired, but I have also observed others which are as dis-
tinctly unpaired. Posterior to the place of crossing of the two in-
dividuals the lateral branches are absent. Here the median trunk
divides into two, one of which retains nearly the median position,
while the other proceeds more laterally towards the ovary. Posterior
to the testis these two branches unite, and thenceforth the intestine
proceeds towards the suckers as a simple unbranched tube, and ends
between and a little anterior to the first pair of suckers, where it
generally presents a rounded enlargement. '• A l'endroit ou les deux
corps s'unissent, les coecums digestifs semblent atrophiés, mais en
dessous de l'appareil générateur, dans le bout postérieur du corps,
chacpie tube présente de nouveau ses ramifications régulières et com-
plètement séparées, comme dans la partie antérieure," says v. Bene-
den,^* and I can confirm his observation with my own on the Euro-
pean species ; but in 7iipponicum I have found this part of the intestine
always sim[)le. The wall of the intestine is destitute of an epitheliinn

1) P. J. V. Beneden— 1, c. p. 40.

174

s. GOTO.

such HS we. niv wont to see in the Distoines. In its .stead, we find large
cells (Figs. 14, 16. 19, de) separatee] from one another by a considerable
interval, and filJed with duvk-I)rown or sometimes even black gra-
nnies. I have not (Jl)ser^•ed any wall or nncleus in these (.-ells, al-
though Zeller'' points to the presence oi" the latter in J'o/ystonium. and
I could distinctly oljserve it in OrtohotJiriiün. These black {»igment-
containin"' cells I hold, m agreement with Taschenbero','^ to be dio^estive
cells, and the pigment-grannies to be tood-particles taken in from the
ca\itv of the intestirje. Digestion, therefore, takes place in the allied
forms intracellulary, as in the Turljellarians. The intervals between
these cells are usually destitute of any distinct meud.u'ane in the
anterior half of the body, so that here the digestive svsteni consists
of mere hollows in the mesenchyma ; but in the p<.)sterior part, where
the intestine is simple, 1 could usually distinguish a more or less dis-
tinct memlirane of compact connective tissue.

VI. The Excretory System.

The excretor}'' system (jf the Plathelminthes has been minute-
ly examined by Fraipont,"' Lang," l-*intner''' R. AVright and
^lacallum,''' and some others. ]>y these investigations two [)oints
seem to have been lirndy established : 1) That the excretory system

1) Zeller — Untersuch, ü. d. Entwiek. n. d. Bau. d. Polystomum integerrimum. Ztschr. f. w.
Zool. Bd. XXir. 1872. p. 19.

2) Taschenber;;- — Weitere Beiträi^e. [> 11.

S) Fraipunt— itecherches sur l'appareil excréteur des 'l'réujatodt s. Archiv, d. Biologie. T.
I. I have not been able to gain access to this work, aud_ am indebted for its account to J. V
Carus's " Zoologischer Jahresbericht" (1880. 1. p. 277) and to Looss (1. c).

4) Lang— Der Bau von Gunda segmentata u. d. Verwandtschaft etc. Mittheil. a. d. zool.
Station z. Xeapel. Bd. III. 1882. p. 187.

5) Pintner -Untersuch, ü. d. Bau. d. Bandwurmkörpers, mit bes. Berücksichtigung etc.
Arb-it. a. d. zool.-zoot. Inst. d. Univ. Wien, etc. Bd. III. 1880. 2. Heft.

C) R. Wright and Macallum— 1. c. p. 20.

ON DIPLOZOOX NIFPONICUM, N. SP.

17n

of this class consists of vessels with :i distiiK-t wall, 2) that these
vessels are of two kiln Is, tlie larger oiies serving iii:iiiily f<jr leading
out the contained fluid, and the cnpillarÎL's which end in fnnnel-sluiped
little bodies shewing the so-called " Wimperflannne " in the interior,
and which are the most important part of the system.

In DipL nipponiciun^ as in Dipl. panidoxum^ two main canals (an
always be distinguished on each side of the body, one of which is
larger than the other and opens to the exterior by means of a circnhir
o])ening on the dorsal side^', close to the lateral mai'gin, a short dis-
tance posterior to the pharynx (Fig. 2, eo). Immediately at the
entrance of the o])eniiig, the vessel presents an enlargement (the so-
called " vSammelrohr "), then proceeds anteriorly to about the level
of the pharynx, whei-e it bends backward and proceeds posteriorly,
winding more or less on the w^ay, and giving off but a few
branches. On reaching the posterior suckers, it bends inward to
them and reaches nearly the posterior margin of the body. Here it
turns on itself and proceeds anteriorly, following closely its former
course, but this time liberally sending ont branches which anastonnjse
with one another and with those from the opposite side of the body.
Anteriorly this main vessel reaches the upper lip of the mouth, where
it divides itself into numerous branches, having also become smaller
during its course. These two main vessels follow closely in their
course that of the ventral nerves, on whose dorsal side they are situat-
ed except where they make windings towards one side or the other.
I have sometimes observed a direct connecting vessel between the two
main ones. Within these vessels as w^ell as the branches that proceed
from them are seen, in a living specimen, active vibratory movements,
which generally come to view only after the animal has been left f)r

1) Cf. Brauu — Uebur (.lie Lage d. Excaetiouspori bui d. ectupar. Trematudeu. Zool. Auz.
Jahrg. XII. 1889. p. 620.

176

s. GOTO.

sometime under the cover-glass, and which are executed in such a
way as to drive the contained fluid towards the excretory pores. These
movements are probahly due to the presence of vil)rati]e flaps in the
wall, but I have not been able to observe them in sections. The
wall is seen, in section, to be f(jrmed by a compact refractive mem-
brane with douljle contour, which does not stain with haematoxylin
(Fig. 16, as). Evidences have been advanced by Lang^* and Ijinia'^
that the excretory vessels of the Turbellarians are " nichts Anderes als
durchbohrte Zellen." In the Cestodes, Pintner^' has observed a well-
developed epithelium on the wall of the main vessels, "das zweifels-
oline als Matrix ihrer glashellen, homogenen Membran aufzufassen
ist." According to Schw^arze*\ the central excretory vessel of the
Distomes is ut first a solid string of cells, which afterward acquires a
lumen. He also supposes that the finer branches originate in the same
way, and that the structureless condition of their walls in the adult
worm may be explained by supposing " dass nach der Resorption des
Inhaltes der primären Zellen keine äussere, muskulöse Zellenlage
gel)ildet wird, somlerti die Wandung sicli, al lein, aus der äusseren ZeUmein-
hraiini der ursprüngliclieu Aidagc zusaunneiisctztr Whether the walls of
the excretory vessels oi Diplozoot is to be regarded as similar to those
of the Turbellarians, with the difference that the protoplasmic remnants
of the original cells have been transformed into a structureless mem-
brane, or whether they had been produced by a distinct epithelium
which afterwards underwent degeneration and finally disappeared, or
whether they were formed by such a process as Schwarze^- supposes

1) Lang— 1. c. p. 212.

2) Ijima — Untersuch, ü. d. Bau u. d. Entwickl. d. Siisswasser-Dendrocoelea. Ztschr. f. w.
Zool. Ed. XL. 1884. p. 397.

3) Pintnei— 1. c. p. 21.

4) Schwarze — 1. c. p. 58. The italics are mine.

5) In the extreme case, viz. wliere the cells are arranged in a si agio row, the view of
Schwarze reduces itself to that of Lang and Ijima.

ON DIPLOZOON NIPPONICUM, N. SP. 177

to have taken place in the smaller excretory vessels of the Distomes,
I must leave entirely undecided, with the single remai'k, however,
that in Diplozoon I have observed no trace of nuclei in the wall.

The capillaries, furnished like the larger vessels with a distinct
wall of compact refractive membrane, proceed from tlie smaller
branches of the main vessels, and continue throughout their whole
course without undergoing any perceptible diminution of their calibre.
They are especially abundant in the layer of the mesenchyma just
under the muscular wall of the b()dy. The}' do not, like the branch-
es of the larger vessels, anastomose with one another, and no vibratory
movement is to he observed within. They branch freely, and each of
the branches ends with a minute funnel-shaped enlargement (Fig. 15),
within which is to be seen an active vibratile flap, the so-called
" Wimperflamme." Various structures have been described in con-
nection with these funnels, but, although I directed my utmost atten-
tion to the point and applied the best lenses at my disposal (Seibert
apochr. syst., 4inmx8),I could not observe any of them. The
majority of the writers wdio have specially investigated this subject
seem to agree in excluding any direct communication between the
cavities of the funnels and those of the surrounding mesenchyma.
In this respect, however, Fraipont makes an exception. He observed
"fenêtre ovale" in the wall of the funnel, by which its cavity was
put in direct connection w^ith the surrounding pseudocoel. I had at
first supposed the end of the fiumel completely closed ; but on
repeated observations with the apochromatic system of Seibert, it
seemed to me very probably open, and to communicate with the
cavities of the mesenchyma. I have not observed any of those peculiar
cells described by preceeding writers.^'

1) On reading Wright and Macallum's description, the question naturally arises if the
writers have not mistaken the ciliated portions of the capillaries, such as have been described

178 s. GOTO.

^ vil. The Nervous System.

With the excellent investi ligation of Lang" on the nervous
system of the Trematodes before nie, I directed my special attention
to this svstem, and can confirm his statement in its general aspect,
though it seems to me to require modification when the writer ex-
tends it to the Trematodes in general. Let us begin with the brain.

As to its position, Lang says, "Ich glaube überhaupt, dass bei
allen Trematoden das Gehirn diese Lage hat, dass es nämlich bogen-
föi-mig über den vorderen Theil des Pharynx verläuft und ich zweifle,
ob sich die abweichenden Angaben bei erneuter, genauer Untersuchung
bestätigen würden." Leuckart-' is inclined to explain those cases where
the brain has been observed behind the pharynx "durch eine Lagen-
veranderuns" " of the latter, " die um so leichter eintreten kann, als
das Nervenband nirgends ringförmig geschlossen ist, obwohl das für
einzelnen Arten behauptet wurde." T find, however, after careful and
repeated observations, with these statements full in view, that in Diplo-
zoon and also in Axhie, Microcotijle, and Oclohothnum^ the brain is a
band-shaped nervous body arching over the oesophagus on the dorsal
aide and heJiind the pharynx. In a fresh specimen, it is seen to be
composed of very thin fibres ; but sections shew that in addition to
these fibres the brain contains a finely granular substance doubtless
identical with the " Punktsubstanz " of the Turbellarians (Fig. 17).
The fibres in the brain are seen to run mainly in two groups, one on

by Looss, for the funnels, and overlooked these latter. I also belive that they go too
far when they endeavor to attribute excretory nature to the large cells observed by Looss and
others in the pharynx of many Tiematodes.

1) Lang — Untersuch, z. vergleich. Anatomie u. Histologie d. Nervensystems d. Plathel-
minthen. Mitth. a. d. zool. Station z. Neapel. Bd. II. 1881. p. 28.

2) Leuckart— I.e. p. 22.

ON DIPLOZOON NIPPONICUM, N. SP. 179

its dorsal side, the other more on the ventral side, close to the dorsal
side of the oesophanfus. These 1 have marked in the figure with a
lightei' shade. They unite at the two ends of the brain where the
nerves take their rise. The l)rain is traversed by numerous dorso-ven-
tral bundles as already inentiiuied.

From the brani are given off nerves both anteriorly and posteri-
orly. One pair (Fig. 2. nai) yn'oceeds anteriorly near the median
line embracing the pharynx, near the nnterior ])art of which it is lost
in the mesenchyma. A second p;nr (nnc) ]^r(U'eeds more laterally, and
can be followed as far as the suckers, externally to which it proceeds
and there withdraws itself from view. These internal and external
anterior nerves are connected with each other by a commissure at a
little distance from their origin in the brain.

Two pairs of nerves also proceed posteriorly, one of which may
be called the ventral pair and is by far the stronger ]iair. The other
pair (nvl) may be called tiie ventro-lateral nerves and proceeds
])osteriorly just at the angle between the ventral and lateral borders of
the body, and can be followed as far as where the two individuals cr(~>ss
each other. The ventral nerves (nv) take their rise in the brain at its
])ostero-lateral corner, and ran be followed to near the posterior bor-
der of the body, rhey become, liowever, more and more indistinct
as tliey proceed ])osteriorly, and finally become invisible at about tlie
level of the hinderniost pair of suckers. They closely follow in their
course the main excretory vessels, (^n whose ventral side they are
situated at a little distance from the muscular layers. At tlie place
where the two individuals cross each other and where the ventro-
lateral nerves withdraw themselves from view, the ventral nerves take
a more lateral position, and this position they keep throughout the
remainder of their course. The ventral nerves are connected with
each other and with the ventro-lateral nerves by a number of com-

180 :.': S- <5oto.

missures occurring nearly at regular intervals ; and in such a
way that each commissure between the ventral nerves lies in a line
with that between them and the ventro-lateral nerves. The ventro-
lateral nerves, atrain, sends ont branches towards the lateral margin
of the body, just at those points where they receive the commissures
from the ventral nerves ; so that all the nerves form a regular I'ect-
angular net-work, and divide the whole ventral surface of the body
into a number of distinct areas. At the points where the commis-
sures cross the main nerves, the course of the fibres is interesting.
From any main nerve, namely, which we may be considering, fibres
are given otf on both sides to the neighl^oring nerves. Beside these,
there are also fibres coming from the latter and proceeding directly
past the main nerve without mingling with its fibres, so that the
four main nerves are probably ])ut in direct connection with one an-
other. I ha\e counted as manv as thirteen commissures in the an-
terior half of the body, in addition to the pair of commissures between
the anterior nerves. In the posterior half the commissures seem to be
less numerous. I have l^een able to count only a few ; but this is
perhaps due to the presence of the strong fields already mentioned
and the special development of the diagonal muscle in this region,
which greatly increases the difficulty of following the course of the
nerves. I have not been able to make out the ])lexus whicli the
nerves probably f )rm on the dorsal side.

As to their histological character, tlie nerves present typical
"Balkenstränge" (Fig- 16)- In some of the meshes are to be seen
sections of nervous fibrils as exceedingl}^ miiuite dots, which are
visible only in the most favorable cases. Pintner'* maintains that
the " Biilkchen selbst" which form ilie mesh-work are the sections of
the fibrils which are probably arranged "reihenweise, nebeneinand-

1) Pintner — I.e. p. 71.

ox DIPLOZOON NIPPOXICUM, N. SP. 181

erstehend." Poirier^ ^ describes the nervous fibres oi' D slnnnm clavatum
as filling up tlie entire cavity of the meshes. But an examination of
the nerves in a fi-esh state shews very distinctly the exceedingly fine
fibrils. They do not seem to be so regularly arranged as Pintner
supposes, and are not at all large enough to fill up the entire cavities
of the meshes. Without doubting the correctness of Poirier's obser-
vation, I am convinced that in Diplozoon the nerves consist of a
frame- work of connective tissue, in the meshes of which i"un tlie true
nervous fibrils. I liave not observed any of the nervous cells describ-
ed by Larig and others in the ner\es. This set me to a careful
searcli after ganglion cells, as these were not also to be found in the
brain, where in other species they make such a conspicuous figure
especially in the peripheral portion. But I have ncjt been able to
find out any to which 1 could decidedly pinnt as nervous cells {Vide
supra p. 171).

VII. The Reproductive System.

We now C(jîne to the consideration of the most complicated
system, the reproductive organs. Of these the female portion consists
of the vitelline body, the ovary, the oviduct, and rhe uterus, with a
"connecting canal" the natiu'e of which is not at all clearly known.
The male portion consists of the testis with a single vas deferens.
I shall beo-in with the latter.

The Male Organs — The testis is a nearly globrdar or ovoid body
situated about midway between the point of crossing of the two in-
dividuals and the posterior niargiji of the body, and is composed of
many lobes. Each hjbe is separated from its neighbour and from

^) Poirier — Coatribution h l'histoire naturelle des Tréiuatodes. Arch. d. zool. expérimen-
tale 2e. Série. T. III. 1885. p. 603.

182 s. GOTO.

tlie .surrounding- niesendn-niM by a layer of dense connective tissue
(Fig. 2, t& Fig. 18). During the winter season, it is a solid mass of
vesicular cells that have assiuned a polyhedral form by their mutual
pressure. Each cell encloses a conspicuous round nucleus, which
seems to be provided with a wall of its own, and in which numerous
chromatin particles are to be observed. The cytoplasma is a hyaline
fluid which scarcely takes up any color. The nuclei are of various
sizes in the same lobe, some being very small, leaving abundant
space for the cytoplasma, while others are of such a size as nearly to
h II up the entire cavity of the cell.

From the anterior end of the testis proceeds a single vas
deferens, which passes anteriorly in a straight coiu'se dorsal to the
oviduct and ventral to the yolk-duct. Daring the first part of its
course, it lies ventrally to the uterus ; but at al)out the level (^f the
anterior end of the ovary, it turns dorsal to it and opens info the vitel-
line duct of the other indiridaal a little more anteriorly than the an-
terior end of the ovary. Zeller" represents the vas deferens of one
individual as standing iu direct connection with the " Laurer's canal"
of the other. But my observations contradict this view entirely. 1
have traced the course of the vas deferens in more than one series of the
sao'ittal sections of the worm. One of these series is reproduced with-
out interruption on PI. XXIIT. The opening of the vas deferens of one
individual into the yolk-duct of the other is seen in Figs. XI. &
XIII. By the same series of sections, the opening of the connecting-
canal of the oviduct into the intestine is distinctly seen (Figs. V &
XIX). The vas deferens is destitute of any distinct wall of its own.
It seems to be merely a continuous tube-like cavity in the general
mesenchyma., and to collapse entirely during the winter season.

^1 Zeller — Ueber deu Géschlechtsapparat des D/pZ. 'parado.vuni. Ztschr. f. w. Zool. Bd.
XLVI. 1888. p. 233.

ON DIPLOZOON XIPPONICUM, N. SP. 183

The Female Organs — The ovary (Fig. 2, ov) is a long conico-
cjdindrical body which is doubled on itself by its middle portion, so
that the two ends come close to each (jther, and placed on the dorsal
side of the body just anterior to the testis, to which its smaller end
is closely applied ; the anterior end where it is doubled on itself
reaching as far as where the dorsal jjapilla formerly was. From its
larger end, wdiere ripe ova are found, proceeds the oviduct. As we
approach the other end, the ova become smaller and smaller until
finally we see a mere assemblage of round nuclei imbedded in a com-
mon mass of protoplasm. The whole ovary lies in a mere cavity of
the mesenchyma without any distinct wall of its own. A section
through the larger end (Fig. 21) shews the ovary to be a solid
body consisting of large ova which are either polygonal or wedge-
shaped according to the direction of the section. Each ovum is des-
titute of any membrane, and consists of a mass of homogeneous deeply
stained protoplasm, in which lies a large vesicular nucleus provided
with a distinct wall and containing a hyaline fluid in which float
numerous deeply stained dots, the chromatin {)articles. Each nucleus
again encloses a large deeply stained nucleolus in which are again to
be observed one large or a few smaller vacuoles. Zeller^' mentions
and figures in the ovum of Pokjstomum and Diplozoon a thick, elastic
" Hiille " ; but I have no doubt that the ovarian ova of Diplozoon are
destitute of any membrane. This is also the case in Axine, Micro-
cotijle. Octohothrium, Dactylocjijrus^ in fact in all the species of ectopara-
sitic Trematodes I have hitherto examined. Willemoes- Suhm^' men-
tions no '' Dotterhaut " in I'ohjstomuin occllatam, ; Taschenberg"^* asserts

1) Zeller— Ztschr. f. w. Z. Bd. XX 11. p. 5 & 169 foot-note; Bd. XLVI. p. 235. Is not the
elastic membrane the result of fertilisntion?

-) Wiilemoes-Sahui — Zur Naturgescn. d. Polyst. integerrimvim u. Polyst. ocellatum.
Ztschr. f. w. Z. Bd. XZII. 1872. p. 33.

^) Taschenberg — Beiträge, p. 36.

184 s. GOTO.

the absence of any membrane in the ovum of Tristomum ; so also
Wierzejski'' in that of Calicofiße Kroijcri ; and I believe the same is true
of the ova of all ectoparasitic Trematodes. As we proceed nearer the
smaller end of the ovary, the ova and their nuclei become smaller
and smaller, the vacuoles within the nucleolns disappear and finally
the nucleolns itself, until we see only spherical nuclei crowded to-
gether and surrounded by a common mass of uniform protophi>;m.
Fig. 22 shews a section through this part.

The oviduct proceeds from the larger end of the ovary and takes
its course posteriorly and to the right, ventral to the vas deferens and
the testis. At a short distance from its origin, it receives a canal
(Fig. 2, cc) which proceeds anteriorly and, after making a slight wind-
ing or two, opens into the intestine (Figs. V & XIX). This is the
" Laurer's canal" of Zeller which he represents as standing in direct
connection with the vas deferens of the other individual. In Poljisto-
inuni integer rimuur^ he asserts a direct connection between the ovary
and the testis ; and in proof of this he alleges his observation of the
ova passing through the oviduct and Lanrer's caiial and entering the
cavity of the testis. But it has been pointed out by Ijima'"' that the
canal in question distinctly opens into the intestine, and that a similar
canal is present in many other species of the group ; and I can con-
fidently state from my own study that the ''dritte Dottergang" of
Lorenz in Axineixwà Microcotyle distinctly opens into the intestine. A
similar connecting canal is also present in a species of Octohothrinm
which I have examined."*' The fact cited by Zeller can be explained

1) Wierzejski— 1. c. p. 558.

2) Zeller— Weiterer Beitrag z. Kenutniss d. Poljstomeen. Ztschr. f. w. Z. Bd. XXVII.
1876. p. 245.

3) Ijima — Über den Zusammenhang d. Eileiters mit d. Verdauungscanal bei gewissen
Polystomeen. Zool. Anz. Jahrg. VII. 188i. p. 635.

4) Voeltzkorw (Arb. a. d. zool.-zoot. Inst, in Würzburg. Bd. VIIT. 1888. p. 267) describes
an evidently homologous canal in Aspidogaster conchicola. According to him it ends blindly
near the dorsal surface of the worm. He calls it Receptaculum vitelli.

ON DIPLOZOON NIPPOXICUM, N. SP. 185

if we consider that the intestine is destitute of nny distinct
wall, and that when the testis is nearly empty there is ahnost nothing
that would prevent the entrance of the ova into the cavity of the
testis by way of the intestine. I therefore believe, notwith-
standing his positive statement to the contrary, that the canal in
question opens also in Polystomum into the intestine at the point where
he represents it as arising from the testis. In Dactijlogijrus a similar
canal opens externally on the dorsal side, at a short distance from
the right lateral margin of the body. In Dipl. paradoxum this canal
is very long and undergoes numerous convolutions, but in nipponicum
it is shorter and nearly straight, and the internal surface is clothed
with cilia. Its nature and function, if it has any, I hope to be able
to treat of later. At a little distance from the point where
it receives this canal, the oviduct receives also the yolk-duct (yd).
After this it continues its former course, and then, making a sudden
turn anteriorly, opens into the uterus.

The uterus, under which I include both the " Ootyp "' and the
" Eiergang " of the German writers, is a cylindrical tube with a dis-
tinct wall which is thickly beset for the greater part of its length with
long cilia on its internal surface. It shews an ovoidal enlargement at
its origin, the " Ootyp," then diminishing in diameter proceeds an-
teriorly, following the same course as the vas deferens, and opens ex-
ternally by a small aperture on the ventral side just at the angle
formed by the ventral side of one individual with the dorsal side of
the other, at the top of a conical elevation which is sometimes very
small, sometimes larger and very conspicuous. Just before opening,
it presents a second enlargement in which a single egg is usually found
during the period of reproductive activity. "II y a à l'origine de ce
conduit (i. e., of the uterus) une sorte de pylore," says v. Beneden.^'

1) V. Beneden — 1. c. p. 43.

186

s. GOTO.

This is caused by the. opening at this point of numerous flask-shaped
unicellular glands (shg), fjie shell-glands. The wall of the uterus
proper (Ootyp, Fig. 24) is lined by a distinct epithelium, whose
cells contain each a round nucleus projecting into the internal cavity.
The protoplasm is granular and no cell-boundaries are to be seen.
The epithelium sits on a distinct basement membrane and is destitute
of cilia. The remainder of the uterus (Eiergang) is provided with a
similar wall (Fig. 23), with the nuclei, however, more separated
from one another. Here, as already stated, the wall is beset with
long cilia.

The vitelline body is an extensive lobed body (Fig. 2, vb) situat-
ed exclusively in the anterior half of the body, all around the
intestine both on its dorsal and ventral sides. In specimens in wdiich
reproduction is going on, each lobe is seen, when fresh, to contain a
dark granular mass. Sections (Fig. 19) shew that each lobe consists
of a number of cells containing numerous yellowish granules,
each with a nucleus and a nucleolus in the centre, and a thin cell-
wall. These are the ripe yolk-cells, îmd when freed take up a globular
form. In the peri|)heral portion iu*e seen smaller cells wdth a deeply
stained protoplasm, a nucleus and a nucleolus. The protoplasm is
homogeneous, finely or coarsely granular according to their different
stages of development. They are the young yolk-cells ; and there are
also to be observed cells intermediate between these two kinds — cells
one half of whose content has already been changed into yellowish
yolk-granules while the other half still consists of granular protoplasm.
During the winter months, the yolk-cells present a quite different
appearance (Fig. 20). Tliey are then scarcely to be distinguished
from the cells of the mesenchyma of certain species of Microcotyle.
They are then of a polygonal form, with a distinct cell- wall, a round
nucleus and nucleolus, and a granular protoplasm which stains very

ON DIPLOZOON NIPPONICUM X. SP. 187

well. In this granular protoplasm there are fibrons structures radiating
from the central nucleus to the cell- wall and more or less forminsr a
net-work. The steps hy which these cells are changed into ripe yolk-
cells and the origin of the deeply stained young yolk-cells I must
leave at present unexplained.

As will he immediately seen from the above investigation, the
union of Diplozoon is, as Zeller maintains, a permanent copulation.
But the relation in which he has represented the parts of the two
individuals to stand to each other reqaii-es correction. We have seen
that the vas deferens of one individual opens into the yolk-duct of
the other. This is well in accordance with the probable mode of
copulation in some allied forms. In ]\[icrocotiile, which seems to be
very closely allied to Diplozoon, there is a dorsal vagina which leads
into a canal opening into the yolk-duct. In this canal I have often
observed spermatozoa, and as during the period of reproductive activi-
ty yolk-cells are constantly going down the yolk-duct and push down
before them anything that might come u}) from below, it is very prob-
able that these spermatozoa had found their way here from the dorsal
vagina. Hence the supposition is very natural that in copulation the
penis of one worm is directly applied to the dorsal vaginal opening of
the other. Now if this very probable supposition be true, and if
we further imagine such a relation to persist permanently, we
should have j List the case that we actually see in Diplozoon, with the only
difference that the copulation is not cross-wise. Whether in Micro-
cotijle also, as in Poliistoniiun, the copulation is normally cross-wnse
and mutual is well worthy of our attentive observation, since if this
be the case, the copuhition of Diplozoon would l)e nothing more or
less than the regular mode of copulation in. allied forms made per-
manent.

188

s. GOTO

In conclu sien I wish to express my best thanks to Prof. K.
Mitsakiiri and Prof. C. G. Knott for kindly looking through my
paper and making suggestions.
Tohjo, October 1890.

ON DIPLOZOOX NIPPONICUM, N. SP. 189

Explanation of Figures.

Abbreviations common to all the figures.

as ascending stem of the excretory vessel (according to the

direction in which the contained fluid moves).

hv brain.

cc connecting canal between the oviduct and the intestine.

dem dorso- ventral muscle.

ds descending stem of the excretory vessel.

dc digestive cell.

eo excretory opening.

int intestine.

mce external circular muscle ]

, [ of the phai-yiif'-eal wall.

mci internal ,, ,, i

mo mouth.

nae external anterior nerve.

nai internal anterior nerve.

nv ventral nerve.

nvl ventro- lateral nerve.

ov ovary.

ovd oviduct.

oe oesophagus.

ph pharynx.

pph prepharynx.

of the posterior suckers.

19Ö S. GOTO.

pm median piece

ppa paired anterior piece

ppp „ posterior ,, ,

pp process of the posterior piece j

sa anterior sucker.

sp posterior sucker.

sg sticky gland.

shg shell gland.

t testis.

ut uterus.

vd vas deferens.

vh vitelline body.

yd yolk-duct.

All the figures, if not otherwise stated, were drawn with cam.
lue, Zeiss Ex 2.

PI. XXI.

Fig. 1. — Dipl. nipponiciuu killed with boiling sublimate; free-hand,
surface view, x about 14. The black dots represent the digestive
cells seen throuo'h the tissues.

Fig. 2. — The same, free-hand, from a specimen killed under the
cover-glass, shewing the internal organs, half-diagramatic. The
right antérieur half [)resents the ventral, and the corresptjnding
posterior half the dorsal aspect ; and vice versa with the other
individual. The nerves are colored yellow ; the excretory ves-
sels indigo- blue.

Fig. 3. — Chiti nous frame- work of the posterior sucker as seen in a
specimen under the cover-glass.

Fig. 4. — Section of the postericn' sucker in the direction indicated by
ab in Fig. 3.

By inadvertciicc of tlie printers, the micle(ili in Fi^'s. 8, 9, 11,
lo, Iß, 17, 18, 19, 20, 23, and 24 are represented either as lyino-
outside the nuclei or quite eccentrically in Tliem, whereas they
ou<i'ht to occupy more central positions. Their true positions are
indicated in most of the fi<,aires hy weakly shaded dots.

ON DIPLOZOON NIPPONICÜM, N. SP. 191

Fig. 5. — Section of the same in the direction indicated by xy in

Fig. 3.
Fig. 6. — Hooks between the posterior suckers.

PI. XXII.

Fig. 7. — A part of a cross-section of the worm passing through one
of the conical elevations of the epidermis. It also passes
throug-h one of the transverse folds into which the surface of the
body is thrown when the animal contracts ; hence the longitudinal
muscles are separated from the circular by a rather thick layer
of connective tissue.

Fig. 8. — Section of the sticky gland, from a cross-section of the
worm.

Fig. 9. — Section of the anterior sucker, from a cross-section of the
worm.

Fig. 10. — To shew the direction of the diagonal muscular fibres,
from a horizontal section of the worm.

Fig. 11. — Cross-section of the pharynx.

Fig. 12. — Diagram shewing the muscles accessory to the pharynx.

Fig. 13. — A part of a cross-section of the worm, from the posterior
half of the body, a little posterior to the testis ; to shew the
character of the mesenchyma.

Fig. 14. — The portion of the mesenchyma situated between the ovary
and the testis, from a sagittal section of the worm.

Fig. 15. — Excretory funnel, Seibert apochr. sys. 4 mm. x8.

Fig. 16. — Cross-section of the ventral nerve.

Fig. 17. — The brain, from a cross-section of the worm.

Fig. 18. — The testis, from a longitudinal section of the worm.

192 s. GOTO.

Fig. 19. — The vitelline body, from a longitudinal section of a worm

collected in September.
Fig. 20. — The same, from a cross-section of a worm collected in May;

the yolk-cells not ripe.
Fig. 21. — Section of the ovary near its larger end.
Fig. 22. — The same near its smaller end.
Fig. 23. — Longitudinal section of the uterus (Eiergang).
Fig. 24. — Cross-section of the uterus proper (Ootyp).
Fig. 25. — A cross-section of the worm through the region of the

pharynx ; to shew the peculiar gigantic cells. Zeiss Dx2.

PI. XXIII.

An uninterrupted series of sagittal sections of the worm. To
the respective abl)revnations is subjoined the letter r or I according as
the parts belong to one or the other individual. Zeiss Bx2.

Jour. Sc. Coll. Vol. IV. PI. XXI.

Fig. 1

m "

Piih

Ph

S. Goto del.

Fi;,

«V

Jniir Sc. Cntl. Vnl IV. PI ïyi.

;?s

A New Species of Hymenomycetous

Fungus Injurious to the

Mulberry Tree.

by

Nobujirö Tanaka.

With Plates XXIV— XXVII,

In Japan the mulberry tree has been widely cultivated, from time
immemorial, for rearing: silkworms. Althouo'Ii the methods of its
culture have much improved, yet its diseases, especially those caused
by fungus parasites, have been overlooked even by skilful cultivators.
The chief reason f )r general neglect ivgarding these points is the
want of accurate knowledge of the nature and biology of fungi.
One of the most serious diseases of this kind is that which is known
under the name of " Mompa-byö."* This disease has produced much
distress at various intervids for about eight years, in the experimental
farm of the Agricultural College at Komaba, Tokyo. Some dis-
tin^fuished biologists and au'riculturists have investifrated its nature,
and stated that it was due to the ravages of a sterile mycelial stage
of a fungus of some other form, but its true nature has never yet been
fully explained. I have lately had the good opportunity to study
this disease under the direction of Prof. R. Yatabe. The object of
this paper is to deal with one or two of the unsettled questions
regarding it, namely, the morphology of the perfect condition of the
fungus which causes the disease, and its systematic position.

* Mompa, a kind of nappy cotton-cloth ; byö, disease.

194 N. TAN AK A.

Towards the end of last year, I obtained specimens of mul-
berry trees attacked by the disease, but unfortunately the speci-
mens were so far advanced in decomposition that the course of the
mycelium of the fungus in its relation to the internal tissues of its
host was not clearly definable, and also the fructification of the fungus
could not be found. Since then I have examined many other speci-
mens, up to the beginning of April of this year, and at length found
the perfectly developed condition of the fungus. Its specific characters

are as follows : Pi lens sessile, resupinate, somewhat orbicular

or oblong, often iircgnlarly lobecl, 5-10 cm. across, 2-4 mm.
thick, at first velvety and membranaceous, then subcoriaceous, some-
what convex, incrustate, purplish brown, at length albo-pruinose ;
hymenium wdiite ; basidia curv( d, 1-o-septate, tetraspored ; sterig-
mata elongated ; spores ovoid, curved, hyaline, 10-12 n. long, 5-7
U. broad.

By the above characters, especially ])y its peculiar form of b.asidia
and by its nature, I consider that this fnngus belongs to the genus
Helicohasidium in the fiimily Thelephoreœ of the Hiimenomifcetes. It
has much resemblance in its characters and hal^it to many species of
its allied genera ; but it can be distinguished from Thelephora and
Gorticeum chiefly by having an intermediate stratum in the pileus, and
from Stereumhy hny'ing a usually superior hymenium. Of the s[)ecies
of the genus Hclicobasidiuni liut few are known ; in Saccardo's Sylloye
Fungonim * only two species, H. purpureum (Tul.) Pat. and H.
cirratum Pat. et Gail.^ ru-e given. By comparing my description oî
the fungus with that of the above named species, it can be distin-
guished from the forniei- chiefly by the colour of the pileus and the
number of spores borne on a basidium, and, from the latter, by the
diameter of the pileus, the number of spores borne on a basidinm,

* Vol. VI. p. 666.

A NEW SPECIES OF MULBERRY FUNGUS. 195

and their size. An allied fungus on the mulberry tree in South
Carolina, North America, was described by Prof. Berkely under the
name of Stereum moricolum ; and two other species of Sterevm^ viz.
S. suhcruentatum, B. et C. and .S'. contvariiim Berlc, are given in
Saccardo's Sylloge* These are Japanese species, but unfortunately
I have never yet found them. They must, however, be very distinct
from my species. For these reasons i venture to call it Helicohasidium
Mompaf from the well known Japanese name of the disease.

The fungus at first attacks the root of a living tree, and the
diseased tree shows external symptoms of the disease on portions above
ground : usually the growth of shoots is arrested, the newly develop-
ed leaves become gradually smaller and at length die otf ; then the
lower part of the shoots begins to die, though the bark higher up
may preserve its normal appearance. It takes a tree one or two
months to reach this state, after it has first shown the external symp-
toms of the disease.

On uprooting a young mulberry tree badly attacked by the
fungus, the roots are found to be killed from below upwards, and
present the appearance represented in Plate XXIV, Fig. I. The
tree figured there is three years old ; the roots marked a have grown
three years, and those marked h and c are of this year. The
portions marked a' are dead roots, whose bark was already severely
injured and so loose that it was separated by the act of uprooting.
As these dead roots were of no use to the tree, it produced the new roots
h to absorb nourishment from the soil. But the newly formed roots
were also injured as the disease advanced, and became unfit to per-
form their function ,• and at length another crop of newer roots c was
produced higher up, by means of which the tree was enabled to

* Vol. IV. pp. 507 and 579.
t See p. 193.

196 N. TAX AK A.

-sustain its life. In the state just described no fructification of the
fungus is yet observable, although its subterraneous vegetative
mycelia are actively growing.

After the fungus has Ijeen growing in this manner for some time,
flat irregular disks of mycelia begin to form under certain circum-
stances on the aerial portion of the tree at the bases of the shoots.
These disks are the first stages of the |)ileus. The successive stages
-of growth of the pileus are shown in Plate XXIV, Figs. 2, 3, and
4. It first appears as a thin etfused mass of mycelia of a dark purplish
brown colour, haviug a paler margin of definite outline, and present-
ing a smooth velvety appearance (PL. XXIV, Fig. 2, a). It surrounds
the basal [)art of the shoots of the diseased tree to a height of 15
cm. or more, sometimes leaving here and there small narrow portions
unG<wered. It often encloses in its embrace some extraneous matter,
such as decayed leaves, branches, and the like, together with particles
of soil. As it gradually develops, it forms generally an irregular
r(3undish flat disk, one part of which stands out at right angles from
the surftce of the shoot, while the other remaining part is firmly
attached to it. The projecting part of the [)ileus then expands
laterally either on one side of the shoot or on both sides ; and as the
shoot is usually bent horizontally at the base, the pileus becomes also
horizontally expanded. The hymenium is produced on the free sur-
face of the pileus, on the u{)per and lower sides of the projecting
parts, as well as on the exposed side of the part fastened to the shoot.
The fully developed pileus is of a whitish colour tinged with violet ;
the projecting part is about 5 mm. thick, and its upper surface is more
uneven than its lower surface (PL. XXV, Figs, 1, 2).

By carefully detaching the young pileus from the substratum,
numerous mycelial strands of unequal thickness may be observed on
its lower margin (PL. XXV, Fig. 3). These strands are found on

A NEW 8PECLES OF MULBERRY FUNGUS. 197

almost every portion of the diseased roots, forming irregnlnr networks
of various complexity (PL. XXV, Fig. 4). They are ;^-l mm. thick
and of a purplish brown colour like the young pileus ; and as to tlieir
mode of ramification there seems to be no regularity. Without des-
troying even their finest brandies, they can be very easily detached,
with a needle, from the roots upon which they grow, t() a length
of several centimetres (PL. XXV, Fig. 5). They are often found
free, either forming large groups in spaces left between the partly
detached cork layers of old diseased roots, or solitarilv in the soil.

The microscopical structure (jf the mycelial sti-and is ditferent
from that of Agaricus iiielleus, whose minute details are now well
known from the excellent description given by the late Prof. De
ßary.* Li the ])resent species the axial portion of the mycelial strand
consists of thick-walled hyplue, 3 U. in diameter, mixed with a few
finer ones ; and the peripheral portion consists entirely of finer hyplue
(VL, XXVI, Fig. 1). Ill the transverse section of the strand this is
more clearly seen (PL. XXVI, Fig. '2). In the mycelial strand of
Agaricat) vielleiis the hyjiliœ are so compactly arranged as to form a
tissue as is clearly seen in the cross section ; + but in the present
species the hyphas composing the strand are so loosely put together
that they easily separate from one another, and in the cross section
they present a circular and not angular form, since they are not
pressed together so as to assume the latter form. Moreover the form
of the cross section of the strand in Agaricus inellcas is round, but in
this species it is flattened. The thickening of the strand is ett'ected
either by the copious branching of a single hypha (.)r by the coales-
cence of two or more strands. In the group of hypha; formed by the
first method, there is always an axial or original thick hy[)lia

* De Bary, Vergl. Morphol. u. Biol. d. Pilze ; Eng. traus. p. 2H-29
t See Fig. 11, p. 2t, of the baiue book.

19S

N. TANA K A.

surrounded by finer ones wliich liave been produced by its ramifica-
tion (PL. XXVI, Fig. 3). As the strand grows, the branches of the
original hypha also ramify ; and the secondary branches thus produced
surround the primary branches, just as the latter surround the
original hypha. In this way branches of higher orders are succes-
sively produced, and surround the branches of the next lower order.
Ordinarily the branches of the liyjiha^ grow in one direction, but
occasionally there are found those that grow in two opposite directions
from the point of origination (PL. XX\^I, Fig. 4). The older
hyphai or those lying towards tlie center of the strand are much
more darkly coloured than the younger or those of the periphery. The
mycelial strand of the fungus is found only on the surface of the host.
When it makes its way into the tissues of the latter it usually forms
longitudinally elongated masses, such as are seen in the interstices be-
tween the cork layers of the host (PL. XXVI, Fig. 9). Similar mass-
es are also found on the surface. These masses of the hypha3 spread
widely in the cambium zone and in the yourjg bast, forming membrane-
like expanded networks of wliitish mvceli:i. These mycelia send out
single colourless hypha*, L5-1 U. in dinmcter (PL. XXVI, Fig. 5),
into the rind and wood, and especially into the dotted vessels. They
also send out masses of coloured hypha3 to the surface of the host,
from which are again developed ordinary external mycelial strands.

Crystalline spheres of calcium oxalate, ^-\ mm. in diameter
(PL. XXVI, Fig. 6), are found in great numbers on those places
where the white mycehal membranes abound. They consist of an
enormous numher of somewhat radially arranged wedge-shaped
crystals (PL. XXVI, Figs. 7, 8), ench of which is 20-30 u. long and
10-15 n. broad. If we examine one of these crj'stalline spheres
under the microscope, taking care not to crush it, we see only the
sides and broader ends of the wedge-sha[)ed crystals ; and by crnsliing

A NEW SPECFES OP MULBERRY FUNGUS. 199

it we can recognize the radial arrangement of tiio crystals. Prof. De
l>ary has described crystalline spheres of a similar nature found in the
narrow cylindrical hyphaä of the mycelium of Fhallus caniims.*
Crystals of calcium oxalate of other forms, such as regular quadrate
octohedra, rod-shape, &c., are also found in great abundance in the
same place where the crystalline spheres are found.

The mycelia of the fungus form an em^rmous nnmber of
sclerotia in all parts of the diseased portion of the roots (PL. XXVII,
Fig. 1, (t). The sclerotia are irregularly roundisli bodies 1-4 mm. in
diameter, and are dark purplish brown in colour. If tlie nourish-
ment in the sap-containing layers of the host plant becomes scanty by
the parasitic action of the fungus, and also when the vegetative
activity of the host plant is diminished in autumn, the interior of the
lenticels and the interstices between the cork layers become filled with
the sclerotia of the fungus, while the mycelial strands which remain
outside spread widely on the surface of the roots. By carefully
detaching the mycelial strands we can ascertain that they have no
direct communication with the sclerotia. The number of sclerotia is
different in different parts of the roots, according to the degree of the
injury done by the fungus ; and the greater the degree of the injury,
the greater the number of the sclerotia. The formation of sclerotia
does not take place on the outside of the host plant, but always in the
inside or in the spaces partly exposed by the formation of fissures
(PL. XXYII, Fig. 2). The sclerotia have a dark l)rown rind (PL.
XXVII, Fig. 3, i), and a medulla of white soft tissue (Fig. 3, (i) with
a few air-conducting passages. The hyplia? of the medidla are
cylindrical and septate, anastomosing with one another in a rather
loose manner (Fig. 4, ci), and are 4-5 U. in diameter. Towards the
surface of the sclerotia, the medulla passes gradual I v into the rind,

* De Bary, Vergl. Morphol. u. Biol. d. Pilze, Eng. traas. p, 11.

200

N. TANAKA.

which consists of thicker- walled nnd shorter-celled hyphœ, forming a
compact tissue without interstices (Fig. 4, b). In its younger stage
the surface of the rind is felted over wnth the remains of dead hyphae
(Fig. 4, c). A series of five ditferent colours — white, yellow hrown,
dark brown, rose violet, and dark violet hrown — may he seen in the
ordei- stated, from the centre outwards in the section of the Sclerotium.
As the mycelial strands gradually grow upwards, they aggregate
into a few flat thick strands, more than 1 mm. broad. These strands
spread themselves from the a|)ices and unite into a thin broad layer,
consisting of reticulated hyphal filaments and covering the base of the
shoots of the host plant. As the development of this layer proceeds,
the pi lens is formed from it. The pileus is an irregularly roundish
flat disk witli a smooth velvety surface, and takes a purplish brown
colour, leaving its margin whitish (PL. XXIV, Fig. 2, a). Thin
radial sections of a ftdly developed pileus, show that its medullary
stratum is composed of loosely anastomosing branclied hyphœ, dark
violet brown in colour, and 3-4 u. in diameter (PL. XXYII, Fig. 5).
Towards the outer surface of the pileus these hyphte take a vertical
position, and produce short and blunt branches (PL. XXVII, Figs.
6, 7). These branches of hyphœ are colourless and shortly septate,
and form the hymenial layer. Some of them elongate here and
there, and form the basidia, which are curved and 5-8 u. in diameter.
From the convex surface of the basidium are produced four sterigmata,
which are pointed, slightly curved and 6-10 U. in length (PL.
XXVII, Figs. 8, 9, 10). The spores are formed singly on the apices
of the sterigmata ; they are ovoid, curved, 10-12 u. long and 5-7 u.
broad (PL. XXVII, Fig. 11). The portion of the pileus attached to
the substratum produces hairs or rhizoids on its inner surface, which
penetrate into the substratum. But the horizontally projecting part
of the pileus produces the hvmenium on both surfaces, when it does

A NEW SPECIES OF MULBERRY FUNGUS. ^^01

ii«)t lie uni on the üronnd. The internal .structure of these two
]VM-ti(^ns is, however, essentially the same.

In the merlullary stratum of the pileus which lies on the an^i'ud,
an iiumense number of minute algae, belonging to the genera Conferva
and Profococrus (PL. XXVII, Fig. Iß) are found in groups, very
much like the gonidia of Lichens. On the higher parts of the stems
and branches of old mulberry trees, are frequently found orbicular and
lirownish purple patches, from 1-10 cm. in diameter ; they are com-
monly called " Köyaku-byö " * of the mulberry tree. They resemble
very much in their structure the young ])ileus of the speciees of
Helicohafiidium in question, except that the hyphfe in the pileus of
the foriuer are more slender than those of the latter, being only 2-8
U. in diameter (PL. XXVII, Fig. 12). The sterigmata of the former
are also very minute; and I hv.\e not been able clearly to determine
their number on a basidinm (PL. XXA'II, Figs. 13, 14). Besides
the ordinary slender basidia. 3 u. in diameter, much thicker and
segmented basidium-like extremities of hyphîe bearing no sterigmata
are often seen in the hymeninm (IM.. XX VII, Fig. 15 h Whether
the orbicular patches just described simply represent a form of tlw
present species or not can only be determined after further investiga-
tion. But I venture to say that it is probably a poorly nourished
form of the latter.

In conclusion, I wish to express my thanks to Prof. K. Yatabe
who has helped me throughout my work with valuable suggestions.

* The Japanese word kaijaku means a medical plaster; byô, disease.

202 X. TAXAKA.

Explanation of Figures in Plates XXIV — XXVII.

Plate XXIV.

Fiq. 1. Sketch of tlie base of a young mulberry tree, injured by
the disease at the roots a, }>. The upper portion a' and the roots c are
free from the disease ; the lower portion a of the roots a is completely
disorganized. Beduced.

Fig. 2. Portion of the base of a shoot, showing the young
pileus a of the fungus. Natural size.

Fig. 3. More advanced stage of a similar pileus with its pro-
jecting parts a. Natural size.

B'ig. 4. Mature form of a similar pileus ; a its projecting part;
b its basal ])art. Natural size.

Plate XXV.

Fig. 1. Mîiture form of the pileus of the fungus, showing its

upper surface. Natural size.

Fig 2. Lower surface of the same. Natural size.

Fig. 3. Young stage of the pileus carefully detached from its

substratum. Natural size.

■/'':: Fig. 4. Portion of a diseased root, with mycelial strands of the

iiTDgus. Natural size.

- - Fig. Ô. Portion of the mycelial strands detached. Natural size.

Fig. 6. Group of mycelial strands. Natural size.

Plate XXVI.

Fig. 1. Hyphee of mycelial strands. x 440.

F'ig. 2. Cross section of the same. x440.

A NEW SPECIES OF MULBHRRY FUNGUS. 203

Fiff. 3. Hypha3 of mycelial strands, showing the mode of
ramification. X440.

Fig. 4. A kind of branching in a similar hyphn. x 440.

Fig. 5. White hypha3 in the tissues of the host plant.

X 440.

Fig. 6. Crystalline spheres of calcium oxalate. xö.

Fig. 7. A similar sphere much magnified. x240.

Fig. 8. Wedge-shaped crystals B of the same ; A showino-
their radiating structure. x240.

Fig. 9. Ma.sses of coloured mycelia a in the interstices of cork
layers b. x 10.

Plate XXVII.

Fig. 1. Portion of a diseased root, with numerous sclerotia a of
the fungus. Natural size.

Fig. 2. Longitudinal section (jf the bark of a root, showing the
formation of sclerotia. x 5.

Fig. 3. Vertical section of a Sclerotium; a, medulla; ^, rind ;
c, remains of hyphas. x50.

Big. 4. Portion of the same, showing its tissues ; the letters
correispond to those in Fig. 3. x440.

B'ig. 0. Hyphas in the medullary stratum of the pileus.

x440.
Figs. 6', 7. Hypha^ in the hymenial layer of the pileus.

x440.
Figs. 8, 9, 10. Basidia with sterigmata and young spores.

X440.
Fi(j. 11. Mature spores. x440.

Fig. 12. Hypha' in the medullary stratum of the orbicular

204 N. TANAKA.

p<atcheH on the higher parta of the stem aud branche« of an old
mulberry, tree. x440.

Figs. 13, lé. Basidia of a similar patch. x 440.

Fig. 15. Basidium-like hypha of a similar patch. x440.

Fig. 16. Alga3 in the medullary stratum of the pileus ; A, Con-
ferva; B, Protococcus.

Jour. Sc. Coll. VoJ. IV. PI, XXIV.

N. Tanaka del.

Jour. Sc, Coll. Vol. IV. PI. XXV.

X. Tiiiinka do

Jour. Sc. Coll. Vol. IV. PI. XXVI.

Fig. -5.

N Tanaka .1h|

Jour. Sc. Coll. Vol. IV. Pi, XXVII.

N. Tniiaka <lel.

Notes on the Irritability of the Stigma.

by
M. lYliyoshi, Rigakiishi.

Witli Plates XXVIII— XXIX.

It is filready known l)y the researches of Meckel* and others
that the hifid stigmas of certain plants, such as Martijnia, Biguonia,
and especially of some Scrophularinece, e.g. Miinuhis, Torenia, Gratiola,
are irritable to touch. But as our knowledge on the subject is still
scanty, it will not be superfluous here to state some of my observa-
tions on this subject. The plants I studied were 3[azus rvgosus, Lour.,
var. macrantlia, Fr. et Sav., Miimdus iiepalensis, Benth.. M. sessifolms,
Maxim., M. 'mnschatiis, Dong. I shall give in detail only the case of

* • Mazus rugosns, Lour., var. macrantha, Fr. et Sav.
(PI. XXVIIl, Fig. 1.)

The plant belongs to Scrophnlarinfce, and may be briefly de-
scribed as follows : —

A low annual. Branches prostrate, 5-40 cm. long, often r(^(^ted
at the nodes. Leaves exstipulate, sparingly hairy, coarsely and
irregularly dentate ; the radical sessile, cuneate-spathulate, 1-4
cm. long, 0.5-1 cm. broad ; those of the branches opposite, some-
times alternate, obovate, narrowed int<3 the cuneate base, smaller
than the radical leaves. Flowering stems, erect, more or less

* E. Hecke] :— Du mouvement dans les stigmates bilobiés des Scrophularinéps, des Bigno-
niacées et des Sésamées. (Comptes rendus, t. LXXIX. 187-4. No. 12, P. 702-704)
** The Japanese name of this plant is Sagigolie.

206

M. MIYOSHI.

]'ube>Gent. 5-25 cm. hiii'h. Flowers distant ; ])edicel8 bracteate,
minutely piil;escent, 1-2.5 cm. long- ; bracts minute, scaly, acute.
Calyx (Fil!-. H) 5-lobed, cainpanniate, persistent, 0.7 cm. Ioii<>- ;
lobes ovate, acute, at length spreading. Corolla (Fig. 2) bilabiate,
light blue or often deeper-coloured, sometimes snowy white.
Upper lip erect, or curved upward, bifid at the apex. Lower
lip deflexed, o-lobed, 1-1.5 cm. l)road and long; the two lateral
lobes broader than the mi<ldle ; its palate convex, beset with
delicate hairs, whitish or yellowish, with yellowish brown or deep
brown S|)ots. Stamens (Fig. 2) 4, didynamons, inserted in the
tube of the corolla, distinct at hrst, eacli pair connivent and adher-
ing by the anthers at maturity ; anthers wliitish, 2-celled. PolJen-
grains (Fig. 6) whitish yellow, elliptical, with 3 longitudinal
grooves. Style (Fig. 3) longer than the stamens, ascending under
the upper lip of the corolla. Stigma (Fig. 3) 2-lobed, lobes
semicircular, 1 mm. long. Ovary superior, globose, 2-celled,
Capsule compressed, loddicidal. Seeds numerous, minute, brown.
The plant is common everywhere, especially on sunny lawns,
and bears flowers from early spring to mid-summer. When I hap-
pened to notice* the irritable property of the stigma and began my
observations early in April, I visited daily certain spots in the Uni-
versity grounds where I found the plant in profusion, some growing
in positions \ery convenient for examination.

To observe the phenomenon, take the flower of this plant, and
touch the lower lobe of the stigma with the point of a needle or the
like ; we shall then see the affected lobe move steadily upwards with
uniform speed until it comes in close contact with the upper lobe (IM.
XXIX, Figs. 7, 8, 9). We may cause the same action with the least

* A few weeks after, I was iuformed by Mr. T. Yoshiuaga of Tosa, of the same fact whieli
be had himself observed.

NOTES OX THE IRRlTABltlTY OF THE STIGMA. 207

possible toucli as, for example, witli the tip of a l)ri.st]e or hair. On
the other hand, plaeing a small drop of water on the stiginatie lobes
or blowino- upon them does not iiidiiee the motion. Again, mere rub-
bing on the style or on the (joter surface of the hjlies does not show
even the least sign of motion, though a slight touch on the inner sur-
face is very effective. Moreover, this curious property is not contined
to the lower lobe only, as may at first sight appear, but it is possessed
by the upper lobe as w^ell. Since the lower lobe is widely reflexed, the
motion there is very manifest ; but the upper one being nearly in the
same line with the style shows no decided motion other than a
slight bending down.

I made these experiments on the natural position of the flowers,
and measured the time required for the closing and reopening of the
lobes. The results varied not only in the flowers of different stocks,
but in diff'erent flowers of the same stock, even in the same flower in
different stages of development, in different hours of the day, and
also in different states of temperature and weather. Generally speak-
ing, the closing and reopening in a given flower are more rapid at
the middle of a clear warm day than at other times and in other
states of weather. Complete closing is performed usually in
3-6 seconds, but may sometimes take 7, even 10, seconds. Com-
plete reopening takes place usually after 7-12 minutes, but some-
times sooner, sometimes later. Some flowers which I examined on a
very warm day, reopened only after 5 minutes. I also found that in
young flowers, the closing is more rapid, while the reopening is much
slower, requiring about 13-15 minutes. But in mature flowers,
chasing takes place in the usual interval of time, while reopening is
quicker (7-10 minutes). In all cases the movement of closing may
easily be observed, but that of reopening is so gradual that we cannot
recognise it without careful observation. The experiments may be

208 M. MIYOSHI.

repeated several times in a given flower apparently without any
sign of decrease in irritability. The experiments may also be made
on the plants kept in the house with just as good results as on those
in their natural habitats. Of the flowers detached from the shoot,
the same holds good as long as they are prevented fi-om withering.

*The stigmatic lobes, when magnified are seen to be made up of
bundles of filaments (^Pl. XXIX, Fig. 10, 11, loos, tiss.) composed oi
cells full of granular protoplasm. The filaments are very loosely
acro-regated, passing below to the closer conducting tissue (cond. tiss.)
of the style. The inner sui-face (Tig. 10) of the lobe is quite naked
but studded with many papillœ (pap.) or the clavate apices (clav. ap.)
of the above-mentioned filaments, among which the pollen-grains (pol.
gr.) take lodgement. The outer surface (Fig. 11), on the contrary,
is loosely covered with a very thin layer of epidermis (directly con-
tinuous with that of the style), the cell-walls of which are more or
less cuticularized and marked with minute longitudinal wrinkles
(Fig. 11). Besides, there may be seen differences in the outlines of
the component cells of the epidermis, as we pass from the lobes of
the stigma (stig.) to the stylar portion (styl.) below — those of the
former being irregular and sinuate, while those of the latter are almost
rectangular.

As has been pointed out by Pfeffer. Sachs, and others, cells form-
ino" irritable parts of plants, when acted on by external stimulus,
allow water to pass out of their protoplasm, thereby suffering
diminution of volume ; and this contraction affecting the exten-
sive and elastic cell-walls makes the motion visible to the naked eye.
This, I believe, may also explain the irritability in the present case,

* The sfcruetares o£ the style and stigma have been studied by J. Behrens. (Untersuchungen
über den anatomischen Bau des Griffels und der Xarben. Göttiugen, 1875 )

XOTES OX THE IRRITABILITY OF THE STIGMA.

209

althoLio-li I am as yet unable to detect any decided structural
pecLiliarity.

The folli^wiiig observations were mnde to ascertain the signifi-
cance of the movement and to know in what relations, if any, it stands
with respect to the visits of insects.

April 16, 17. Rainy. I visited certain spots wliere the
plants were abundant. Many flowers were open. I saw no single
insect near, and the stigmatic lobes of almost all the flowers were
deflexed.

April 18. Clear warm day ; 22°C. at noon. At one o'clock
p. M. I went to the same places and found that many of the
flowers had their stigmas closed. Soon I saw two or three bees
come with a buzzing note. They alighted on some of the flowers,
thrusted their mouth-parts deep into the throat of the corolla
which had honey stored in the basal part of the lower lip. In so
doing the heads of the insects unavoidably struck against the open
lobes of the stigma wdiich at once closed. The heads were then
thrust in deeper and came in contact with the anthers. In a few
minutes thev visited no less than a liundred flowers and then
flew away.

At 3 p. M. On the same day I revisited the same places
and found a similar occurrence.

At 6 p. M. Comparatively small number of flowers (about
one-third) had their stigmas closed ; no insects were flying about.

At 9 p. M. Dark night. The flowers did not close, and the
stigmas were wide open.

April 19. Foggy morning. At 7 a. m. I saw the stigmatic
lobes cpiite re flexed.

At 9 A. M. A few insects were found entering the flowers.
April 20. Clear but very windy day. At noon I visited

210 M. MIYOSÖI.

the same place without noticing a single insect, and most stigmas
were o|)en.

During these days I likewise examined tlie same species in the
Botanic Garden of the University at Koishikawa. and found almost
the same state of things.

In all cases I observed that those growing in shady places and
those kept in the house had their stigmas always open, while those
on open sunny lawns had the parts mostly closed, — the differences
seeming to be due to the relative frequency or total absence of the
in sect -vi si tors.

These insect-visit(3rs belong almost exclusively to the Hymeno-
ptera, a species of Eucera (PI. XXVIII. Figs. 4, 5) of Apidœ,
identified for me by Dr. C. Isliikawa, being the chief visitor. The
visit of this bee, however, is not confined to the flower of Mazus, for
I often noticed that the insect burdened with yellow pollen dusts of
other flowers, probably of Taraxacum, thrust its body into the lips of
the flower smearing the stigma as well as the corolla with the golden
yellow powder.

So far as my observation extends, I may conclude that the
iri-italjility of the stigma of this plant is not for the pm-pose of
protection against wind and rain, of which the stigma may be
tolerably well kept out by the overhanging n[)per lip of the corolla,
but — as has been suggested by Hermann Müller* in the case of
Miinulm liiteus — for a more important purpose, i.e. ïov cross-fertiliza-
tion, which no doubt takes place in the following manner.

A bee laden with the [)olleu of one flower enters another flower
of the same species for honey, and thus comes with its liead in
contact Avith the lower lobe of the stigma which just overhangs the

* Die Befruchtung der Blumen druch Tnsecton und die ireoenseitig-eu Anpassungen beider.
Leipzig, 187:î.

NOTES OX THE IRRITABILITY OF THE STIGMA. 211

stamens. Soon after the contact (hy which the stigma receives the
pollen), the lower lobe folds up, opening' the way for the bee which
tlien enters deeper and becomes dusted with a new sii])ply of pollen.
That reopening of the lobe takes place in abcnit 10 minutes after
the closing seems to be well adapted to the reqnirement of the rase,
when we consider the intci-Nal of time which usually elapses before
the bee revisits the same ünwer. Tlie nsual deep bhiish purple, or
rarer snowy wliite, colour of the corolhi serves no doubt to attract
the insects, while tlie hairs on the floor of the lower lip seem to assist
the visiting insects in aHgliting.

In 3Ii)}nihts tirixilciisis, Benrh.. M. sessifolius ]\faxim.. and M.
moschatiis, Doug., all of wiiicli I liaNe observed, the mechanism is
precisely similar and adapted for the same purpose as Mazus, so
that it is hardly necessary to enter into details.

212

M. MIYOSHI.

Explanations of Plates XXVIII and XXIX.

Indications of Reference Letters.

il. I., upper lobe of the stigma ; /. /., lower lobe of the stigma ;
sty., style; stig., stigma; loos, tis., loose tissue of the stigma ;
cowl, tis., conducting tisssue of the style ; pap., papillœ ;
pol, gr., pollen-grains.

Plate XXVIII.

Fig. 1. Mazus rugosus, Lour., var. niacrantlnis, Fr. et Sav. (natural
size).

Fig. 2. Coi'olla cut open along the middle line of the central lobe
of the lower lip. showing 4 didynamous stamens (magni-
fied 3 times).

Fig. 3. Calyx cut open showing the pistil and its bilobed stigma
(magnified 3 times).

Fig. 4. Euccra sp, whicii visits the flower of Mazus. (magnified
1.5 times).

Fig. 5. Upper and lower wings of the same, showing the veins
(magnified 3 times).

Plate XXIX.

Fig. 6. Pollen-grains in different positions (magnified 540 times).

Fig. 7, 8, 9. Stigmatic lobes in the successive stages of closing
(magnified 22 times). Fig. 7, at the moment of a shock
given. Fig. 8, after 3 seconds. Fig. 9, after 5 seconds.

XOTE.S ÜX THE IKRITABILITY 0¥ THE STIGMA.

218

Fig. 10. I*i)i'li()ii of the inner .surfaccof Hie .stigmatie lobe showing-
i\iv papilltc or eiavate apices of the loose filameîits togethei-
with .some pollen-graius develojiing j)ollen-tii])es. The
Illimité globules in the cells represent the granular asj)ect
of the proto[)lasm (magnitii'd 540 times).

Fi(i. 11. Outer surliice of the stigmalic lobe togetlier with a portion
of the style; the epidermis is shown as broken along the
îniddle h'ne so as to sliow tlie loose tissue inside. The
cells of the e])idermis on the stigiiiatic portion are sinuous,
those on the stylar portion nearly rectangular (iiiagnitied
•!?){) times).

Jour. Sc, Coll. Vol, IV. PI. XXVIII.

Fig. 4

Fig. 5

Fig. 2

W

Jour. Sc. Coll. Vol. IV. PI. XXIX,

lOOis

Fig. 10

Fig. 11 7oo«. tu.

'^Wiiid tis

i1 i ^ty

Cond tis

i^'ig- 7

Fig. 8

u. l.

Fig. 6

Notes on the Development of the Suprarenal
Bodies in the Mouse.

By
Masamaro înaba, Rié^kushi,

With Plates XXX-XXXI.

It hns long lieon known that the suprarenal bodies of the ver-
tebrata consist of two substances, the medulla and the cortex. As to
Innv these two substances arise and in what relations they stand to
each other the o])inions of previous investigators are divided. During
the academic year, 1888-89, I studied the development of the organ in
the common dimiesticated mouse, a variety of ^lus musculus, and
came to the conclusion that the cortical cells are derived, as
Janosik stated, from the peritoneal epithelium, and the
medullary substance arises, as described by Mits uk uri,
from the sympathetic elements. The following is a brief
account of my investigation. I nmst here express my sincere thanks
to Profs. Mitsukuri and Tjima, f)r tlieir constant encouragement
and valuable suggestions, without which I could not have finished
the work.

As to the method of investigation I preserved after Selenka the
specimens, young and adult, in Kleinenberg's picro-sulphuric acid
mixed with chromic acid in the ratio 8 : 1. Some of tlie adult speci-
mens were also preserved in bichromate of potash, but as Gottschau
justly remarked, it is n<)t necessary to use the chromic acid, at least in

216

M. IX AB A.

tlie case of tlio monso, to demonstrate tlie distinction hotween tlie medul-
lîiry and (-(^rtical elements. In the preparations of the chromo-picro-
sidphnrie acid the mediilJa i« not coloured Ijrown ; this seems to he due
partly to the shortness of the interval during" which the emhryos
were exps^sed to the action of the reagent (1'/^ liours) and partly to the
presence of the [)icro-sul[)huric acid. To stain emhryos, I used a weak
solution of Kieinerdjerg's hematoxylin, as it gives the clearest and
most ditferentiated figures. With picrocarmine I also ohtained good
preparations of the suprarenal Ijodies of the young mouse. The
objects were stained in foto before imbedding in the celloidin paraffin.
In all cases I took pains to stain (h^eply and to cut sections as lliin
as possible.

I am not C(uite snre of the age of the embryo, since I could not
observe any actual c<^-]inlation. After the method of Selenka, I separat-
ed the individuals of two sexes for from ten to fifteen days, then put
a ]inir together f >r a night, and separated them again the next morn-
ing. I connt(M:l the day of separation as the first day of gestation, the
next the second day. and so fu'th. Frcnn a number of preserved
embryos I determined the approximate size (from the tip of the head
to the root, of the tail) of the embryo in each stage as follows :

nth day 3-4.5 mm.

12th day 4.5-6 mm.

13th day 6-8 mm.

14th day 8-10 mm.

15th day 10-12 mm.

In cases of embryos older than this stage, I opened their ab-
domen as (piickly as possible before immersing them into the killing
fluid, and could not make any relialile measurement.

Suprarenal ])odies of the Mouse, from the new- born
to the adult. — I commenced my study with the young mouse about

ox SUPRAKEXAL BODIES IX THE MOUSE, 217

one month old. In thtse specimens, the two substances of the
suprarenal bodies are already well marked. In cross sections, the
organ is elliptical, consisting of two concentric zones (PI. XXXI. tig.
21); the inner central zone (med.) stnins somewhat less than the outer
zone (cor.). Under a high power, the central zone is found to be
composed of irreü;ular cord-like cell-a^'o-reo-ates. each of which is

i O Co o

bounded by strong connective tissue fibres. The cell-protoplasm is
faintly stained ; the nuclei are large (6 U- on an average) and slightly
granular. The nuclei of the cells of the outer zone are smaller in size
(5 /z.) and highly granular. Their cells are smaller than those of
the central z(3ne ; this is especially the case in the middle portion of
the outer zone where the cell-protoplasm is stained deeper than in
any other part, so that the outer zone is subdivided into these minor
concentric zones. But these three zones gradually merge one
into another without presenting any distinct limit. The transition
from the <3uter (cor.) to the central zone (med.), on the other hand, is
very sudden ; the limiting line is distinct and tolerably even, forming
an elli[)tical outline. Evidently the central zone is the medulla, and
the outer the cortex.

Turning now to the mouse ten days old (PI. XXXI. fig, IS), a
considerable difierence is observed in the structure of the medulla.
The medullary substance (med.) projects irregularly into the cortex
(cor.), and the boundary is not yet even, though its elliptictd outline
can already be made out. The cells and nuclei of the medulla are
stained deeper than befjre, so that the distinction of it from the c(.)rtex
is obscure in some parts where the farmer projects int(j the latter. The
difficulty is further increased by the fact that the cord-like arrangement
of the medulla is as yet very weakly developed, and the respective
sizes of the nuclei in the two substances are approximately equal.
Put tracing carefidly the nnirgin of the medulla, we can tind here and

218 M. IXABA.

there the distinct groupiiigs of its cells into cords (tig. 19), where
the nuclei are larger and the protoplasm is less stained, than in the
adj(jining cortical cells. This stage seems to be the formation of the
medullary cords. The three minor zones of the cortex are already to
he found, thouoh less distinct than in the staii'e described before.

In the mouse three da^^s old (hg. 1(5). the medulla is very
irregular in its outline. Alon£>' its maro-in the cells are "'reatlv
minüded with the cortical cells, luit the distinction is cleai-, the cells
and nuclei of the medulla being stained more deeply and packed more
closely, than in the cortex. The three minor cortical zones are not
yet distinguishable.

In the newly-born nKjuse (wood-cut 1 and PI. XXXI. hg. 15),
the medulla no hunger forms any compact mass, but has cortical cells,
intermixed throughout its substance. The distinctions between the
two substances can however be easily made out as before.

The relative size of the nuclei in the two substances is interesting.
In hgs. 15 and 16 (PI. XXXL), the nuclei of the medullary cells are
evidently smaller than those of the cortical cells, while in tig. 21, the
case is reversed. I measured the nuclei of cells in the two sub-
stances near their boundary line at various stages. The following
gives the average size (in u) of those nuclei.

1 day old. 3 days. 10 days. 29 days. adult.

Medulla 5.2— 5.6— :^.Q+ (5— 6 —

Cortex G.5 — 6. — 5.4— 5 — 5 -f-

It will be seen from the tal)le that for about a month after birth, the
cortical nuclei are gradually decreasing in size ; at the same time the
medullary nuclei are growing though very slightly. This is, I believe,
due to the formation of the cord-like arrangement on the part of the
medulla, and of the z<_)iia reticulata <_)n the part of the cortex.

ÜX SÜPRAKENA BODIES IX THE MOUSE.

219

Woodcut 1.

From a mouse one day old. The left .suprarenal body is represented. Ao=Aorta,
B v. = Veins, Cor.=cortex, Med. = Medulla, Mes.=Mesentery, s. s. = Maiu mass of
Sympathetic ijanglia, sy. <j:.— gang-lion of sympathetic origin. 2xBB.

So far US trart'd, the niediilla i.s always di.stiiK-t fVum the cortex,
and its origin cannot be decided. lUit some interesting (and evident-
ly a little ahnornial) cases were met with. In one mouse jast born
(woodcut 1), the roughly eliiptictil medulhi (med.) .situated in the
centre of the organ sends otf an (jifslxjot at one place toward the
medial side, actually reaching the C(jmiective tissue capsule. Outride
the organ hes a large ganglion (sy. g.), which is found on tracing-
sections to Ije continuous with the main sympallietic system (s. s.).
The medullary cells of the suprarenal body and the true ganglion
cells are very similar in their size and colouration. This condition
was observed only on the left suprarenal.

In a three-day (jid mouse (woodcut 2), again on the left side, I
obser\'ed an actual ctjnnection of the medidhi with the gangli(jn. In
tig. 17 (PI. XXXI.), which is a more magnified tigure of the wootl-
cut 2 I), a mass (jf cells with small and deeply stained nuclei (med.)
eads out of the organ, and directly joins the ganglion cell mass (sy. g.)

22Ö

M. IX ABA.

lyiii«^ close to it. In another mouse at the same stage, a simihir condi-
tion Avas observed; the gan<^lion besides Ijeing joined Ijy a nerve com-
ing from tlie neighbourhood of the kidney.

Woodcut 2.

■i snccessiTL- sections (not cunsccutivu) fiom the posteriui' cud of the left suprarenal,
a :3-<lay old mouse. Horizontally sliadrd part— Cortex. Vertically shaded part-
Medulla.

riuided by these facts, I examined again tlie ten-da^^s old mouse,
and found in one case the medulla projecting on its medial side and
actually t(3U(-hing the connective tissue capsule (PL XXX f. fig. 1^0.)
but it Avas not traced to the sympathetic ganglion. These facts
plainly show that the medulla is derived from the ganglion colls.
When and how the ners'ous elements enter the organ, will be describ-
ed below.

ON SUPRA-REN'AL BODIES IN THE MOUSE. 221

In passing, it may be remarked tliat in wciodcut 2, a small
portion of the rortical substance is projecting far posteriorly and is
separated from the main mass h\ the sympathetic g'angli(Mi. In
fig. 17, the part (ac. ror.) is distinctly separated from tlie main mass
hy strong connective tissne cells. This is the so-called accessory
suprarenal body. From the mode of the entrance of the nerve into
the organ, as seen in this and other rases, I am inclined to be-
lieve that the introduction of the nervous elements into the organ
greatly influences the formation of the accessory suprarenal body,
though it may not lie the sole canse.

Uf the adult suprarenal body (PL XXX. hg. 22). I have little to
say, as it does not differ much from that of the ono month old monse
(hg. "21). One feature interesting from the embryologienl point of
view is the occasional ]jresence of the ganglionie remnants.
In one specimen (hg. 23), T f>und at the murgin of the medidla on
its medial side, a mass (sv. g.) of indistinct cells, highly granular and
deejdv stained. Their nuclei are smaller than those of the mc(bdla
or cortex cells but decidedly larger ihan those of the connective tissue
cells. P)V tracing sections, [ fonnd the mass t(^ project pyramidally
into the cortex and finally reach the capsule. In comparis(^n with
the ten-days old suprarenal l)ody (tig. 2»») this mass may be considered
as a part of the nei-vons elements, wliich has not been transformed into
the true medulla. Of the laro-e ganulion cells such as seen outside
the adult suprarenal body. \ could find none pressent within the
adnlt organ.

I)(Melopment of the HeduUary Substan<'c, in the
1 3 t h - 1 8 t h day E m b r v o s . — lîalf )ur ' remarked i n his mono -
graph on elasmobranch fishes that the suprarenal bodies of

1. Olilei* erature I had not access to.

■ 222 M. IN ABA.

Vertel)rate,s consist of two substances distinct in tlieir ovig-in. This
Braun " hns confirmed in Reptiles, and Mitsiikiiri ' in Mammalia.
Mitsukuri says that in the 16th day embryo ral)bit the medullary
substance is already distinct ; sympathetic nerve cells closely applied
to the inner side of the suprarenal blastema send in a process partly
composed of nerve fibres into the ventral end of the suprarenal ; the
cells thus carried in become gradually transformed into the medulla,
(iottschau* and Janosik'' dispute this statement. TlnMigh these
authors do not deny the entrance of the nerve fibres into the
suprarenal, they state that the two jiarts of the suprarenal substance
cannot be distinguished at the time of the entrance, and the medullary
substance is gradually ditterentiated from the cortical substance at a
considerably later stage, Gottschau even states that in some mam-
mals the medulla is developed only after birth. Yet fivMn the descrip-
tions of the two authors, the exact nnxle of the formation of the
medulla is not yet clear, and it is also necessary to trace the ultimate
fate of the nervous fibres sent into the suprarenal Idastema.

The suprarenal blastema is already distinct in the 13th day
embryo. It is a somewhat elongated mass of cells lying between
the 16th and 17th l)ody-segnients, just behind the lobes of the
lun^'s. The anterior end of the l)lastema lies on about the same
level as the 2nd tubule of the mesonephros, while the 3i'd segmental
tul)ule lies on ab(out tl>e middle portion of the suprarenal. In
cross sections ( wooden r 3 and V\. XXX. fig-. 8), the blastema

2. F>au und Entwicklung der Nebennieren bei Reptilien. Arb. aus dem Zoo]. Zoot. Inst,
in Wurzburg. Bd. V. 1882.

;î. On the Development of the Suprarenal Bodies in Mammalia. Quart. Journ. of Mic-
roscop. Science, XXII. 1882.

4. Structur uad embryonale Entwiclclung der Xe))ennieren bei Silugethieren. Arch. f.
Anat. u. Physiol. 18S3.

5. Bemerkungen libn- die Entwicklung der Xe})enniere. Arch. f. Mikr. Anat. XXII.
1883.

ox SUPRARENAL BODIES TN THE MOUSE

223

(s. r.) is seen ns n rounded mass (al)out ^ mm. tliiek) of cells lying
between the aorfa(Ao.) and the mesonepliros (st.), immediately below
the cardinal veins (v. car.). Already at this staL>,"e, a blood vessel
(e. V.) is seen in the posterior portion of the blastema, coming from
the cardinal vein ; this vein is ultimately transformed into the central
vein of the adult suprarenal. The suprarenal blastema (s. r.) is
distinguished from all neighl)ouring tissue cells by the densely packed
state of its large and faintly granular cells. Cell boundaries within
the blastema are only faintly indicnted, but a careful observation
sliows that relis are collected into irregular groups, separated by
scanty connective tissue cells. The cell nuclei are slightly granular
and their size varies between 5-7 u. These characters of the cortical
cells are retained during the subsecpient developmental phases and aie
useful in distinguishing them from the medullary cells.

Woodcut 3.

UigJit

A cross section taken near the posterior end of the suprarenal bodies. — 13th day
embryo. Ao.=aorta, c. v.=eentral vein of the suprarenal, Gr. 0.=generatiye organ,
s. r. = suprarenal blastema, sy. g. ^sympathetic ganglia, v. c =vena cava, v. car ;=
cat'dinal vein. 2xaa.

The sympathetic ganglia (woodcut 3 sy. g.) are well developed
on the upper lateral corner of the aorta, ami a strong branch from

224

M. TNABA.

the spinal nerve enters each ganglion. The ganglia send out branches
downwards between the aorta and the cardinal vein, but they are
ver^' tine, often consisting of a single row of cells and cannot be
clearly traced. Yet on the medial side of the suprarenal blnsteinn,
closely applied to it, there is seen a small irregular grou]) of deeply
stained cells (fig. 8, sy'. g'.), whose nuclei are a little smaller and
more granular than those of the suprarenal, and similar to the cells
of the sympathetic ganglia. Probably these cells are of the nervous
nature.

Woodcut 4.

Migli <- —

A cross section taken near the posterior end of the suprarenal 1)odi»js.— Later stage
of the 13th day. Ao.— aorta, Bv.=veins, G. O.^geuerative organ, s. r.=suprarenal
blastema, s. t.=segmental tubiilus, sy. g. = symi)athetie ganglia, v. o.-- vena cava, v.
car.— cardinal veins, "W. D.=WolfRan dnct. 2xaa.

Towards the close of the 13th day (woodout 4), the cardinal
veins greatly retrograde, on the right side almost completely. Thus
the central vein of the right suprarenal becomes now the direct
continuation of the vena cava, and the left central vein becomes a
side branch from the great vein. The suprarenal blastemas of the
tw^o sides are now placed not ventrally, but laterally to the aorta.
The mesonephros is pushed laterally and Miiller's duct is distinct. In

ON SUPllAREXAL BODIES IX THE MOUSE. 225

the 14th day embryo, the blastemas have a considerable size, a little
projecting into the coelom cavity. The kidneys appear at the post-
erior and dorsal side of the suprarenaL By dissecting the embryo, the
suprarenals are seen as a pair of oval shaped bodies, flattened antero-
posteriorly as if pressed by tlie developing kidney. The inner end of
each su])rarenal is attenuated and thus overlaps the anterior inner
corner of each kidney, — a state of things retained and more distinctly
seen in later stages. In the 15th day embryo (woodcut 8), the
suprarenal bodies liave shifted their ])osition, farther dorsalward,
being now placed just laterally t(j the ^'ertebral body and dorsally to
the aorta. Thus at no stage, are the suprarenals of the two sides con-
nected together as some writers state. As Mitsukuri and Gottschau
well remarked, it is the ganglion placed inside of each suprarenal,
which is posteriorly joined to its fellow by a cross bar.

The nerves sent out from the sympathetic ganglia are distinct in
the later stage of the loth day (woodcut 4). Two or three branches
are successively given out from the ganglia and all are united into
the splanchnic plexus lying inside of, and closely applied to, each
suprarenal. A Ijranch is further sent downwards from the plexus to
the front <jf the aorta, where it is connected (in the next day) with
its fellow of the other side. From the 14th day onward (woodcut
5), we can distinguish in each splanchnic plexus at least two
ganglia, the larger anterior and the smaller p(3sterior ones. The
posterior <j;ano^lion on the ri^'ht side is elono-ated and becomes con-
tinuous with the cieliac ganglion, no that the latter may be said to be
the direct continuation of the right splanchnic plexus. From the
ganglion closely applied to each suprarenal (that is the second
ganglion of the plexus), some fibres enter th(3 organ. Though very
fine, these fibres can be traced for a certain distance within the organ.
Woodcut () and tig. 9, taken from the 14t,h day embryo, repre-

226

M. IXABÀ.

sent the state of things, when the nervous elements are just entérina-
the organ. It is seen only for one section.

Woodcut 5.

^'2')lanchnic GinujUa VenaCara Aorta Siilanchnic Gaii/jlia Left Siqirarriiitl

Eight Su2:>rctreiiai

Central Vein

J?. Siiprai'. Artery
Ren. Art.

Ren. Vein

L. Siqn-ar. Artery

Central Vein

Ren. Art.

Ren. Vein

Ureter Vena Cava Aorta Ureter

Semi-diagramatio figure, showing relations of suprarenals to ganglia and bloodvessels.

Woodcut 6.

From a 14th day embryo, representing the right suprarenal. The place marked x is
more magnified iufig. 9. (PI. XXX ). Ao.=aorta, Bv. —veins, G. 0.=generative organ,
s. r. = suprarenal blastema, s. s. — main mass of sympathetic ganglia, sy. g. = ganglion
of the sympathetic origin.

In I he loth dav enibrvo, the nerve fibres within the oriran are

stronii'er and more easilv to be ascertained. Tiiest; branches are

ox SUPHAKEXAL IJüDIEö IN THE .MOUSE.

1^27

tolerably constant in iiuin1)er. tJcucrally into tlie left snprarena]
(woodcut 8), one ^'el•y .strijng Ijumlle enters at about the middle and
ventral portion of its inner margin. At the corresponding point of
the right suprarenal (woodcuts 7 and 9 A) a strong bundle (but
nnjre slender than that of the left side) is seen ; on the same level and
S(Hnewhat dorsal to the one just mentioned anotlier smaller Ijundle runs
in from the same ganglion. Besides these, a small bundle may S(jme-
times be seen entering the organ at its p(jsterior end (woodcut 9 B).
All these bundles are very delicate, and can be seen only for three (jr
four consecutive sections.

It will be necessary here to describe the characters of the ner\ous
cells to distinguish them from the cortical cells. The protoplasm
in these cells is not so rich as in the cortical cells, and is very
granular ; their nuclei are comparatively small (4.5 u on an average),
thickly packed, and deeply stained due to the presence of many
granules.

Woodcut 7.

•5'y. 9.

A cross section takeu from a loth daj embryo, right suprareual. S. r. = suprarenal
blastema, Sy. g-. = <>-aug-lion of sympathetic origin. 2 X B.

The place marked ^ is uiDre uiagnified iu fig. 10. A. (PI. t.) The jilace marked * is
more magnified iu fig. 10. B.

In tig. 10 A (which rei)resents a portion of tlie woodcut 7
under a higher ))ower) taken frc^m a 15th day embryo, a mass of

228

M. IN ABA.

nervou« cells is seen insinuutini^' itself intcj the cortex. The other
smaller bundle (marked in the woodcut witli a *) is interesting. It
is very delicate and scarcely visible, running deeply into the cortex,
and iinally ending in a small cluster of cells, which are distinctly of
nervous nature (Pi. XXX. tig. 10 B).

Woodcut 8.

A cross section of a IGth day embryo, left side. Ao.= aorta, cor. = cortical substance,
med = medullary substance, Sy. g. = ganglion of symijathetic origin. 2xBB.

In the 16th day embryo, the nervous elements carried in the
organ are consideral)le (woodcuts 8 and 9). They form now a
reticulated network imbedded between the cortical cells, appearing in
sections as small scattered groups of cells. Though the main mass of
the nerve cells is clustered in the centre, some cell groups (PI. XXXI.
tig. 11) are fjund in the peripliery of the organ at its medial side
and send out their tibres, which actually piercing tlirough the con-
nective capsule become continuous with the ganglion near the organ.
In others (tig. 12), altlnjugh the tibres pierce through the capsule.

ON SUPRAKEXAL BODIES IN THE MOUSE.

229

they can not be traced to tlie ganglion, but are lost on the way ; iti
otliers again, they are lost in the connective tissue capsule of the organ.
rro7n this stage onward we can call the nervons elements within tlie
suprai-enal more appr(^i)riately the medulla. I l)elieve this and tlie
previ(Mis stage are snfticient to show the nature of the medullary
substance. Probal)ly these two stages were not observed by
(rottscliau and Janosik. who thus concluded that the medulla is
differentiated gradually from the cortical substance.

Woodcut 9.

Cross soction takon from a Itltli day embryo, rif^lit side suprarenal,

A, at the mid<llo of the or^-au. P.. near the posterior end.

c. v. = central vein. oor. = cortex, med. = medulla, sy. "•. = o-ano'lion of sympathetic

origin. 2xBB.

As to the further growth of the medulla, I have little to describe.
It consists merely in the increase oi' the medullary cells which
gradually form a compact mass in the centre of the organ, the cortical
substance becoming in consequence scanty in the centre and pushed
to the; ])eriphery. (PI. XXXI. 14)

The severance of the nervous connection commenced in the
previous stage is usually complete in the 18th day embryo, which

230 M. IXABA.

wns t'iie oldest one I investigated. The process takes place simply by
the growth of the connective tissue capsule around the piercing rierve
whicli is consequently reduced to a narrow neck and hnady cut ott'
(fig. 13). Still the direct connection of the niedulhi witli the
syni{>athetic ganglion is retained in some cases, especially on tlic left
side, fn all such cases observed, the connective link which persists
is enorinc^uslv strong, so much so that sometimes the ganglion itself
may be iimnersed in the organ. This is one reason why the
connection j)crsists longer. Fiu'ther ns before stated, on the left side
the nervous fibres entca- the organ mostly as a single conspicuous
bundle, whih^ on the right side they are usually divided into several
smaller clusters, which will more easily be cut off. Hence the
connection when it persists in tlie newly born mouse is always found
on the left suprarenal as before described.

As to the general aj)pearance of the Idstological elements of the
suprarenal bodies in this stage, it does not much differ from those of
tlic newly born animal.

Development of the Cortical Substance in the 11th-
1 :^ th day Embryos. — As regards the (n-igin of the cortical sul)stance
the attention of earlier writers has been principally directed to the in-
ditferent mesoblast. Kiilliker''' stated that the siipi-arenal bodies in the
rabbit first ap|)ear in the 12tli or 13th day embryo as masses of some-
what large round cells on each side of, and ventral to the aorta, on the
inner side of the Wolffian bodies and doi-sal to the mesentery. Mi-
tsukuri continued this and added that dors;dlv tliis mass is tolerably
distinct from the <)ther mesoblastic cells, but \entrallv its termination
is indetiiiite. Ihaimi', 13raun, and more i-ecentlv Gottschau derived

f). Entwieklmio-sovscliiehte des Menschen uml der li('>lieren Thieve. 1S79.
7. Eiu Beitrage 7.nr Kentniss des feineren Baues und der Entwiclduno'sgeschichte der
Nebennieren. Arch. f. Mikros. Auat. VI IT. 1S72.

ON' SUPIiAREXAL P.ODIES I\ THE MOUSE. 231

the cortical cells from the mesolilast, l)nt in connection with the walls
of the blood vessels (atn-ta, cardinal veins, vena cava, or vena renalis).
Recently for the hrst time Janosik stated that the suprarenal
body takes its origin from the peritoneal epithelinm, and it is in fact
in tlu' ch^sest connection witli tiie beo-inniiiL»- of the sexual oro-an :
this connection pt-'rsists i'nv a tolerably long time nntil it is cnt otF
by the entrance of b]o(-)d vessels, especiall}' the vena vertebralis
|)osteriori and other veins emptying into the same from the Wolffian
liodies. Weldon,"" on the other hand, derived the blastema from the
Wolffian 1)odies. According to his statement, a cell-mass proliferates
from tlie walls of the glonieriilus and separates into two masses : the
one travellirjg liackwards becomes the suprarenal Ijody, the other
growing downwards and entering the sexual organ becomes the
tubuli semirjiferi (in the male). Mihalcovics" also affirmed like Weldon
the connecti(Mi of the suprarenal Ijlastema with the sexual " strano* "
(=:segmental " sträng" of Jîraun), wdiidi he dci-ives, however,
from the germinal epithelium. At this point he agrees with JaiK^sik,
but ditfers in the statement that the suprarenal b(xly is only tlie
undiiferentiated anterior continuation of the sexual or(^an. In front
of the anterior end of the generative ridge the suprarenal cells are said
to be directly proliferated from the peritoneal epithelium, and
posteriorly they are said to he continuous with the sexual sträng but
not in direct connecti(^n with tlie peritoneal e[>ithelium. In Ijirds
and mammals, the direct proliferation of the peritoneal epithelium to
firm the suprarenal blastema is said to be confined to a very small
tract, so that it might be overlooked if series of sections were not
studied.

8. Suprarenal Bodies of Vertebrates. Quart. Jour, of Micros. Scieneo XXV. 1885.

9. Untersuchuuo-en über die Entwicklung des Harn- und CJe.selileehts-apparates der
-imnioten. Inter. Monatschr. f, Anat. u. Hist. II. 1885.

-•">- M. IXAKA.

To trnoo the on'o-in of the cortical snhstanco is in far-t extroniel}^
difficult, as its cells are faintly distinguished from the other tissue
cells. The cortical hlastema in the mouse is tolerably v:e]] seen in
the early stage of the 12th day of gestation. The mesonephros in
the monse is very weakly developed. Only the aiiterior two or three
segmental tuhiiles actna]l3M-)pen to the Wolffian duct ; iolh^.wing these
ran be traced five or six blind tubules, vhicli lessen in size one after
another, until finally no tul)ular structure is seen beyond tlie 8th or
9th one, the cells being merely clustered in ])r(^per ])laces. The
supi'arenal blastema extends from about the middle of the anterior
two segmental tuljules to about tlie 6th or 7tli tubule. In ci'oss
section, it is large anteriorly and gradually lessens in size posteriorlv.
It is placed just at the angle of the mesentery (PI. XXX. figs. .") and n\
occupying the space enclosed by the aorta and the cardinal vein on
the medial and dorsal side, and liy the mesonephrc^s and the genera-
tive organ on the lateral size. ^Fediallv the [blastema is distinctly
liounded by connective tissue cells. Where the S-shaped segmental
tul)ules are projected in medial direction, they approach the dorsal end
of the suprarenal blastema ; in other cases they are far removed from
the suprarenal. In no cases do the tubules send out cells medially.
The walls of the cardinal vein show no signs of proliferation.
]>ranches of the vein to the suprarenal are not yet developed.

The relation of the suprarenal blastema with the beginning of
the generative organ is interesting. These two blastemas are ]il;iced
side bv side, their anierioi- extremities reaching about the same level,
but posteriorly the generative blastema extends far beyond the
end of the suprarenal. The cell elements of the two are very similar,
consisting of large cells with large round nuclei, which are stained
slightlv deeper than those of the connective tissue cells. Ihit the
two blastemas are separated from each other in all places, except at

ox SUPRAKENAL BODIES IX THE MOUSE. 233

the anterior parts, bv an intervening tliiu septum of connective tissue
cells. Tliis septum, consisting of the two or three rows of cells,
runs from the peritoneal epitheHum in dorsal direction, and finally
separates itself into two branches, tlie one bending laterally and
covering the generati\e (jrgan, the other bending medially and cover-
ing the dorsal end of the suprarenal.

The cells of tlie peritoneal epithelium which touches the supra-
renal blastema are arranged in a single row (fig. 6). But as we
proceed anteriorly (fig. 5) the epithelium cells are evidently
proliferating 5 they are actually pushed upwards and are even con-
tinuous with the suprarenal blastema. Tracing sections still anterior-
ly, the connection becomes more intimate, till near the anterior end
of the suprarenal (TL XXX. fig. 4) the |)eritonea] epithelium cannot be
distinguished from the suprarenal blastema itself. Here the septum
no longer exists between the suprarenal and generative organs. The
cells of the two blastema are laterally continuous with each other, the
two being indicated only Ijy the two rounded eminences projected
dorsal ward; ventrally they are both seen to be the proliferation of
the peritoneal epithelium.

In a stage somewhat earlier than that above described, the sup-
rarenal blastema is not yet so distinct. Figs. 1-3 were taken from an
endjiyo in the later stage of the 11th day of gestation. Fig. 2 taken
frcjin near the anterior end of the left suprarenal blastema corresponds
with fig. 4, and figs. 1 and 3 taken on both sides at the middle
ot the organs corres[)ond with fig. 5. From the somewhat
detailed description of the previous stage, any further remarks will
not be needed. Only it may be added that the proliferating cells are
xevy indistinctly bounded dorsally, but a carefid study shows that
they are proliferated from the e[)itheliuni. Why I do not consider
these ])roliferatinL!; cells as the sole bei^'inniiiL;- of the u'enerative oru'an

234 M. IX AB A.

is simply that tlie position <jf (hat oro^an is always in the following
stages a little removed fr<jni the angle of the mesentery. Further in
figs. 1 and 3 the proliferation of the peritoneal epithelium can be
roughly se[)arated into two parts, the medial and lateral.

From the above description, I think that Janosik's statement as
to the origin of the cortical cells is quite correct. My figure 1 corres-
ponds with his figure 1. The only ditference is that the mesonephros
in the mouse is not so well developed as in the case of the pig.
Thus Junosik stated that the cells proliferate in the medial direction
to the aorta, ^vhich condition is observed in the mouse only on the
right side. The mesentery in the mouse being shifted from the
medial line a little to the right side, its angle on the left side is
carried far to the medial line, so that on this side the suprarenal
blastema is projected upwards and a little lateralwards in the direction
of the mesonephros (compare figs. 1 and 3). I camiot determine
whether the suprarenal body is really the anterior continuation «jf the
fj-enerative ridge or not. The state of things as seen in the tigure
given by Mihalkovics from a sheep embryo (his fig. 1(37) I could not
find at the corresponding point of the mouse. Dut from the fact that
the peritoneum is proliferated and tlie su[)rarenal blastema is placed
side by side with the generative organ in its entire length, it is more
likely to be the lateral separation, and not the anterior continuation
of the generative organ.

Further growth of the suprarenal blastema consists simply in its
separation from the peritoneum and clustering into a more compact
round mass, as will be seen in fig. 12. The pr<diferation of the
peritoneum, though slight, is still observed towards the close of the
12th day. Beyond the anterior end of the suprarenal bodies, a slight
proliferation of the peritoneum w.is sometimes observed (Fl. XXX.
fi^r. 7). I think that the com[)act suprarenal blastema is formed

ox SUPRARENAL BODIE« IX THE MOÜSIv 1^35

rom the main mass of the proliferated cell«, while a small portion
may be left behind, which .seems linally to disappear without eDterin<i-
into the formation of the suprarenal bodies.

T o b u m u p :

1 . T h e M e d u 1 1 a a n d t h e cortex are distinct in their
o r i g i n .

2. T he CO r t i c al b 1 a s t e m a a p p e a r s in t li e later stage o f
the 11th day of gestation, as a proliferation of the perito-
n e u m at t h e angle o f t h e m e s e n t e r y a n d laterally continu-
ous with the beginning of tlie generative organ. The
separation from this connection is complete on the 13th
day.

o. The medulla is derived from the sympathetic
t! 1 e m e n t s , w h i c li enter the o r g a n in t h e 1 4 1 h d a y e m Ij r y o .
They i n crease a n d f o r m a reticulate d mass at the n entre,
from which the cortical cells are gradually pushed aside.
The connection with the sy nipathe t ic sy s tem is usually
cut toward the close of gestation, but in some may be
re t a i n e d until a f t e r birth.

2S6 M. IX AB A.

Explanation of Figures.

tic. cor. = accessory suprarenal. A()= Aorta. Art. c.=cœliac artery.
Bv.=N'ciiis. cor.=cortical cells, c. Y.=central vein of Suprarenal
l)0(]ics. l)iai;'.=Dia[)liragni. <i. o.=Generative organ. Kid.=
kichu'v. Med.=Me(lulhu"y cells. Mes.=Mesentery. S. r.=Supra-
renal body. IS. t.= Segmental tuljules. Sy. f.=Synipatliedc nerve
fibres. Sy. g.=Synip;ithetic ganglion cells, v. car.=cardinal veins.
V. c.= \ ena ca\ a. \V. D. = Wolffian duct.

Fig. 1. From the 11th day embryo. Right side. Taken from

the level of the ind segmental tubule. 2 x E.

Fig. rJ. From the 11th day endjryo. Left side. Taken from

near the 1st segmental tubule. 2 x E.

Fig. 3. From the 11th day embryo. Leftside. Near the 2nd

seirmental tubule. 2xE.

CT

Fig. i. From the 12th day embryo, early stage. Left side.
Near the anterior ends of the suprarenal and generative organs.

2xE.

Fig. Ô. From the 12th day embryo, early stage. 10 sections
behind. 2 x E.

Fig. (J. From the 12th day embryo, early stage. About the
level of anterior one third of the left suprarenal. 2xE.

Fig. 7. From the 12th day embryo, late stage. Left side.
l>eyond the anterior end of the suprarenal bodies 2xF,

Fig. (S'. From the loth day embr^^o, early stage. Right side.

2xE.

Fig. L>. From the 1-lth day embryo. Right side. Tlie [)lace
marked x in woodcut G. 2xF.

ON SUPHiVTIEXAL P.ODIES IX THE MOUSE.

287

Fig. 10. From the loth day embryo, llioht side. A. tlie
place marked ^ in the woodcuf 7. 2 x E. P». tlie pl.ice marked ^^

2xF.

i'Vy/. ii. From the IGtli (hiy embryo. Left side. More
maonitied tii^nre of woodcut 8. oxJ)D.

t'ig. 12. From the Killi (biy embryo. Rio-ht side. M..re
magnified fionre of woodcut Î) A. 3 x J^D.

I'ig. IP). From tlie 18th day em1)i-yo. Fi-om the posterior
part of the left .suprarenal. 2xE.

Fig. 11. From tlie IStli day embryo. Fi'om another em1)rvo.
Central portion oi a .section, taken near tlie j^ostci-ior end of the
right suprarenal. ox])D.

Fig. IT). From the 1 day old mouse. Right su])rarenal.

3xD.

Fig. IC). From the o days old mon.se. 2xF].

Fig. 17. From the o day.s oM mouse. Another speeimen.
Rosterior end of the lefi suprarenal. Move magnified fioufe of
woodcut 2. R).

Fig. IS. From a mouse about 10 days old. 2 x E.

Fig. 19. From a mouse about 10 days old. Medulla is

weakly developed. 2 x F.

Fig. 20. From a mouse aliout 10 days old, another specimen.
The remnant of the connection with the sympathetic.

2 X E.

Fig. 21. From a mouse about 1 month old. 2 x E.

Fig. 22. From an old wild mouse. 2xE.

Fig. 2S. A p;u't of the right su[)rareiial from an old mouse.

2xE.

Jour. Sc. Coll. Vol. IV. PI. XXX.

/■<</ ')

/■Vi;. ,S,

tCZ). V

/'V9. >î.

l''i-i). 6.

M. Inaba del.

Jour. Sc. Coll. Vol. IV. PI. XXXI.

' e ^,

Fig. n.

Feg-. 14-.

î>o

Med ■

^-

_A\

Sy.f.

«M G

..^>® "o!^'

■ g*0^

Fig /V

?^^^o%.

^■6

St<:,(

. •• St •

:::)«»0

>>"

V • • • ••

•

/

• »

Fig- 13.

S«./.

•••

••V*

.M^((-

F,g. 19.

>• •

Med. ,

../

^t:« >.^

:?:-5r.i;-.>:

••»•• **

■'^st ..

■*^..

Cor.*«

'Ä

f i? 2 0.

i^^^'v •

Cr" o Ô

/■V 21.

: V.

o" »Sr

■M. I.ntba del

EKRATA.

Page 245, line 10, for Prof. Schenk read Schimper.
„ 246, „ 20, „ I). Kochihe read B. Kochihei.
„ 300, ,, 5, from end, for " nickel " read " bisrüuth."

On some Fossil Plants from the Coal-bearing
Series of Nagato.

By

Matajiro Yokoyama.

With Plates XXXri— XXXIV.

Fii tlic :'])i-iii£i' of 1S90, aIc. l\i,('liilM> oi' tlie (k'oloo-icnl .airvev
(lir'ooYcred soihc ])]:iiit i'ciiiüiiis in the cw.'il-lK'nn'n^' icrie.s of Xm^'uIo :it
•A ])]:!(•(' f:i]I('(l V:iin:!iioi, some .']() kiloineters en;'t of the city (jf
Ak:un;iLi-:is(-ki/' Tliese ])]:nits lie ren^o-iiize«! :is ^[c.'ozoic, and sulj^e-
(pieiitlv sent tlieni to nie ior exruninatioii. ()n lookiii^'at tliese ])lant>!,
I was at (^nce struek 1)v the oetanTenee of form.: whieli are (jnite foreio-n
to onr ^riddle dnrassie flora, lately worke(l ont"-' 1)\' myself, and wliidi
arc liirlua'to knoAvn oid\' a,' ocrni'rinL;" in tlie IHuptie. Tiiterested in this
disco^■ery, I visited the locality myself in tlie ninmci' ()f tlie same year,
in ordi'r to ohtain. ii" ])ossil).lc, :i laiycr mnnhcr of species, which, as T
thonu'ht, Avould l)e (jnite indispensahle for the determination of tlieir
exact a_u"e. Soi'ry to ::av, ho\ve\ei', T did ni)t succeed in makino- :inv
LjTeat additions to tlie iiiimher i-f s])ecies, nearly all tlie jilants whicli I
found ha\i:iLi- Ijeen already represented in tlie Ci,IIection of ^Ii-.
i\ochihe. Still my collection ])ro\'ed to l)e \ery useful, f )r I had tlius
a laru'er numhia* of in(h'\ idual.- for comparison.

The coal-heariiii^' série.; of Xag'ato occupie,; a limited area in tlie
.southern ])ortion of that ])rovince bordering' the Iidaiid sea, and con-

1) 'I'liis city is better Icnowu nuilpr the old uauie of SluiiioiKischi.

■1) Yokoyama, Jurassic Phmts from Katja, ITidn, mul Kchi~i')i, -Inininl of the College of Science.
Imperiid University, Jajxin, vol. Ill, part I, ISS'.l.

240 M. YOKOYAMA.

.si;st.^ of :i thif'k oom])lox of sandstone-!, cltiy-sl'itcs iind shnles, witli
subordinate layers of sclmlstciii and autlirncite in it.s lower ])art :ind
of brownco:d in its u])])er p'U't. These strata wliieli form a low
hilly country surrounded 1)\' niount-nns of u'ranite and of l^alaeozoic
formation strike generalh' from east to west, and show steeper dips in
the northern than in the southern part of the district, where they
gently .slope towards the sea. Owing to the repeated foldings to
whicli these strata luive heen subjected, their geoL^gicid structure is
complicated, and hns not yet been clearly mnde out. It will be
onlv added here th:it our fossils were discovered in the lower or
^•cha]stein-l)enriijg ])ai't of this formati(^/n.

The fossil locality lies on one side of a rond wliich lends from the
village of Yamanoi to the town of Habu, in a valley surrounded l)y
liills. Here in a space of al)out 4 meters, I observed four fossil horizons.
The loAvest of them is a yellowish o-rey aru'illnceous sandstone vieldins"
ini]y Dietyopliyllum japom'cum, ])ut in great numljers. The plants of
tliis horizon are easily (hstiuguishal)le from those of the others, being
coloured dark green as if the veget:d)le matter were still remaining (^n
tliem. Tlie next horizon is that of a light greyish argillaceous sand-
stone wliich on weathering also assumes a yellowish colour. In this
horizon all the species below described were found, ]\Ir. K<!cliii)e's
plants ha\ ing been probably taken also from this layer. Tlie two
iij)per horizons have yielded only some fragments of DiclyopJiylhim
japonicwn. I>esides tliese two lu^'izons there is, I ])resume, another,
as I ibund some jnnniv of the snme species in a black slate situated
more to tlie nortli and occupying |)r(jl)ablv a higlier ])osition than the
yaridstorjes. From this, we can see tliat there are several fossiliferous
zones in tlie coal-liearing series of Xagato. Uut at present as the
iiumber of species found in them is very small, it is not possible to
fiiake any palaeontological distinctions in them.

FOSSIL PLANTS FE.OM NAGATO. ^41

Fos.sils, where there they are found in alnindiince, are generally
very well preserved. Owing, hcnvever, to the l)ritt]e nature of the rock
containing them, it i,s very ditticidt to oljtain any large .specimen.

After these brief preliminary remtirks I shall first p:iss to the
description of the species, and then to the conclusions which can be
drawn from them.

Description of the Species.

1. Asplenium Roesserti Fresl sp.
VI. XXXII, Fig. 1-5, Pi. XXXIV, Fig. 2.

Asplen'niin Roesserti Schenk, Fossile Pflanzen aus der Albourskettc gesammelt
von E. Tietze, p. 2, pL I, fig. 2-1, II, 8-10, IV, 19, VI, 33, VII, 3G.

Ayilcnitea llnesscrti Schenk, Foss. Flora d. Grenzschichten d. Keupers u. Lias
Frankens, p. 49, pl. VII, fig. O-Ta, X, 1-4. Zeiller, Examen de la Flore foss. des
Couches de Charbon du Tongking, p. 302, pl. X, fig. 3, 3a.

Chlailtijilt/clii.s itvhliL'Jt.se rar. lloesserti Nathorst, Floran vid Höganas och Helsinrr-
borg p. 42, Helsingborg pl. II, fig. 1-3.

All of our specimens excepting fig. 3, 4, pl. XXXII agree so
well with the figures of Aspleniuiii Eocsseiii given by Schenk and
Nathorst, tliat I have not the slightest douljt about their identity with
this well known species. Tlie pinnules are more or less falcate and
inclined forward, witli secondary veins only once forked. As to the
form of the pinnules, I must say that they are very varialjle, bein"-
sometimes long and fino-er-like, sometimes short and triangular, as
may be sufficiently seen from the spechnens here figured. The
arrangement of pinna' along tlie |)i-inci[)al rhachis is in our specimens
o[)posite or suljoppositc which according to Schenk is said to be the
case in the l(jwer part of the fr(jnd.

242 M. YOKOHAMA.

Specimens represented in li^\ 3, 4, ]>]. XXXH, difler fmm others
in liaving' twice forked secondary veins in s[)ite of the smaller size of
the pinnnJes, nuifli as in fiuiires coinmonly ,u'iven of tlie typical forms
of Asplcnntiii irliitbicnsc^ l>>''jf- (<-'• \X- in J leer's ]]eitr. z. rinraflora
Ostsili, u. d. Amurl. 1876, pi. T, III. and in Schenk's Jurassische
l^flanzen in Ivichthofen's China, vol, J\', pi. LI I.), jhit as it has
been ah-eady sliown by eminent autliorities, that Aspkiiitivi n-ldthicnsc
is synonymous with Alcthopteris indicuin Old. et Morv.^'^) whicli in
turn exhiljits no difference from our Aspltniiiui Bœsscrti,'^^ so it would
be n(3w (phte objectionaljJe to separate tlie abo\'e specimens into dis-
tinct species. StiJl however, as I (obtained no transitional forms be-
tween tlie two, I sh(_)uld prefer to descriljc forms with bifurcate second-
ary veins as Asplcniiun llœsscrtl vav. irltithiciisis.

Asplcniimi Uœs^crli occurs in tlie L ]>per and Lower Gondwjina
System of India, in the Kha3tic of Europe, Persia and Ton<iking, and
in the Lower Oolite of various countries.

This fern is very comiiKjn at Yamanoi. l)eing' the most abundant
fossil next to DictiiopluiUum jupoiiiciiiii.

2. Dictyophyllum ef. acutilobum Bymiit sp.
V\. XXX 11, Fiu-. 6.

J >icti/(ijilii/(hnii. aciitilulniiii Schenk, Fuss. Ptlanzcii a. d. Albouvskctte, p. o, pi.
II, fig. 7. Foss. Flora d. Grenzschichten, p. 77, ph XIX fig. 3-5, XX, 1. Nathorst,
Ftoran vid liuganas och He'singborg, p. ]4, Hngunas aldre p]. I, fig. 8, p. 44,
Hüganäs yngrc, pi. I fig. 10-18, Heisingborg pi. I fig. (J-lO. Zeil]er, Exam, de la
flore foss. da Tongking, p. 311, pi. X, fig. 11.

1) Feistuiantel, Fossil iïlora of tlie Soiilli. Hcirali. OiDuhcana Basin, p. 21», 1S82, Calcutta.
Heer, Bcitr. z'tr Jurajhra Ostsib. u. d. Aiiiiiii. 187G, p. 38.

2) yaporta considers in his '' Piautes Jurassiques " (Palcunt. franc. Terr. Jurass., Végét-
aux) p. 301) C7«l(((io^)/(/c'>/.s ('yi.'./'ic/;/«»;^ Iiöj.<crli Prcsl as ideutlcal with Pecoiilcris ( Asijlcniuia)
whithiensis Br^t.

FOSSIL PLANTS FROM NAGATO. ^43

A fragment of ;i coar.sely toothed pinna, with teeth triangular,
ohtusely pointed at apex and slightly inclined forward, and with
reticulate venation, is undoubtedly a species of Dlctyophylluiii which
is at least very closely akin to DicfijopliiiUniii acntllohuin of the lihcL^tic
of Europe. In our only specimen tlie teeth are closer together than
in most of the figures given of this species, and the secondary veins
slightly zigzag.

Besides occurring in the R luetic of Eiu'ope, this species has been
also described as occurring in that of the Albours Chain in I'ersia and
of Tongking.

3. Dictyophyllum japonicum //. sp.
VI XXXIIL

Although this is the most aJjundant of all the plants found at
Yamanoi, vet not a specimen was oljrained representing a complete
frond, all being isolated ])inna', which may l)e characterized as follows :

Pinnie linear-lanceolate, ljroade:;t near the middle, slightly tai)er-
ing towards both ends, lobed exce[>t near the base where they are
simply wavy (jr entire; lobes more or less inclined forward, triangular
in shape, with the aiiterior margin straiglit or c(jncaAe, with the
posterior margin usually convex, and the apex obtusely pointed.
Ivhachis very stn^ng, straiglit or S(3mewdiat curved, running to the
apex of the pinnae ; secondary veins, coarse, slightly cr(j(jked or zigzag-,
directed forwai'd and going up to the apex of each lobe, thus forming
its median vein ; tertiary veins distinct, somewhat inclined anteriorly
and dichotomizing, the branches forming by their union with those of
the neighb(jurring ones coarse pentag(jnal or hexag(jnal nets, which are
usuallv drawn out in the direction of the median \'ein ; (piaternary
veins very fine, forming secondary nets within the primary ones.

244 M. YOKOYAMA.

A glance at the plate with show th;it a great resemblance exists
Ijetween this species and Thaumatopicris M'dnstcri vav. ahhreviata Gopp.
(Scliimper, Traité de Paléont ^éget, vol. I, pi. XL, tig. 7) from the
Ivhaîtic (3f Franconia. kS(j great is tlxis resemblance, that I was at
tirst inchned to treat the two species as identi'-il ; but a careful
comparison between Schimper's figure and many tens of specimens
at hand seems to show that the secondary veins in (jur plant are
not so strong and rigid as in the European. J^esides, none of our
specimens had the lol^es linear and tinger-like as in the figure of
Schimper, 1)ut always had tliem m(3re or less triangular. Under these
circumstances, I deem it more advisable to treat it as a new species.

DictijophijUum japouicuiii is also not indike Camptoptcrts serrata
Kurr (Xathorst, Floran vid Bjuf, pi. A', fig. o) in the general appearance
of its pinna?. Ihit the latter is said to have very indistinct sec(jndary
veins.

A Spiropteris shown in fig. 5, pi. XX XH', I believe to belong to
DictijnplniUum japuiiicmii^ as it was found in the lowest fossil horizon,
where no other species occur.

4. Dictyophyllum Koehibei //. sp.
VI XXXIX, lig. 1, la.

Pinnœ elongated, deeply pinnatifid ; pinnules ovate or ovately
lanceolate, crenate at margin, obtusely pointed at apex, passing off
either at ri^dit angles from the rluichis, or sliglitly inclined forward.
Khachis moderatelv strong ; secondary veins (piite distinct, somewhat
zio-za"', one in each lobe ; tertiarv veins also distinct, forming- by their
union two to three rows of irregularly polygonal nets ; (piaternary
veins very fine, forming secondary nets within the [)rimary ones.

fJudii'ing from the size of the rhachis and the weaker impression

FOSSIL PLANTS FROM NAGATO. 245

m:i(lo 1)V fliololio:^ on -tone, tliis fern -eeni.- to liave been more delieute
tlinn the ])ref'e(liiio- oiie.

Tlie i»nl\- Jurr<,]'t':in s])ecies whicli ran l)e compared witli it is
Thaniuatdjdcrix Scln'iild Xatli. (= T. Jlraunidna Sdieiik) from the
Hha^tic of Sweden (Xatliorst. Flora \ id 1 [(in-aniis och ?Ielsinf;'h(^r[i', ]).
46 }I()!4-anäs ynu're, ]>I. I, hi;-. 1, 1 h']siiii>-l)()i-o-, p]. fj, fio'. 4) and
Francoin'a (Scluadc, Fli-ra dei- < irenz.-chichten. |). 73, ])]. XA Tld, firr.
1-3,])]. XTX, fii!-. 1.). It lias also erenate ])innn]es; Init these are
U'erierallv linear a.nd inneh lorii^'er, and the erenations finer.

As to Tlie u'eneric deiiomiiiation of our species, I follow Vrof.
Sclienk, wlio considers Thanmatoptcris ^»öpp, :is identical witli Dicfi/o-
pliißluiii Findf ct Iliitt. (liandliiicli dec Falaeontolcio-ie, TI. Ahtlieil.

p. US).

Tlie fiü'mvd s])ecimen is tlie onh' one found.

3. Podozaniites lanceolatus LlnflJ, .s/?.
V\. XXXIV, Fiu-. 3, 4.

I'odnzcDiiitc.s htiiceo'.ntiis Niitlioi'st, Floran vitl Bjuf p. 73, pi. XVI, fig. 2-lOa,
Heer, JavaHora Ostsibiriens, 187(; p. 45, 10(3, pi. I, fig. 3a, pi. XXIII, Ic, 4abc,
XXVI, 2-10, XXVII, 1-8. Beitr. 1878, p. G, 20, pi. V, fig. Ml. Foss. FJora
Spitzbergens, p. 35, pi. 'VII, fig. l-7c,(l. Schmalliunseu, Jtu-aflora Paisslands, p. 2!)
pi. V, fig. 3-5c. Schenk, Javassisclie Pflanzen, in Iliclithofen's China, vol. IV, p,
248, pi. XLIX, fig. 4, 5, p. 255, LI, 3. LII, 8, p. 258. LI, 7, p. 2G1, LIV, 2c.
Yokoyama, Jurassic Plants from Kaga, Hida and Echizen, p. 45, pi. IV, V, VI, 1,
VII, 8b, XII, 18, XV, 12b.

Podozaniites distaiis Zeiller, Exam, flore foss. du Tongking, p. 320, Pi. XL, fig.
2. Nathorst, Beitr. z. foss. Flora Schwedens p. 23, pi. XIII. fig. 0-lG, XV, 20.

ZaDiites distaiifiSGhQuk, Flora d. Grenzscliicliten p. 158, pi. XXXV, fig. 10, XXXVI.

Xow aiid tlien occur leaflets of n Vodozainiir^i wliicli are to he

identified witli tlie \S('I1 known cosmo{)olitan sjiecies ahove name(l.

Our specimens are all in frag-ments, that rejiresented in fio-, 3 heino" the

246 M. YOKOYAMA.

liest, but wnritiiii^' the ti]). Jiidgino- I'roiii its rreneral outline, it seems to
l)e]oi\"- to the vnriety (ffuii'mti ol' lleer in wliidi tlie le-iflets are drawn
out into an acnininate a])ex. Fia'. 4 a])])enrs to liave been nuidi
shorter, and I am not quite sure wliether it really hclono-s liere.

6. Baiera ? sp.
PI. XXXTV. Fio-. 6.

Fra^'ments of loni>', ])'n'allel-sided leaves, ajiparently representing;
lobes of a I'aiera or of a (iinkn'O/. oerau' in some eases tliiekly seattered
on f ices of stone. In one case they were observed arising' from a
commori base, as sliown in the ficaire, each ha\in<i- 8-4 ])arallel veins.
It is much to be regretted that tlie specimens are so im])erfect as not
to allow any ])recise determination.

As to the results to be drawii from tlie study of tlie al)i)ve jilarits,
I must sav that the number of species is yet too limited to allow us to
form any very definite conclusions. Some of them however seem to
lie tolérai )1 y characteristic. DiclijoplinUnvi aciitlloJuiiii, has hitherto
been restricted to the Kluptic of Euro])e and tlie similar formations
of Persia arid Tongking. DicfiicplniUiuii japunicinii, altliough new, ex-
liil)its a great relationship to D. Miuiskri rar. ahhreviaf t (MJp])., Avhich
occurs only in the Rha^tic. A tliird IHctiioplniUmn, 1). Kuchihc^ is
(piite new, showing only a (bstant relation to the Pluetic form V.
Srliniîîi Xath. .sp.. It cannot tlierefore, strictly s])ea.king, lie em-
])loyed in the determination ol' the age. Tlie two other well deter-
minable s])ecies, A^plciiniiii lloa^i^erti and l'oilozaiiiitcs laiicenlatus, are
widely diffused in the Phtpfic as well as in the dnrassic. Thus we
have here two species ])ointing to the Pha'tic, and two species pointing

FOSSIL PLANTS FROM XAGATO. 247

to the liha^tic or to the JurassiV. From these facts, I aiii inch'ned to
believe, at present, that this little flora is somewhat older than that of
the Middle Jurassic of Central Ja]ia]i, r'orres])oiidiiiu- either to the
Liassic or, as it seems more probable, to the ii])])ermost Trias or Rha^ic
of Europe. Ordy the discovery of a o-i-eatei- înmdK'r of species ran
decide the question. Ft is here interesting- to note that a similar fl(^ra
is already known to exist in Ton_o'kin,£i', (Zeiller I. c.) and perhaps also
in China, ^'Xathorst having- recently nieAit'ioncd Dictijophyllum Nihsonl
Brfft. up. and Podo.zamitr.^ lancrolafus (Jisfmis Prest. as occnrrino- in
the " U])pf'i" Yan^-tszi." Anotlier point to be noted in onr flora is
the C(^m])arative impiciicy of sjiecies of Diet ijopluiJl urn ^ a n-enns whicli
had its maximal develojnnent in Europe durinii' tlie bMia^tic time.

1) Xathorst, Om fôreVnnuiten af Dictyophylhim NiUsoni llrtjt. xp. i /v'/ha.^ Kol/nramle Bild-
niufjar. Oefversigt. af Kono-l. V^teuskaps-Akademiens Fnrhaiiillinoar, ISOO. Xo. S.

PLATE XXXII.

Plate XXXII.

Fig. 1, 2, 5. — A.spleiiium lÂœs.serti Fresl. sp.
,, 3, Sa, 1. — Asplenium Kœsserti rrcsl. \nv. wliitbicnsis Bnjt.
5, 6'. — JJictyophyllum ci' aciitilobum Braun sp.

Yokoycnna, Fossil riants.

Jour. Sc. Coll. Vol. IV. PI. XXXII.

Aue tor m lapidejn. ih.l.

PLATE XXXIII.

Plate XXXIII.

Fi(j. 1-7. — Dictyoplivlkini japoiiiciiui //. sp. ; 2 left repreoeuta
the l)asal part aiid 5 the a[>iea] ])ai'( of a piinia.

Yokoyama, Fossil Plants.

Jour. Sc. Coll. Vol. IV. PI. XXXIII.

Au.ctorirtla'pide'm, dd.

PLATE XXXIV.

Plate XXXIV.

Fi (J. 1, la. — I)ictyo]ihylliii!i T\(x-lii1)ei //. sp.

,, 2. — Ayplciiinm Roc:<^erti PresL ftp.

,, 5. 4. — rodoznmites Innceolntns L. el IT. ^p.

,5 Ô. — Spiroptcn'is.

,, 6'. — l);iitT:i ? s]).

Yokoyainciy Fossil Plants,

Jour. So. Coll. Vol. IV. PI. XXXIV.

Auct^T in lap II] Cm V-cl .

Comparison of Earthquake Measurements made
in a Pit and on the Surface Ground.

By

S. Sekiya, Professor,

and
F. Omori, Ri'éakushi.

Imperial University, Japan.

In certain earthquake report« it is stated that there has been com-
paratively Httle or no movement felt at the bottom of a deep pit or
excavation, while 2:reat damasse was done on the surface of the o^round*
and it seems to be generally believed that shocks are felt less intensely in
mines. It is not easy to make instrumental measurements in a mine,
and, in fiict, we have very little exact knowledge of underground
shakings. From a practical point of view, however, with reference t(j
the Imildin"- of houses, it is more interesting' to investio'ate the shakinirs
in pits or excavations such as might be made for foundations. The
only instance of such actual measurements as yet published, as far as
we are aware, is that described by Prof. John Milne in a paper entitled
"On a Seismic Survey made in Tokio in 1884 and 1885 " (Trans.
Seis. Soc. Vol. X.) He made observations in a pit 10 feet in depth,
whose bott(3m was dry and consisted of hard natural earth. Comparing
the maximum amplitudes, maximum velocities and maximum accelera-
tions obtained in the pit during tlie tolerably severe earthquake of

* For iustauce, see Trans Seis. Soc. Vol. VIII. page 98. "The Earthquakes of Ischia."

250

SEKIYA AND OMORT.

March 20th, 1885, with those obtained on the surface ground about
30 feet distant he found that they were in the ratios of 1 : 34,
1:52 and 1 : 82 respectively. But for small disturbances, tlie records
in the pit did not differ much from those on the surface. The
observations we have made are really a continuation of Prof. Milne's,
the same method being adopted in both cases. The results con-
tained in the present paper also show in certain cases some difference
of movement on the free surface and in the pit.

The observations were made in the Imperial University at Hongo,
Tokyo, where the soil is hardened alluvium. The pit is 4 feet square
and 18 feet deep, and is situated only a few yards distant from the
instruments in the Seismological Observatory. Its bottom is ]iaved with
bricks to a thickness of about 2 feet. The soil appears here to be very
homogeneous, so that there will be little difference in earth-shakings
arising from the heterogeneity of ground between the surftice and the
bottom of the pit.

Comparison of the Instruments used on the Surface
and in the Pit.

The comparison in tlie present paper is restricted to the horizontal
components of earth movements. The instruments employed were Prof.
'I. A. Ewing's Horizontal Pendulum Seismographs. For earthquakes
which are not too great these instruments give diagrams which
represent practically absolute motions of the ground.*

The instruments used in the pit and on the surface were made
as much alike as possible. To compare their action, they were
placed on a shaky table, and their diagrams for the same motion were

• See Memoirs of the Science Dep., Univ., Tokyo : No. 9, and the Jour. Science Coll., Imp.
University, Vol. I.

EARTHQUAKE MEASUREMENTS IN PIT AND ON SURFACE.

251

taken. Specimentj of such comparison diagrams are given in PJ.
XXXV. The multiplying ratio of both sets of instruments was intended
to be five. If we go through the diagrams, we see that for moderate
motions both give waves of almost exactly the same amplitudes and
periods. Even small ar^^d irregular ripples are faithfully recorded.
Fig. 1 is for the East-AVest component instruments, and Fig. 2 is for
the North- South component instruments. In the following tables is
given the numerical comparison of the amplitudes of some of the
corresponding waves as recorded by the pit and surfice seisujographs.

For E.W. Component Instruments.

Amplitudes in mm.

GIVEN BT

Ratio.

Amplitudes in mm.

GIVEN BY

Ratio.

The Surface

The Pit

The Surface

The Pit

Instrument

Instrument

s

Instrument

Instrument

s

s.

P-

p

s.

P-

P

1.3

1.4

.9

1.3

1.45

.9

.92

.92

1.0

.9

.9

l.u

.6

.75

.8

1.2

1.2

1.0

.85

.9

.9

2.5

2.45

1.0

1.2

1.2

1.0

2.1

2.6

.8

l.(j5

1.55

1.1

.67

.67

1.0

.4

.4

1.0

1.3

1.3

1.0

.3

.O

1.0

1.05

1.2

.9

.15

.15

1.0

.9

.9

].0

.4

.35

1.1

1.45

1.45

1.0

1.4

1.4

1.0

1.5

1.55

1.0

2.1

2.3

.9

1.45

1.6

.9

2.9

2.6

1.1

1.5

1.5

1.0

1.2

1.05

1.1

.26

.26

1.0

.23

.20

1.2

1.3

1.25

1.0

.15

.15

1.0

1.03

1.2

.9

252. SEKIYA AND OMOUI.

For E.AV. Component Instruments. (Continued.)

Amplitudes in mm.

GIVEN BY

llATIO.

Amplitudes in mm.

GIVEN BY

Ratio.

The Surface

The Pit

The Surface

The Pit

Instrument
s.

Instrument

P

Instrument
s.

Instrument

P

.1

.1

1.0

l.l

1.05

1.0

2

.22

.9

A

.48

.8

2.65

2.85

.9

.1

.1

1.0

2.5

2.35

1.1

.25

.13

1.9

2.2

2.0

1.1

.12

.10

1.2

.65

.85

.8

.2

2

1.0

.82

.7

1.2

.27

.30

.9

2.7

2.55

1.1

.36

.4

.9

3.05

2.8

1.1

.4

.4

1.0

1.75

1.6

1.1

.55

.55

1.0

1.85

2.0

.9

1.9

1.8

1.1

1.1

1.1

1.0

.9

1.0

.9

.18

.17

1.1

1.8

1.8

1.0

1.4

1.35

1.0

2.1

2.0

1.1

1.55

1.55

1.0

1.82

1.9

1.0

1.9

1.9

1.0

.9

.9

1.0

Averag

e of all the ]

L'atios

1.01

For N.S. Component Instruments.

Amplitudes in mm.

GIVEN BY

Ratio.

Amplitudes in mm.

GIVEN BY

Ratio.

The Surface

Instrument

s.

The Pit

Instrument

P-

s
P

The Surface

Instrument

s.

The Pit

Instrument
P-

P

1.9
2.1

1.8

1.7

1.9
1.7

1.1
1.1
1.1

2.9

1.35

1.3

2.75

1.4

1.25

1.1

1.0

1.0

EARTHQUAKE MEASUREMENTS IN PIT AND ON SURFACE. 253

For X.S. Component Instrument«. (Continued.)

Amplitudes in mbi.

GIVEN BY

Ratio.

Amplitudes in mm,

ÜIVEN BY

Ratio.

The Surface
Instrument

The Pit

Instrument

1'

The Surface

Instrument

s.

The Pit

Instrument
P-

1}

I A

1.3

1.1

1.1

1.15

1.0

1.15

1.15

1.0

.58

.55

1.1

\A

1.45

1.0

.7

M

1.1

1.15

1.2

1.0

.92

.89

1.0

1.2

]A

.9

.9

.8

1.1

2.5

2.4.

1.1

.2

.2

1.0

1.7

2.0

.9

.65

.61

1.1

2.1

2.4

.9

.18

.13

1.4

1.2

1.0

.8

.74

.74

1.0

1.5

1.4

1.1

.71

.69

1.0

1.8

2 2

.8

.3

.3

1.0

2.15

1.G5

l.o

.65

.65

1.0

1.3

1.15

1.1

.42

.42

1.0

2.1

1.8

1.2

.4

.4

1.0

1.5

1.7

.9

.31

.31

1 .0

2.0

2.1

1.0

.21

.20

1.1

1.9

1.8

1.1

.16

.15

1.1

1.7

1.7

1.0

.1

.08

1.3

1.9

1.9

1.0

.76

.76

1.«

2.85

2.(3

1.1

Ayerag

<-Q of all the 1

atios

1.04

In the above tables, the numbers are the actual semi-ranges of
motion as recorded by the instruments each divided by 5. These
shew that the two .sets of instruments give on the whole results which
are practically identical, so that their records are at once comparable.

254 SEKIYA AND OMORl.

It should bo «tated tliat the «urface-grouud and the ])it instriunent«
were interchanged with each otlier in June, 1888.

The (j nanti ties calcuJated lor the different earthquakes are : —

(1). The number of waves in 10 seconds, marked n.
(2). Amplitude, (r), or semi-range of motion in mm.
(3). Complete Period, (T), or the time taken to make a complete
for-and-back moticjn of the u'round in sec.

2 TT r

(4). Maximum ^ elocity in mm. per sec, (V), or — ™ — .

V-
(5). Maximum Acceleration in mm. per sec. per sec, ÇA) or — ^.

In (4) and (5), it is assumed as usual that the motion of the
ground is simple-harmonic It is rare, however, that any complete
Avave presents a very good simple-harmonic character during the
whole of its course, but usually differs in extent of motion and in the
corresponding time of describing it in the first and second semi-])hases
of the motion, and so in some cases we have calculated V and A for
the two diffèrent semi-phases oï a wa\'e. Sometimes also we give the
maximum period during the 10 seconds interval.

The East-West and Xorth-South components of the horizontal
motion are not compounded, but the same components in the pit and
on the siu'face are compared separately. It is a well known fact tbat
motions of very quick periods and of small amplitudes generally occur
at the beginning of earthquakes, and in the diagrams appear superposed
on the principal imdulations. In severe earthquakes, such as those of
January 15th, 1887, and of Feln-uary 18th, 1889, these ripples are
very prominent ; and, being very (juick in period, though small in am-
plitude, they have maxinuun accelerations very much greater than
those of the principal waves, which are longer in period though greater
in amplitude. We have also made calculations on some of these ripples,
which can sometimes be identified in the two sets of diagrams. As

EARTHQUAKE MEASUREMENTS IN PIT AND ON SURFACE.

255

may he iningined their calculation is very difficult, especially in the
estimation of their periods, so that any great exactness is not to he
ohtained. The calculation will, however, give some approximate idea
as to the state of things. Hence, for some of the earthquakes, "large
waves " and " ripples " are separately calculated. " Large waves " are
those principal undulations for which calculation is usually made
in earthquake reports, ;and " ripples " are the irregular wavelets
superimposed on them. In douhtful cases the amplitudes only are
given. With respect to n, the numher of waves in 10 seconds, there
is no difference to he found hetween the large waves of earthquakes
observed on the surface and those observed in the pit ; l)ut, for
ripples, the number is often very much less in the pit diagram, because
of the reduction of amplitude and the consequent unification of some
of them amongst themselves. The quantity n is therefore given only
for ripples and not for large waves. The distinction between large
waves and ripples is often very doubtful and does not exist for small
earthquakes.

We may here remark that the maximum acceleration, A, is a
quantity which approximately measures the overturning and fractur-
ing effect of the shocks. In the case of a ripple, whose period is very
short, this effect might probably be also measured by the total amount
of impulse communicated to a body during a semi-phase of the wave,
which is found to be proportional to the maximum velocity.

Records.

For the materials of the present paper we examined the records
of thirty actual earthquakes. Of these, tliree interesting shocks have
their diagrams shewn in PI. XXXVI. and IM. XXX\'II., and their
peculiarities are discussed. The other twenty-seven shocks were com-
paratively small and the different quantities, measured and deduced

256 SEKIYA AND OMORI.

from the actual diagrams, are arranged in tabular form. Notwith-
standing tlie frequent occurrence of earthquakes in Tokyo, simul-
taneous records of the pit and the surface instruments have been
obtained for a comparatively small number of earthquakes. This was
owing to the difficulty of managing the underground instrument.

(1.) — January 15th, 1887. — This was an earthquake of unusual
severity a full account of which has already l^een given.* The beginning
portions of the surface and pit diagrams are given in PI. XXXVI. j^j"
and these for the convenience of comparison are placed side ]\y side.
Fig. 3 is for the E.AY. component, and Fig. 4 is for the N.S. component.
The glass plate whicli received the record of the surface instrument
made one revolution in 128 sec, and that of the pit instrument in 68
sec, so that the latter moved nearly twice as quick as the former.
Such 11 difference of the rate of revolution w<juld however cause no
material difference in the diagram. In these, as well as in the follow-
ing diagrams, the corresponding parts are marked with the same
alphabets, and the short radial lines mark the successive seconds
counted from the beginnings of shocks.

The earthquake begins as usual with tremors. After a few
seconds, the motion liecomes suddenly great. The character of
the motion is striking. The ripples are very prominerit, and these
are superimp<^sed on slower undulations, whose period is alxjut 2 sec.
in the E.W. component, and about 3 sec in the N.8. component.
After a short time the ripples become less evident but the amplitude
of the motion continues to be great, and the maximum displacement
occurs at a point marked o in tlie E.W. component. Comparing now
the surface and pit diagrams, we see that the latter is much smoother

* See the Journal of the College of Science, Imperial University, Japan, Vol. I., Part Til.
or Transactions of the Seismological Society of Jajoan, Vol. XI.

t The complet ' diagram of the surface instrument is given in the same volumf s as cited
a bove.

ÈAliTUQlIAKE MEASÜKEMEX'J'Ö IN PIT AND ON «UKFACE.

257

than ihc lonncr, cspcciallv licar tlicir ])cgiiiiiiiii^s. The iiumhci-s. 1, 2,
3, etc., in the first cohnnn in this and (Jther tahJes are niereJy
uiven for convenience.

(I.) Large Waves. E.W, (component.

No.

Amplitoue.

Pkkiod.

u

AX. VeL.

Max. Acc.

Surf.

Pit

Suvf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf
Pit

Surf.

Pit

Surf.
Pit

1

nun.
l.(J

1 .35

1.2

...

2

1.1

1.26

0.9

• ••

3

1.58

1.45

1.1

4

2.05

1.93

1.1

, , ,

5

1.75

1.54

1.1

...

()

1.7

1.25

1.4

. . .

7

.95

.93

1.0

.86

1.1

0.8

7.0

5.3

1.3

50.

30.

1.7

8

1.05

.8

1.3

.89

.93

1.0

7.4

5.4

1.4

52.

36.

1.4

9

2.4

1.75

1.4

2.0

2.2

0.9

7.6

5.

1.5

24.

14.

1.7

10

3.53

2.65

1.4

2.8

2.0

1.4

7.9

8.3

0.9

o-y

26.

0.9

n

2.2

1.25

1.8

1.3

1.7

0.8

IJ.

4.6

2.4

52.

17.

3.0

12

1.35

.95

1.4

1.5

1.1'

1.1

5.7

■i.3

1.3

24.

19.

1.3

18

2.75

2.55

1.1

1.8

1.7

1.1

9.6

9.4

I.O

34.

35.

1.0

11

1.8

1 .65

1.1

1.6

1.2

1.3

7.1

8.6

0.8

28.

45.

0.6

10

1.4

.65

2.2

.93

1.1'

0.7

9.5

2.9

'5.5

65.

13.

5.0

IG

2.15

1.8

1.2

2.8

l.*^>

1.5

1.8

6.

0.8

11.

20.

0.6

17

.71

.1

7.4

1.1

.6

1.9

4.2

1.0

1.2

21.

10.

2.4

18

1.7

_ . _ • J

0.8

.97

2.7

U.4

11.

5.0

2.1

72.

12.

6.0

19

1.8

1.8

1.0

3.2

2.7

1.2

3.5

4.2

0.8

7.

9.8

0.7

2Ü

1.3

. 5 5

2.4

2.5

1.1

1.8

3.3

2.5

1.3

8.

11.

0.7

21

.7

.1

7.0

1.3

.9

1.4

3.4

0.7

5.0

17.

5.

3.5

22

1.6

.38

4.2

1.0

1.0

1.0

10.

2.4

4.1

63.

15.

4.2

23

1.3

.14

9.0

1.8

.9

2.0

4.6

1.0

4.7

16.

7.

2.4

24

1.65

.44

3.7

2.1

1.3

1.6

5.

2.1

2.1

15.

10.

1.5

25

l.G

1.5

1.1

1.9

1.5

1.3

5.3

6.3

0.9

18.

26.

0.7

26

1.83

.9

2.0

1.7

1 .3

1.3

6.8

|..|.

1.6|

25.

22.

1.1

27

1.65

.85

1.9

3.1

3.2

1.0

3.1

1.7

2.0

/ .

3.4

2.0

28

1.7

.55

3.1

1.5

1.5

1.0
1.2

7.1

2.3

3.1
2.1

30.

10.

3.0

A^

'erao'i

J.

2.3

2.1

258

SEKIYA AND ÔMOKI.

(II.) Large AVuve?^. IST.S. Component,

No.

Amplitude.

Period.

Max. Vel.

Max. Acc.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

1
2

o
1.
5
6

1.12
1.G5
1.65
1.85
1.8.5
1.5

.85
1.25
1.3
2.1
2.4
1.8

1.7
1.3
1.3

0.9
0.8

0.8 1

1.5

1.7

U.9

Ü.3

r,.7

1.0

20.

25.

1.0

A\

^erag"(

-\

1.1

(111.) Jvipple.s. E.W. Component.

No.

Amplitude.

1

^ERlOr

.

Max. V

EL.

Max. Acc.

Surf ' 'P^^

Surf.
Pit

Surf.

Pit

Surf.

1 Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Pit

1

.95

rr .-

JO

1.3

. . .

i ...

9

1.05

.94

1.1

.54

.73

U.7

12.

8.1

1.5

I4U.

70.

2.0

3

.0

.34

1.8

.39

.46

0.6

, 9.7

4.7

2.1

16U.

65.

2.5

4
5

.6
.56

.16

3.8

.29
.25

M

0.4

13.
14.

1.0

9.3

28U.
35U.

14.

25.0

6

.5

.19

2.6

.25

.6

0.4

13.

2.

6.5

320.

21.

15.0

7

1.24

.78

1.6

.34

.6

0.6

23.

8.2

2.8:

430.

86.

5.0

8

.51

not

.45

7.

lOo.

9
10

.92
.75

exist-
ing.

.4
.4

15.
12.

230.
190.

11

1.2 ; .82

1.5

.75

.9

0.8

10.

5.7

1.8

83.

40.

2.1

12

.98 .90

1.1

.4

.7

0.6

15.

8.

1.8

3.7'

240.

73.

3.3

A

verag

0.

1.9

0.6

7.8

KAUTH QUAKE MEASUREMENTS IX PIT AND OX SUTtPACE.

2nd

(I\'.) lvi])})l('s. X.S. (Jdmponeiit.

No.

Amplitude.

Pkriod.

Max. Yel.

Max. Acc.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.

Surf.

Pit

Surf.
Pit

Pit

1

.50

.25

2.2

9

.28

0.7

18.

5.4

3.3

550.

120.

4.6

2

.7

.64

1.1

.28

.55

0.5

16.

7.3

2.2

370.

83.

4.4

3

.32

.29

1.1

.4

.31

1.3

5.

6.

0.8

78.

120.

0.6

4

.59

.:17

1.6

.32

.38

0.8

12.

6.1

2.0

230.

100.

2.3

•5

1.05

.87

1.2

.5

.55

0.9

13.

10.

].3

170.

120.

1.4

ß

.41

.31

1.3

.25

.36

0.7

10.

5.4

1.9

260.

94.

2.8

7

.59

.65

0.9

.53

.8

0.7

7.

5.1

1.4

83.

40.

2.1

A^

'oraf^

B.

1.3

.8

1.8

2.7

III (irr.), tlic two rip]»!*'-^ in;irk('(I 4 mikI n in tlic snrf:i('0-<:.T()imd
<li:i2r;!iii liriNc unite»! into one in tlic ])it dinu'i'iiin, and tliose marked
S. 9, 10 in tiic fornuT do not exist sej);irately in tlie latter.

ri'tliese ealenlations l)e eori'eet, or at least a])])roxiniate, it wonld
a])]iear that tlie maximum Ncloeities and maximum ae<'elerations are
eonsideral)Iy uTeater tor ri])])les tlian lor l:ir£>'e undulations. Sucli a
ditterenee will he found also to l)e tlie ca.se witli otlier severe eartli-
(iuake.s.

(2.)~A])ril 16th, 1S87.— A very small earth(|uake

Max. Ampl.

Period.

Max. Vel.

Max. Acc.

Surf. Pit ^'""^
Pit

Surf. Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

E. W. Cump.
N. S. Comp.

.1 .13 0.8

.1 .15 jO.7

1

.6 11.2

.5

1.1

.8

1.4

12.

4.

3.

260

SEEIYA AND OMOTîT.

(3.) — May 2ii(l, 1S87. — 'l'hi>; is n o-ood ex:nu])]e of :i small carrli-
«[luikc. The motion iii<licatr(l liv tlie pit I'en^rd :i]>|)e:irs lo l)e snuiller
than that indicated hv the surface record.

X. S. (\)]n]")orierit.

)l.

Aver. Period.

Max. Period.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

21.

21.

1.

.5

.5

1.

.7

.9

.8

Max. Ampi..

Period.

Max. Vf.l.

Max. a.

'C.

Surf.

Pit.

Surf.
Pit. j

Surf.

Pit.

Surf.
Pit.

Surf.

Pit.

Surf.
Pit.

Surf.

Pit.

Surf.
"Pit.

.1

.06

1.7

.0

.6

.8

1.0

.6

0^

16.

6.

o
o.

I'l. W. CornjK^neiit. — Maximum am])litude is not greater than
0.1 mm. both in the sin-face .-md ])il di:ioTams. The waves :u-c
1c,o flat h) 1)e C()uiih'(l dcfinitelw

(4.) — May 7th, 1887. — A small earth(juake wliose extent c
motion appears to be rather greater in tlie pit tlian on the surface.

a.

Aver. Period.

Max. Period.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit \^

Pit

E. W. Comp.
X. S. Comp.

22.

19.

19.
18.

1.2

1.

.46

.5

.53

.56

.9
.9

1. .7

Max. Ampl.

1

'ERTOn

Max. Vei,. !

1

Max. Aco.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.i
Pit

Surf.

Pit

Surf.
Pit

.6

1.6

Surf. Pit

Surf.
Pit

E. W. Comp
X. S. Comp.

•1

.15

.1

.13

1. '
1.2^

.6

r-

.4
1.0

1.5

.7

1.

1.4

1.6
.9

11.

13.

25.
6.

.44
2 2

EARTHQUAKE MEARüRElMENTS IX PIT AXD OX SURFACE. '261

(5.) — riiMO 2nt1i. 1SS7.— A small carrtuninkc.

Max. Ampi,.

Period.

Max. Vel. i :\rAx. Ac.

Surf.

Pit

Sui'f.
l'it

Surf.

Pit

Surf.
Fit

.9
•0

Sni-f.

Fit

Surf.:
Pit

Surf.

Pit

Snrf.
Pit

E. W. Comp.
X. S. Comp.

.1
.1

.07
.1

1.4.

1.

I.
1.1

.7
1.:^

A

..-.8

i.n i

2.8

5.

in.

2.8
3.4

1.8
5.

(().) — Till,.' ?,Otli, 1887.— A very small ('artli.|n:.k.'.

K. A\\ ( 'i)]ni)Mii('iit : — Almost insiLi'iiiilcanr, the maxiiimm aiiipli-
tndo lu'iiiL!,' ii«»l UTcatcr tlian .().") mm, in the surface diaLii-am. and
ohscairo in tlic ])ir one.

X. S. (\)m|)(^iioiit : — ill tlio surface (liaaram, rlic maximum
amplifndc is .OH mm., and in tlie pit diaiiram ])r<)l»al)lv n<tt un':il«r
lliaii .07) m]ii.

(7.). -July 2nd, 18S7.— A small eMrlli(inak(

Max. Ampl.

Pertop.

Max. VeI;.

Max. Acc.

Surf.

Pit

Surf.
Pit j

Surf.

Pit

Surf.
Pit

Surf.

Pit «'"■<'

Pit

Surf' Pit '^"^^
i'it

E. W. Comp.
X"^.. S. Comp.

.10

.25

.10
.22

1. i

1.1 i

.0

.7

1.

.9

2.1
2.6

2.1 1.
2.1 1.2

2K. 27.
27. 22.

1.
1.2

(8.) — Inly 22nd. 1887.— An cartlKinake of uwruisc oxivui of
iiK/tion, but of slow period. The charar-ter of this éartlu|nakc is
iiilenvtino'. Inlike mosf oaiilxpiakes, wliieli lu'U'in with (juir-k vil)ra-
tions, <his begins very genfly, witli waves of small amplitude aiid oi'
](^/rio- ])eriod. After .about 10 seconds from the start, the motion
becomes largei' and irregular, and ripples a])pe;u" superimposed on
p"r-incipal undulations. All Ihe irregul.ar \va\elets are t;d"»u|ated.

262

SEKT Y A AND OMORI

n.

Aver. Period.

Max. Pïirtou.

Surf

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

E. Vr. Comp.
X. s. Comp.

3G.

28.

14.

20.

2.6
1.4

.3
.36

.7

.5

.4
.7

1.

1.6
1.7

.6

E. AV. Componeiit.

Xo.
1

2

.->
o

4

•5

6

S

0

10

11

12
13

14
15
16
17

IS
19
20
21

Amplitude.

Surf, Pit

.15
.25
.11
.21
.15

.15

.21

.14

.2

.15

Surf.

Pit

.18 .24

.18 .11

.1 I
.1 I.

. 1 i

.09

13 .11

.12
.14
.15

.08
.15
.08
.20
.21
.16
.21
.34

.12
.15
.15
.05
.12
.05
.19
.15
.23
.25
.31

Averao'e.

1.0
1.2
0.8
1.1
1.0
0.8
1.6
1.0
1.6
1.2
1.0
.9
1.0
1.6
1.2
1.6
1.0
1.4
0.7
0.8
1.1

1.1

Period.

Surf. Pit

.0
1.1 1
.64
.95

1.2
1.2
1.1

.93

.83

.83
1.1

.83

.74
1.0

.7
1.3 1
1.1 1
1.3 1
1.3 1
1.5 2

Surf.
Pit

0.6
1.1

1.0
1.0
1.0
0.9
1.1
1.1
1.0

Ô1.1

.94

.0

.94

.8

.9

.64

.5

.4

.4

.6

.0

0.9
1.1
.9
0.9
1.1
1.1
0.9
0.8
0.9
0.8
0.8

1.0

Max. Vel.

Surf.

Surf. Pit

Pit

1.9 1.2 1.6

1.4 1.3 I.l

1.1 1.4

1.3 1.4

1.4 1.4
1. 1.2
1. .6

.8 9

0.8
0.9
1.0
0.8
1.7
.9

) 1.5

.9

.9

.8
1.1 1

.7

.9

1.1
1.1
.9
1.1
1.8
9 .8 1.1
7 .5 1.4
1. .8 1.3
1.2 .7 1.7
.8 1.0 ,0.8
1. 1.0 'l.O
1.4 1.0 1.4

1.2

Max. Acc.

Surf. ! Pit

24. ! 9.4

8.5 I 8.
U. 14.

12.
5.

5.

10.

13.
6.
3.6
4.6 5.5
6.5 4.4
7.4 i 7.7
7.
4.6
8.7

0.

6.
7.
3.
6.
5.

4.7 3.4
7. I 3.1

3.8 4.t

5. 3.8

6. 3.1

Surf.
Pit.

2.6

1.1

0.8
0.8
0.9
0.8
1.4
0.8
1.5
1.0
1.4
0.8
1.2
2.0
1.0
1.2
1.4
2.2
0.9
1.3
2.0

1.3

EARTHQUAKE MEASUKEMEXT.S IX PIT AND OX SUKFACE. 2ßo

X. S. Coiiiponeiit.

No.

Amplitude.

Period. ; Max. Vel.

Max. Ago.

Surf.

Pit

Surf.
Pit

Surf.

Pit

^-^- Surf.
Pit

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

1

2

.37
.13

.35

2

1.0

.7

.7
.G

.9
1.5

.8 3.3
.4 1.3

2.5

.8

1.3

29.
13.

18.
3.5

1.6
3 7

Ave

rage.

.9

.6 1

1.5

2.7

Tn the latter part of the inoticju, the amplitude seems to he Jar"er
ill the pit diagram than in the surfaee diau'ram. Dut the period
was mueh loiiij'er in the former than in the latter.

O

(9.) — Septemher 25th, 1887. — A moderate earthipiake, like the
preceding" one. The extent of motion a] »pears to l)e larger in the ])it
than on the surface, and consecpiently also the duration of motion
is lono-er in the former than on rhe latter.

n.

Aver. Period.

Max. Period.

Surf.

Pit

Surf.
Pit

' Surf. Pit S"*-^
! Î Pit

Surf.

1.1
1.

Pit

Surf.
Pit

E. W. Comp.
N. S. Couip.

28.

3U.

10.
17.

2.8
1.

.35; 1. .4
.oo .(J .6

1.8
1.

.0
1.

E. ^v. Coi

iipoiicnt.

No.

Amplitude. ; Period.

31

ix. Vj

;l.

Max. Acc.

Surf.

Pit

Surf.
Pit

Surf. Pit , '"^""■^■•
Pit

Surf.

Pit

Surf.
X^it

Surf. Pit ^"'■^•
Pit

1

.25

.46

0.5

1.5

1.5 1.0

1.1

1 .0

0.6

4.1 8.

0.6

2

.08

.1

0.8

.6

.7 0.9

.S

.1»

0.9

8.8 8.1

1.1

3

.09

.22

0.4

1.4

1.5

0.9

.4

.9

0.4

1.8

3.9

0.5

4

.09

.15

0.6

1.2

1.3

0.9

.47

.73

0.6

2.5

3.6

0.7

5

.12

.21

0.6

.73

.76, 1.0

1.

1.8

0.6

9. 15.

0.6

6

.05

.07

0.7

.5

.57 0.9

.6

.8

0.8

7.4 8.5

0.9

A^

'erag'c

i.

0.6

.9

0.7

0.7

26 i

SËKIYA AND OMOKI.

X. tS. Couipoiicnt.

Xu.

Amplitude.

J

^ERIüU.

Max. Vel.

Max. Acc.

1

Siuf. Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

1

.17 .19

0.9

.56

.69

0.8

1.9

1.7

1.1

21.

15.

lA

2

0

.25

0.8

.51

.42

1.2

2.5

3.8

0.7

31.

58.

0.5

'o

.1

.15

0.7

.46

l.O

.5

1.4

.95

1.5

19.

6.

3.2

4

.08

.1

0.8

.54

.6

.9

.9

1.

0.9

11.

11.

1.0

Ö

.lo .15

0.9

.7

.66

1.1

1.2

1.4

0.9

11.

13.

0.8

6

.1 .1-1

0.7

.6

.66

0.9

1.0

1.3

0.8

11.

12.

0.9

7

.1 1

.15

0.7

.8

1.0 0.8

.9

.95

1.0

6.9

6.

1.2

8

.05

.1

0.5

.44

.53

0.8

.7

1.2

0.6

10.

14.

0.7

0

.17

2

0.9

M

.6

1.1

1.7

2.1

0.8

17.

22.

0.8

10

.1

.15

0.7

.6

.6

1.0

1.0

1.6

0.6

11.

17.

0.6

11

.19

.18

1.0

.5

.48

1.0

2.4

2.4

1.0

30.

32.

0.9

12

.06

.08

0.8

"5

.57

0.9

.8

.9

0.9

10.

9.7

1.0

13

.00 M

1.0

.45

.43

1.0

.8

.9

0.9

12.

13.

0.9

M

.09 .!<■)

0.6

.6

.71

0.9

X)

1.4

0.6

10.

12.

0.8

\ô

( 1 7 aliuobt
•'-" nul.

...

.45

.41 1.0

1.

13.

1.1

A^

eragc

.8

<'.9

0.9

(10.) — J)c(viii])L'r IGlli, 1887. — -An ctirllKiiüike of inodcmk' in-
Icii.sily, At ;i lilaiKV, \\\c iiioiioii on tlic surface appears to be larger
aijd more irregular than that in tlie pit.

n.

Aver. Period.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

E. W. Comp
N. S. Comp.

41.
41.

20.
18.

2.

.24
.24

.0

.67

.5
.4

EARTU(^UAIv[': 3IEA!SUJiEMENTS IX PIT AND OX H'JRFACE.

•J{)D

K. W . ( oii»[)()iieiit.

Xo.
1

AmPIjITUDE.

1

'ERIOr

IVJ

AX. Vet,.

Max. Acc.

Siu-f.

Pit

Siui.
Pit

Surf.

Pit

Sni-f
Pit

Surf.

Pit

Surf
Pit

Su if

Pit

Surf.
Pit

.77

.7

1.1

•9

.9

1.0

5.4

4.9

1.1

38.

34.

1.1

2

0

.25

0.8

.5

.8

O.G

2.5

2.0

1.2

31.

16.

2.0

.-b

09

1.1

.65

.8

0.8

2.4

1.6

1.5

23.

22.

1 0

4

.07

no.

2

2.2

...

70.

5

.1

.14

0.7

.5

.7

0.7

1.3

1.3

1.0

il 6.

12.

1.3

6

.14

.14

1.0

.7

.7

1.0

1.3

1.3

1.0

12.

12.

1.0

7

.15

.18

0.8

.0

1.2

0.5

0,8

l.G

■

1.8

17.

5.

3.4

A\

erag-c

Î.

.9

1.3

1 .6

]S . S. Coinpoueiit.

Xo.

Amplitude.

Period.

Max. Vei,.

Max. Acc.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pic

Surf

6.5

Pit
6.3

Surf.
Pit

Surf.

Pit

Surf.
Pit

1

.55

.53

1.0

.53

.53

I.O

1.0

77.

75.

1.0

0

.3

.24

1.2

.55

.72

0.8

3.4

2.1

1.6

38.

18.

2.1

3

.33

.05

6.6

.4

.42

1.0

5.2

.75

7.

82.

11.

7.5

4

.25

.25

1.0

.5

.54

1.0

3.2

2.9

1.1

41.

34.

1.2

5

.25

.15

1.7

.55

.8

0.7

2.9

1.2

2.4

34.

10.

3.4

6

.24

.1

2.4

.8

.67

1.2

1.9

.95

2.0

15.

9-

4.7

7

.16

.03

5.3

.25

.26

1.0

4.0

.73

5 5

100.

18.

5.5

8

2

.11

1.8

.42

.83

0.5

3.

.84

■3.6

45.

6.4

7.0

A^

'crag<

ü

2.6

0.9

3.0

3.7

L^GG

SEKIYA AND OMOKl.

(11.) — Jauiuiry lltli, 1888. — A very .small eurtlKjajike. In
each ctjiuponeiit on the surface, the maximum amplitude is 0.1 mm. ;
while for tVie motion in the pit, it is not greater them .06 mm. The
motion seems here to he much more pron(junced on the surface than
in the pit.

(12.) — April 5th, 1888. — A tolerahly severe earthcjuake, in Avhich
the amplitude is not very large, hut the viljrations are very (juick.

The difference of appearance between the surface aud the pit
diagrams is striking, the small shtu'p waves whicli exist in the former
beini»' mostly flattened in the latter.

n.

Aver. Period.

Surf.

Pit

Surf.
Fit

Surf.

.18
.19

Pit

Surf.
Pit

E. W. Comp.
X. S. Comp.

57.
54.

25.
25.

2.8
2.2

A
.1

.5
.5

E. \V. ( 'ompoiieut.

Xo.

Abiplituue.

Per. OD.

Max. Vel.

ÄIax. Ago.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

1

2
3
4
5

.4

.05

.o

.35

.35

.35
.37
.1

.25

l.i

1.7

3.0

1.4

.7
.52
2

.33
.24

.8
.05

• O

.42

0.0

0.8

0.7

0.0

3.6

7.8
9.4

0 7

9;.

2.7
3.6
2.2

3.8

1.3

2.2
4.3

2.3

32.

95.
300.
120.
240.

22.
34.

50.

00.

1.5

2.8
6.0

4.0

Av

eragt

\

1.8

0.8

2.5

'èA

EARTHQUAKE MEASUlîEMEXTS IX PIT AND ON SURFACE. 2^7

N. S. C'oiiipoiiont.

Xo.

Amplitude.

, 1

'eriod.

M

AX. Y

EL.

Max. a

:c.

I Surf,

Pit

Sin-f.
Pit

Surf.

Pit

Surf.
J it

Surf.

Pit

Surf.
Pit

jSurf. Pit

1

Surf.
Pit

1

.4

.38

1.1

.43

.53

0.8

6.

4.5

1.3

j 90.

53.

1.7

2

.3

.1

3.0

.21

1

.27

0.8

9.

2.3

4.0

270.

53.

5.1

3

.22

.08

2.8

.2

.o

0.7

7.

1.7

4.]

220.

35.

6.3

4

.65

.35

1.9

.71

.7

1.0

1 ^ _
5.^

3.2

1.8

50.

29.

1.7

5

.5

.26

1.9

.47

.7

0.7

0.7

2.3

3.0

90.

21.

4.3

G

.2

.26

0.8

.27

.56

0.5

4.7

2.9

1.6

110.

32.

3.4

7

.15

.15

1.0

.24

.35

0.7

3.9

2.7

1.5

100.

49.

2.0

8

.24

.15

1.6

.24

.44

0.5

1 6.3

2.2

2.9

165.

32.

5.2

9

.18

.23

0.8

.24

.45

0.5

4.7

3.2

1.4

120.

45.

2.7

10

o
.•J

.16

1.9

11

.31

0

1.0

.7

.75

0.9

1 2.8

1.7

1.7

25.

14.

1.8

12

9

2

1.0

.24

.56

0.4

1 5.8

2.3

2.3

140.

25.

5.0

13

.25

.15

1.7

.47

.8

0.6

Î 3.4

1.2

2.8

46. 9.

5.1

14

.4

.18

2.2

15

.34

.2

1.7

16

.22

.25

0.9

, . .

17

.3

.18

1.7

.73

.7

J.O

2.6

1.6

1.6

22.

15.

1.5

Av

era Of c

1.6

0.7

2.3

3.6

(18.)— April 29tlu 1888.— A severe enrtlKHinke. Tliis i.. very
like the precedino- (^ne. liuf mur-li more intense. The beo-hinino-
portions of liotli sets of diagrams are given in D. XXXA^TT, Fig.
5. and Fig. G, The glass ]i]ate of tlie surface-ground instrument
made one revolution in 88 sec., .and tliat of tlie pit instrument in
70 sec. In tlie early ])art of the shock, the vibrations are very (juick,
and witli the exception (^f the wave marked A in the E. W. com-
p )n('nr llierc is no promincntiv lai'i^'c wave, ihoiigli ihc ripples :\vo

268

SEKIYA AND OMORI.

distributed more or less in groups. Here ngain the pit diagrnm
appears niucli smoother than the surface one ; compare, for instance,
the portions marked a, b, c, d, e, f, g, iii the E, AV. comjionent. To-
wards the end, tl\e motion becomes slow.

E. AV. Component.

H.

Aver. Period

Surf.

Pit !^"."f-
1 Pit

Surf.

Pit

Surf
Pit

49.

30. ! 1 .0

2

.33

.0

(I.)— Ripples.— E. W. Component.

No.

Amplitude.

Pebiot

).

Max. Yel.

Max. Acc.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

1

.55

.21

2.2

.23

.22

1.0

15.

ß.

2.5

410.

1 70.

2.4

2

.3

.04

7.5

.2

.19

1.0

j 9.5

1.3

7.3

300.

4U.

7.5

3

.27

\

2

'

8.5

\

270

j

4
5

.4
.35

.55

.2

.25

.8

13.

8.8

'4.3

3.0

400.
220.

34.

12.0

6

.25

1

.22

1 7.2

210.

7

.5

.35

1.4

.17

.47

0.4

19.

4.7

3.9

690.

63.

11.0

8

.4

.25

1.6

.3

.23

1.3

8.4

69

1.2

180.

190.

1.0

9

.3

.28

1.1

.44

.43

1.0

4.3

4.1

1.1

62.

60.

1.0

10

.5

.4

1.2

.25

.8

0.3

13.

3.2

4.0

320.

25.

15.0

11

.55

.48

1.2

.27

.47

O.G

13.

6.2

2.1

300.

80.

3.7

A\

^orao'c

2.3

0.8

3.2

6.7

EARTKQÜAKE MEASUREMENTS IN PIT AND ON SURFACE.

269

(IT.) T>aro;e AV:i\ps. K. AV. ( 'i)iii]M)iierit.

No.

Amplitddk.

Period.

Max. VEr,

M

.ax. Ace.

Surf.

Pit

Surf.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Pit

1

2_

1 .G.J

1.2

1.

.8

1.3

13.

13.

1.0

80.

94.

0.9

2

.42

.3.-)

1.2

.5

.5

1.0

5.3

4.4

1.2

G (3.

55.

1.2

o
•J

.57

. /

0.8

.93

.9

1.0

4.

5.

0.8

26.

34.

0.8

1

.63

.85

.7

1.

1.

1.0

4.

5.3

0.8

25.

.>r>

0.8

5

.53

.5S

.9

1.2

1.1

1.1

' 2.8

3.3

0.9

15.

19.

0.8

A^

-e rag-

0.

1.0

1.1

0.9

!

0.9

(III.) ]'i])])les. X. S. Component.

No.

Amplitude.

I

^ERIOD.

Max. Vel.

Max. Acc.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf,

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

1

.27

.18

1.5

.22

.23

1.0

7.7

4.9

1.5

220.

1 30.

1.7

2

.6

.55

1.1

.32

.43

0.7

12.

8.

1.5

240.

120.

2.0

3

.55

.31

1.8

o
.•J

.35

0.8

12.

5.6

2.1

240.

100.

2.4

4

.37

_2

1.9

.25

.35

0.7

9.3

3.6

2.6

230.

65.

3.9

5

.54

.04

14.

.4

.3

1.3

8.5

0.8

11.

130.

16.

8.0

6

.58

.15

3.9

.24

.3

0.8

15.

3.2

4.9

400.

68.

6.0

Averao-e.

4.0

0.9

3.9

4.0

270

SEKIYA AND OMOl.'T.

(T\ .) Tjîiiyc AVavos. X. S. r(^in])()iieiit.

In (T), tlie ripple-^ mimbered 3, 4, 5, (! whicli are distinct on the
surface have united int(i one wave in tlie ])it. In taking- the
ratios of niaxinuun velocities and maximum accelerations, this sin^'le
wave is compared w^th the greatest of the C(^rresponding' ripples.

(14.) — June 8rd, 1888 — An eartliqu-dve of moderate amplitude.

N. S. Coiujionerit.

n

Aver. Period.

Surf.

Pit

Sarf.
Pit

2_

Surf.

Pit

Surf.
Pit

33.

17.

.•J

.<3

.5

EAliTlK^UAKE MEASUÜEIMENTS IX PIT AND UX SUKFAOE.

•271

Xo.

Amplitude.

Period.

Max. Vel.

Max Acc.

Surf

Pit

.4

.45

.32

Surf.
Pit

0.8

1.1

1.4

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

1
2

3

.3

.5
.46

1.2
1.5
1.

1.2

1.

1.

1.0
1.5
1.0

l.G
2.1
2.9

2.1
2.8
2.0

0.8
0.8
1.5

8.5

8.8
18.

11.
17.
13.

0.8
0.5
1.4

Av

erao-e

].]

1.2

1.0

0.9

E. \y. Coiiipoiioiit.

Max. Ampl.

Surf.

Pit

Surf.
Pit

l.l

.95

1.1

(15.)— ()ctol)er i'Otli, 1888.— A small cnrtli(juake. In this case
the amplitude seems to be mueli u'reater in the })it tliaii on the surface.

/(

AvEE. Period.

Max. Period.

Surf.

Pit

Surf.
Pit

Surf

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

E. W. Couip.
N. S. Comp.

31.

28.

20.
21.

1.6

i.o!

.33

.45

.5
.5

0.7
0.9

.0

.5

.9
.6

0.7
0.8

Ti'I

ÖEKIYA AND UlNlOßl.

E. A\. Couipoiieiit.

Xo.

Amplitude.

Period.

Max. Vel.

Max. Acc.

Surf.

Pit

Surf.
Pit

Surf.

.7
.O

Pit

.Ü

.7

Surf.
Pit

Surf. Pit

Surf.
Pit

S art.

Pit

Surf.
Pit

1

2
3

5
6

.12
.1

.06
.18
.11
.14

2
2

.21
.26
.21
.21

0.6
0.5
0.3
0.7
0.5
0.7

0.8
0.7

1.1
1.3

...

1.4
1.8

0.8
0.7

10.
16.

10.
16.

10.
10.

Av

crag-

e.

O.Q

0.8

0.8

1.0

N^. S. C(Jiiipoiieiit.

No.

Amplitude.

Period.

Max. Vel.

Max. Acc.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.

Surf.

Pit

Siu-f.
Pit

Pit

1

2
3
4
5

Ü

7

8

9

10

.16

.1

.1

.1

.1

.1

.3

.11

.1

.1

.15
o

0

.16

.16

.21

.1-

.28

.27

.25

1.1

0.5
0.5
0.6
O.G
0.5
0.8
0.4
0.4
0.4

.4
.4

.7

.4
.5
.5

.5
.6
.5
.5
.5
.5

0.8
0.7
1.4
0.8
1.0
1.0

2.5
1.6
0.9
1.6
1.3
1.3

1.9
2.1
2.5

2.0
2.0
2.6

1.3

.8
0.4
0.8
0.7
0.5

39.

26.
8.

26.
16.
1 6.

24.
•)2

31.

25.
-5.
32.

1.6
1.2

0.3
1.0
0.6
0.5

A^

^erag(

=»

0,6

1.0

0.8

0.9

(10.) — Xoveinbe'v 2ii(l, 1888. — A small L'artli([uake. Tlic pit
(liaiJfrjim is much suKJotlier than the surface one.

E.VIiTIigUAKi: MEASUKEMENT.S IX PIT AXD OX öUlil'ACE.

21\

Max Ampl.

Surf.

Pit

Surf.
Pit

E. W. CumiJ.
N. S. Comp.

.19

.15

.16

.14

1.2
1.1

(17.)— Jaiiuarv Isl, 1S8D.— A small o:irlh(iiiakev

Max. Ampl.

Surf.

Pit

Surf.
Pit

X. S. Comp.

.04

.05

1.

(18.) — Fcl)riiary IStli, LSSÜ. — A severe eai'tli(|(iakr, in which
there was a considerahle aiiiouiit of vertical motion. The earlier
p(jrti<)ns of the cliagrains of the E. AV. component are given in PI.
XXXVII, Fig. 7. The glass plates of the surface and ])it instruments
made revolutions in 108 sec. and 95 sec. res])ectively. The periods
of the ^ibration are very short and the motion on the surface seems
to be much sharper than in the ])it.

E. W. Comp.

X. S. Comp.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

"5

40.

35.
20.
10.
12.

28.
21.
15.
19.
11.

1.5
1.7

1.3
1.0
1.1

50.

49.
39.
20.
15.

30.
23.
14.

20.
16.

1.7

2.1

2.8

1.3

.9

In this table ii^ nl .... are the number of irregular wavelets in
the successive 10 sec. intervals.

'274:

yEKIYA AND ÜMüKl.

(I.) Uipple«. E. W. Component.

Xo.

Amplitude.

Period.

Max. V]

EL.

Max. Acc.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

1

.24

.23

1.0

.28

.27

1.0

5.4

5.4

1.0

120.

130.

1.0

2

1.05

.8

1.3

.7

M

1.1

9.4

7.7

1.2

84.

74.

1.1

3

1.35

1.3

1.0

4

2.4

2.17

1.1

5

.73

.4

1.8

.

6

.95

1.05

0.9

...

7

.3

.06

5.0

8

.82

.52

1.6

9

.8

.32

2.5

10

1.3

.8

1.6

11

.7

.35

2.0

.24

.25

1.0

18.

9.

2.0

180.

220.

2.2

12

.05

.25

2.6

.32

.23

1.4

13.

7.

1.9

250.

190.

1.3

13

1.05

.8

1.3

.35

.73

.5

19.

/ .

2.8

340.

60.

5.7

14

.31

.2

1.6

.27

.29

0.9

7.2

4.4

l.ß

170.

100.

1.7

15

1.15

.72

1.6

.55

.54

1.0

13.

8.4

1.6

150.

100.

1.5

16

.8

nul

.5

10.

130.

17

1.2

.7

1.7

.33

.3

1.1

23.

15.

1.6

440.

320.

1.4

18

.4

nul

.18

14.

490.

19

.85

mil

20

1.73

1.5

1.1

21

.78

.6

1.3

.26

.64

0.4

19.

6.

3.2

460.

60.

8.0

22

.75

.45

1.7

.36

.6

0.6

13.

4.7

2.8

230.

50.

4.7

23

.3

.00

5.0

2

2

1.0

9.

1.9

4.7

280.

60.

4.4

24

.87

.75

1.2

25

.71

.6

1.2

.3

.7

0.4

15.

5.4

2.8

320.

50.

C.o

26

.95

.79

1.2

.6

.8

0.8

10.

6.2

1.6

100.

50.

2.0

27

.88

.75

1.1

.55

.58

1.0

10.

8.2

1.2

110.

90.

1.3

28

.5

.55

0.9

.3

.58

0.5

11.

6.0

1.8

220.

6ß.

3.3

29

.47

.3

1.6

.5

.32

1.5

6.

6.

1.0

80.

120.

0.6

30

.84

.56

1.3

.6

.8

0.8

8.8

4.4

2.0

92.

35.

2.6

31

.32

.23

1.4

.4

.5

1.0

5.

3.n

1.4

78.

56.

1.4

A^

'^eragc

3.

1.7

.9

2.0

2.8

EARTHQUAKE MEASUREME^sTS IN PIT AND ON SURFACE.

275

(TT.) Large Waves. E. W. Component.

No.

Amplitude.

Period.

Max. Vel.

Max. Acc.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

1

4.1

2.

2.1

2

1.7

1.4

1.2

1.5

1.7

0.9

7.1

5.2

1.4

30.

19.

1.6

3

1.75

1.5

1.2

.72

.79

0.9

15.

12.

1.3

1 30.

m.

1.3

4

2.

l.G

1.2

2_

2

I.O

6.3

5.

1.3

20.

16.

1.2

5

1.4

1.

1.4

2.5

2.4

1.0

3.5

2.6

1.3

9.

6.8

1.3

G

.92

.95

1.0

1.0

1.0

1.0

5.8

6.0

1.0

■Jt .

38.

1.0

7

.8

.92

.9

1.4

].3

1.1

3.6

4.5

0.8

16.

22.

.7

8

1.4

1.1

1.3

1.4

1.3

1.1

6.3

5.3

1.2

28.

26.

1.1

9

1.8

1.35

1.3

2,

1.9

1.0

5.7

4.5

1.3

18.

15.

1.2

10

2.05

1.75

1.2

3.9

3.7

1.0

3.3

3.

1.1

5.3

5.1

1.0

11

1.45

.9

1.6

2

1.9

1.0

4.6

3.

1.5

15.

10.

1.5

12

1.2

.8

1.5

1.7

1.7

1.0

4.5

3.

1.5

17.

11.

1.5

13

1.4

1.3

1.1

2.7

3.

0.9

3.3

2.7

1.2

8.

5.6

1.4

14

1.65

1.

1.7

3.

o.

1.0

3.7

2.1

1.8

8.3

4.4

1.9

15

2.2

1.25

1.8

2.4

2.6

0.9

5.8

3.2

1.8

15.

8.2

1.8

Aa

^erao-

e.

1.4

1.0

1.3

1.3

(III.) Ivipples. X. S. Coniponeiit.

No.

Amplitude.

Period

Max. Vel.

Max. Acc.

Surf.

Pit

Surf.

Surf,

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Pit

1

.85

.7

1.2

.43

.52

0.8

12.5

8.5

1.5

180.

100.

1.8

2

1.65

1.2

1.4

.6

.56

1.1

17.

14.

1.3

500.

150.

3..3

3

.4

.09

4.4

.2

.23

0.9

13.

2.5

5.0

400.

70.

5.7

4

.72

.65

1.1

.3

.3

1.0

15.

14.

1.1

320.

290.

1.1

5

.4

nul

.15

17.

700.

276

SEKIYA AND OMOEI.

Ripples.

N. S. Component.

(Continued)

No.

Amplitude.

Period.

Max. Vel.

Max. Acc.

Surf.

.85

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

6

.78

1.1

7

.75

.85

0.9

.33

1.0

.3

14.

5.

2.8

270.

34.

8.0

8

.1

.14

.25

0.6

14.

3.

4.7

600.

60.

10.

0

.4

.6

0.7

.17

.6

0.3

15.

6.3

2.4

550.

66.

8.3

10

.05

.55

1.2

.1

.o

.5

0.6

14.

7.

2.0

290.

87.

3.3

11

.4

.12

">.3

.14

.25

0.6

18.

6.0

800.

80.

10.

12

.5

.56

0.9

.46

.4

1.2

7.

9.

0.8

93.

110.

0.8

13

.0

.4

1.5

.46

.45

1.0

8.

6.

1.3

110.

80.

1.4

14

.6

.5

1.2

.4

.4

1.0

0.

8.

1.1

150.

120.

1.3

15

.5

.3

1.7

.44

.4

1.1

7.

5.

1.4

100.

74.

1.4

16

.65

.45

1.4

.73

.7

1.0

6.

4.

1.5

48.

46.

1.0

17

.94

.74

1.3

18

.66

.33

2.0

19

.28

nnl

.24

7.

...

190.

20

.99

.89

1.1

.53

.58

.9

12.

9.7

1.2

1 10.

no.

1.3

21

.65

.59

1.1

.24

.4

.6

17.

9.

1.9

450.

150.

3.0

22

.67

.42

1.6

.24

.3

.8

18.

8.8

2.0

460.

180.

2.5
3.8

Ay

eragM

-}

1.6

.8

2.2

(FV.) l.:n*o-(' Waves. N. S. Component.

No.

Amplitude.

Period.

Max. Vel.

Max. Acc.

Surf.

Pit

Surf,
i'it

Surf

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

1

2.75

2.3

1.2

1.6

1.6

1.0

11.

9.

1.2

41..

35.

1.3

2

2.9

3.3

.0

2.0

1.9

1.0

9.1

11.

0.8

29.

•>7.

0.8

3

.8

7.

1.1

.()

.42

1.4

8.4

11.

0.8

88.

170.

0.5

4

.5

.25

2.0

6

.7

0.9

o.o

2.3

2.3

56.

21.

2.7

EARTHQUAKE MEASUREMENTS IN PIT AXD OX SURFACE.

277

(TV.) Large Waves. X. S. ( '(^niponent. (Continued.)

No.

5

Ampeitdde.

Period.

1

Max. Vel.

Max. Acc.

Siu-f.

Pit

Surf.

Surf.

Pit

Surf.
Pit

Surf.

7.2

Pit

0.

Sui-f.
Pit

1.2

Surf.

Pit

Surf.
Pit

1.1

Pit

.85

M

1.4

.74

.60

1.1

61.

57.

6

.85

.8

1.1

.6

.8

0.8

8.9

6.3

1.4

93.

50.

1.9

7

1.4

1.25

1.2

.8

.8

1.0

11.

9.1

1.2

87.

72.

1.2

8

1.4

.7

2.0

1.1

1.2

0.9

8.

3.7

2.2

46.

19.

2.3

9

1 .05

1.4

1.2

.77

.8

1.0

14.

11.

1.3

110.

87.

1.3

10

1.15

1.1

1.3

Am

^era,^(

i

1.3

1.0

1.2

1.3

The ri])])les numbered 16, 18, 19 in (T) aiid ihose imtnlx'red 5,
19 ill (HI), wliieli are (li.-;tiiici on tlie surface, «Id not exi-r in llie ])it,

(19)-- May 6th, 1S89. — A small e.arth(|nake, on wliosc iindii-
lations are sii]>erpose(l minute irregularities. ITerethe motion appears
to \)v ratlier greater aii<] oi' longer dni'ation on tlie surl'aee than in the
pit.

n.

Aver. Period.

Surf.

Pit

Surf.
Pit

Surf. Pit

Surf.
Pit

E. W. Comp.
X. S. Comp.

43.

50.

33.
22.

1.3
2.3

.23 .3
.2 .4

0.8
0.5

278

SEKIYA AND OMOK'T.

E. W. Component.

Xo.

Amplttdde.

Periob.

M

AX. Aoc.

M

AX. Acn.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

1

.1

.1

1.0

1.

.7

1.4

.6

.9

0.7

4.

8.

0.5

2

.05

.05

1.0

1.3

1.7

0.8

.2

.2

1.0

1.

.7

1.4

3

.07

.07

1.0

1 1.6

1.4

1.2

.3

.3

1.0

1.

1.5

0.7

4

.08

.09

0.9

1.3

1.4

0.9

.4

.4

1.0

2.

2.

1.0

Ö

.0.5

.07

0.7

1.3

1.4

0.9

.24

.32

0.8

1.

1.5

0.7

6

.05

.00

0.8

' 1.

1.2

0-8

.3

o
.'J

1.0

2

2

1.0

A^

'era,i^(

3.

0.9

1.0

0.9

0.9

X. S. Coniponpiit.

Max. Ampi,.

Perfod.

Max. Vel.

INFax. Ace.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

1.6

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

.13

.09

1.4

.8

.5

1.

1.1

0.9

8.

14.

O.G

(20.)— :\[:iy nOth, 18S1).— A smnll otirtliqiinke.

)i.

Aver. Period.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

X. S. Comp.

32.

21.

1.5

.1
.o

.5

0.0

E. W. Comp.
X. S. Comp.

Max. Ampl.

Surf.

0.8

.3

Pit

Surf.

Pit

0.8

2.0

Period.

Surf.

.33
.9

Pit

Surf.

Pit

.46 0.7
.4 2.3

Max. Vel.

M

AX. Acc.

1

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

1.5

2.1

1.4
2.4

1.1

0.9

28.
15.

20.
38.

1.4
0.1

E.VUTllgUAKb: MEASUKEMEN'TS IX PIT AXD OX SURFACE.

279

(21.) — Tune 1st, 1889.— A very small e:irtli(|iiake.
Ill the E. W. and X. S. Ci)Uipoiients ofljoth tlie surface and pit
diau'ranis, the niaximuin amplitudes are aliont .Oo mm. and .02 mm
respectively.

(22.) — dime ord, 1(S89. — A very small eai'tli(piake.
In the E. W . Component of the l)()th the surface and pit diiiü-rams,
the maxinunn am}>Iitude is about .03 mm.

X. kS. Component.

Max. Am PL.

Period.

Max. Vel.

Max. Acc.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

1.

.1

.05 2.

1.

.1

1.4-

.6

.5

1.

4.

4.

(23.) — dune 15th, 1889. — A small earth(juake.

Max. Ampl.

Period.

Max. Vel.

Max. Acc.

Surf.
.07

.07

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

,,, s^

Surf.

Pit

Surf.
Pit

E. W. Comp.
X. S. Comp.

.07
.05

1.0
1.4

.8
1.0

.8
.6

1.0
1.6

.55
.44

.55
.53

1.0
0.8

4.3

2.8

4.3
5.G

1.0

0.5

(24.) — June Kith, 1889.— A small eartlnpinke.

I']. ^\ . ('omponenl. Uji the smvfiice, the maximimi atii])lilude is
not greater than .02 mm., and in the pit if is al)out ,0') mm.

X. S. Cijmponent. On the surface, the maxinumi amjtlitude is
aljout .07 mm., and in the ])it it is aljout 0.1 mm.

(25.)— June 20tli, 1889.— A smalf earthcpiake. In this case,
the amplitude seems to be rather greater in the pit than on the surface.

280

SEKIY.V AND OMOKl.

n.

Aver. Period.

Max. Period.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pif

Surf.

Pit

Surf.
Pit

E. W. Couip.
N. S. Couii>.

35.

48.

25.
21.,

1.4
2.3

.3

.4

.5

.8

.4

.(3

.7

.9

Max. Ampl. '

1'eriod.

Max. Vel.

1

Max. Acc.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.

Pit

j

Surf.

13.
17.

Pit

Surf.
Pit

E. W. Couip.
X. S. Comp.

.08

.07

.14
.11

0.6

O.G

.5
.4

.7
.5

.7
.8

1.
1.1

1.

1.4

1.

.8

12.
18.

1.
1.

(26.)— June '27\\\, 1889.— A .snu.ll earthquake.

Max. Ampl.

Period.

Max. Vel.

Max. Acc.

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
I'it

Surf.

Pit

Surf.
Pit

Surf.

Pit

Surf.
Pit

E. W. Comp.
N. S. Ouujp.

.1

.1

.05
.03

o
3.

1.4

1.2

1.2

.5

1.2
2.4

.5
.5

.3

.4

1.7

1.3

2.

2.7

1.3

4.8

1.5

0.0

(27.)— July ord, 1881).— A small eartlujuake.

Til the E. \\ . Coiiipoueut, the uiaximuiu aiiijviitude is .13 iiiiii.
on the surface, and .1 nun. in the pit.

(28.)_Fel)ruarv loth, 1890.— A very small earllKiuake.

E. W. Component, lioth on the surface and in the pit the
maximum amplitude is not greater than .05 mm.

N. S. C(miponent. lîoth on the sm^fuce and hi the pit, the
maximum amplitude is ahout 0.1 mm.

EARTHQUAKE MEASUREMENTS TX PIT AND OX SURFACE.

^29.)— April 11th, 1890.— A siualJ earthquake.

281

Max. Ampl.

Surf.

Pit

Suvf.
Pit

E. W. Comp.
X. S. Comp.

.13
.1

.15

.or,

0.9
1.7

(^30.)— A])ril 18th, 1889.— A very Hinal] earthquake.

Max. Ampl.

Surf.

Pit

Surf.
Pit

E. W. Comp.
X. S. Comp.

.07
.00

.04

very
small.

2.

Summary of Results.

It is .generally heUeved that the earthquake motion is considerahly
less m a pit tlian on tlie surface. From the fore,<?<^ing calculati.^ns it
.seems prohahle that this is true for some earthquakes and not true f.n-
.^tliers. Among the thirty earthquakes we examined, there are three
which were especially severe. These are (1), (13), and (IS). The
rest are small earthquakes of the kind that daily occur in Japan.
The ratios of the amplitudes, periods, maximum velocities and

282

SEKIYA AXD OMORI.

maximinn nccelerntioiis f(ir some of these Isitter era'tlujiinkes :is observed
on the free snrfice ground to those ol)served in the ])it are collected in
the following t:i1)le, average values being nsed when a number of
waves liave been calculated for a single earthquake.

(\o.)

Katio op
Amplitudes

Katjo of
Periods

Ratio of Max.
Vel.

Eatio of
Max. Acc.

E. W.

X. 8.

E. W.

X. s.

E. W.

X. s.

E. W.

X. s.

{■■i)

0.8

0.7

.5

1.4

3.

(3)

1.7

.8

2.0

3.

(4)

1.0

1.2

1.5

.7

.6

1.6

.44

2 2

(-')

1.4

1.0

0.9

0.5

1.6

2.3

1.8

5.0

(7)

] .0

1.1

1.0

0.9

1.0

1.2

1.0

1.2

(8)

1.1

0.0

1 .0

0.6

1.2

1.5

1.3

2.7

(9)

o.n

0.8

0.9

0.9

0.7

0.9

0.7

1.1

(10)

0.9

2.6

0.8

0.9

1.3

3.0

1.6

3.7

(12)

l.S

i.n

0.8

0.7

2.5

2.3

3.6

3.6

(14)

1.1

1.1

...

1.2

1.0

...

0.9

(15)

(».<)

0.6

0.8

1.0

0.8

0.8

1.0

0.9

(IG)

1.2

1.1

(19)

0.0

1.4

1.0

1.6

0.9

0.9

0.9

0.6

(20)

0.8

2.0

0.7

2.3

1.1

0.9

1.4

0.4

(22)

2.0

1.4

1.0

...

1.0

(23)

1 .0

1.4

1.0

1 .6

1.0

0.8

1.0

0.5

(25)

Of)

0.6

0.7

0.8

1.0

0.8

1.0

1.0

(26)

2.0

3.0

1.2

2.4

1.7

1.3

1.5

0.6

(27)

1.3

.

(29)

0.0

1.7

...

Ay en

Ige. ...

1.00

1.4

0.9

1.1

1.2

1.4

1.3

1.9

Average for bot]
Components.

1

2

1.

0

1

3

1.

6

EARTHQUAKE MEAöUKEMEXTsi IN I'lT AXiJ OX 8UKFACE.

1>83

This tabic .'seem.s In shew that li»i- . mnll (':iflh(jiiak(,'.< the aiii|)htiide
aiid the period are oii the \vh(j]e nearly tlie same on the free surface
and in the pit, there Ijcing a slightly greater motion on the surface.
This confirms the result which Prof. Milne previously obtained. In the
above are not included those very small earthquakes, whose measure-
ments are difficuH ; these li(_)\vever shew tliat the motion in the pit
is also small when tlie motii^n obser^■ed on tlie surf see ground is small.

It must l)e noticed that the diagram taken in the pit appears
always to be smofjther tlian that obtained on the surface, and //, or
the number of irregular wavelets occurrinii' in 10 seconds, is f )und
in every case t(j l:ie greater for the latter, l)eing often twice as manv as
for tlie former. 'J'his is very remarkaldy shown in the three severe
earthquakes mentioned above, for which calculât ioii;; ha\e Ijcen made
separately as regards large undulations and small ;;uperposed ripples.
The ratios of the amplitudes, peric)ds, maximum velocities, and
maximum accelerations for the surface and pit motion of these three
eartliqiiakes are gi\en in the fjlloAving tables.

(I.) Large undulations.

(Xo.)

Ratio of
Amplitudes

Eatio of
Pkriods

Ratio of Max.
Vel.

Ratio op Max.
Ace.

E. W.

X. s.

E. -W.

X. s.

E W.

X. s.

E W.

X. s.

( 1)
(13)
(18)

2.3
1.0
1.4

1.1

1.2
1.3

1.2

1.1

l.(J

I.l

1.0

2.1
0.9
1.3

1.0
1.2

2.1

0.9
1.3

0.9
1.3

Aver

\ge. ...

l.G

1.2

1.1

1.1

1.4

1.2

1.4

1.1

Average for bot!
Components.

1.

4

1.

1

I.

3

1.

3

284

SEKIYA AND OMORL

(ir.) IJipple-s.

(No.)

Ratio ok
Au)j)litiKles

Ratio of
Perious

Ratio of Max.
Vel.

Ratio of Max.

Ace.

E. W.

N.S.

E. W.

X. s.

E. W.

N. >S.

E. W.

N. 8.

( 1 )
(13)
(18)

1.9
2.3
1.7

1.3
4..0
1.6

o.r,

0.8
0.9

0.8
0.9
0.8

3.7
3.2

2.0

1.8
3.9
2.2

7.8
6.7
2.8

2.7
4.0
3.8

ig-e. . . .

2.0

2.3

0.8

0.8

3.0

2.6

5.8

3.5

Average for both
Comj)oneats.

2.2

0.8

2.8

4.7

It will 1)e tliiLs observed tliut for principal iindiiLitions of severe
eartli(|iiake.5 tlie raiii^e of motion is somcwliat o-reater on the surface
than in the pit, Ijiit there is no great difference of inaxiniiiin velocities
and niaxiinuni accelerations Ijetween the two sets of observations.
This seems io he due to the fact that for the larger nndulatic^ns tlie
period will .'.omewhat increase with the amplitude. In fict, tal)le (I.)
would ap[)ear to iiidic:he some slight increase (of period on the surface.
Tlie case is different with ripples, for which tlie results are more
uniform and llie difference of surfice and imderground effects more
decided. From Taljle (H.) the average extent of liorizontal motion in
the pit is only half that on the surface gnnind, and the period for
the former seems rather greater than f )r the latter, which arises
from the fict that veiy many of the ripples disappear in the pit. Tlie
maximum velocities and maximum accelerations on the surface are
respectively ahout tliree and live tunes those in the pit.

Our conclusion tlien is that for small eartli(|iudves there is no
practical difference between the surfice and undergr(3uiid ol^servutions ;
for the principal undulations of severe eartlujuakes this difference
may exist, L>ut not to any marked degree ; but for the small quick

EARTHQUAKE MEASUREMENTS IX PIT AND OX SURFACE. 285

vi1)ratioiis tlic difference is coiLsi(leral)Je. Xow, tlioiigli only ap-
proximate tlie calcultition for the ripples may l)e, their maximum
velocities and mtixinuim acceljrati<jns are found to he very _UTe:)t, and,
in fact, many times greater than those for the principal undulations.
And thus', if these ripples are really in great part smoothed away
in the ])it, it is \cvy likely that in times of such severe earthquakes as
discussed aho\'e, there might ])e less destructive action in deep pits
than on the free surface.

We shall n(jt venture here to discuss what these ripples may be.
They exist only in the early part of the sIkjcIvs and seem to be the
continuation of the tremors which occur at the beginninii' of earth-
quakes. The aj>})earance of the diagrams of the severe earth(|uakes
is very much like that of the disturljances in the sea where minute
ripples are siiperp<r>ed ou large undulations. If the ripples l3e regard-
ed as waves travelling on the surface, then the whole thing will admit
of an easy explanation,

AVe must state ho\ve\er that these observations were made at
li(jngo, where the gr()und is hard, and it is needless to say that tlie
character of the earthquake motion depends in a great measure on the
nature of the soil. Hence it is ([uite possible that observations in
different places may lead to scnnewhat different i-esults than those
oljfained here. Thus, for instance, at I liiotsul)ashi, where the soil is
soft, the range of miction is two or three iimes greater than that at
llong(j, and yet the e;!rth(juake diagrams oljtaiiiecl ihere seem to be
conq)aratively free of superposed wavelets.

In the above the observations were confined ro the horizontal
conqxjnent motion alone. The usual argument lljr tlie supposed
sniallness of the motion at a suljterranean ])oint is derived from the
behaviour oi' a row oî ivory balls in ctjntact with each otlier when one
at the end is sharply struck. This argument appears to apply rather

â86 SËKtYA AND OMÔEt.

to the vertical component than to the horizontal. It is our intention
to continue these observations we have been making and in addition to
investigate the nature of the vertical motion in the pit.

Laboratory Notes,

By

C. C. Knott, D. Sc, F.R.S.E.

Professor of Physics.

1. Electric Resistance of Cobalt.

The manner in wliich the electric resistance of cobalt varies with
]\\(j:]\ temperatures does not seem to have been studied with any great
care. The peculiar beha^'iour of nickel and iron as regards their change
of resistance with temperature is now well knowri'^ With a view to
see if cobalt presented anv similar peculiaritv, I set ^Ir. Omori, one of
the graduating students in Physics, to investigate the (juestion.

The piece of cobalt used was cut from a sheet of rolled cobalt
which had been given me by l^rofessor Tait. T)r. E. Divers, F. R. S.,
kindly determined its composition by an analysis of a very small
quantity (about 20 grains) sup])h'ed lu*m. Tlie result of the analysis
is as follows :

Carbon ft)und 0'77 7o ^^"^^.Y ^^ ^^ much as 1-00%

Silicon 0T5

Iron 0-73

with a minute quantity of manganese and perhaps ^vVo *^^ ^ metal
undetermined. Dr. Divers regarded it as of remarktd)le purity for a
furnace product.

I) See my pajDer On thn Electric Resistance of Nickel at ITitjh. Temperatures, Trans. Royal Soc,
Edin , Vol. XXXIir (1880)— also abstract in the Journal of the Colleije of Science, Tokyo, Vol. I.

288 0. G. KNOTT.

Tlie method of experiment was essentially the same as that des-
cril)ed in my earlier paper on nickel. Four stout copper rods, 60 cm.
]ony and 0'7 sq. cm. cross section, were fixed in a vertical position some
little distance apart. Their lower extremities were joined in pairs hy
two coiled wirei^', one of which was a specimen of nearly pure platinum
and the other the cobalt strip that was the special ohject of investiga-
tion. The upper extremities of the rods were joined 1)y stout copjX'r
sti"i])s to a commutat(^i', which was in connectiori with a Wheatstone
Th'ido'e resistance l)Ox of ordinary construction.

Tn one series of experiments the lower ends of the rods with their
c(ii meeting wires were dipped in a vessel of i^il wliich could he heated
u]) to a temperature of nearly 240° C. A thermometer, centndly
placed so that its Indh lay at the mean level of tlie ]>latinum and
rohalt coils, was used for measuring the temperature. The <^il was
heated very gradually and was kept briskly stirred until a few seconds
before a rending was to be taken. One of the wires was niennwliile
thrown irito the Wheatstone Bridge, and the resistance adjusted
slightly in advance. The temperature was then allowed to rise very
slowly until reversal of the commutator in the galvanometer branch
gave no deflection. AVhen the equilibrium was thus attained the
thermometer reading was noted. In this experiment chief attention
was given to the cobalt ; a few measurements of resistance were made
with tlie platinum, sufficient to give the most important temperature
coefficient.

The resistance curves for the cobalt and the pl;itiiuun are shown
in the diagram (p. 298), Curves Nos. 1 and 2. All corrections have
been carefully applied and tlie resistances are in legal ohms.

lîy interpolation amongst a number of contiguous measurements
the resistance for each of the temperatures 100°, 140°, 180°, 220" C.
was calculated as shown in Table I.

LABOliATOKY NOTES.

289

Table I.

Resistance of a Cobalt Strjp in Legal Ohms at

DIFFE RENT TeMPEEATÜKES.

Température. Resistance.

1

First Differenoe.

Ratio.

1 00° C.
1 40
180
220

.12:340
.1:3094
.15210

.10851)

.ol;354
.01510
.01 049

1.1097
1.1109
1.1084

Since the second differences have appreciably ditferent values, it
i> iinpcjssihle to represent the law oi" chani^'e by means of a paralxjiic
liinction. ]>ut the reniarkaljle ('oustancy oï the ratios of successive
pairs of resistances sugi^'ests an exptjnential function of the temperature
as the expression for the resistsiuce.

Thus we may put

it. — il,, a
from which we tind, if / is the temperature in degrees Centigrade,

k = -OOi^GOo, R, = -09519
The measured resistance at 7°*5 C. was 0*09604, which doe.> not diilcr
ivom the value given by the f)i'nuil;i 1)V more than 1 ])er cent.

In my ])aper already referred t(j \ fxuid that the same form of
ex})ression held for the case of <jne of the nickel wires investigated, the
only essential difference Ijcing in the vahie of the coefïicient k, which
for the nickel was .003.

The resistance of cobalt therefore d«)es not change so quickly with
temperature as does the resistance of nickel.

Tn the second series of experiments the lower ends of the rods
with their connecting wires were insei-ted into a porcelain vessel.
Asbestos was wra])ped r<nmd the wires ; and the whole was heated in

290

C. G. KNOTT.

u charcoal furnace. The observations of resistance were made as the
system was coohng, the Cfjljalt and platinum being thrown alternately
into the Wheatstone Bridge. The instants at which the Ijalancing was
effected were carefully noted, so that it was an easy matter to inter]:>olate
between two successive measurements for the (3ne wire that resistance
which correspc^nded to the intermediate measurement for the other.

By this means more than twenty distinct pairs of measurements
were obtained, every C(jbalt resistance having its corresponding plati-
num resistance. After all corrections were made the platinum resis-
tances were di\'ided l^y the resistance of tlie |)latinum at 7°C. ; and
similarly all the cobalt resistances were di\ideil by the resistance of
the cobalt at this same tempérât lu'e. The numbers were then classified
into groups so as to afford the means of calculating by interp(3lation
the cobalt resistances which corresponded to assumed convenient values
of the platinum resistances. These are the numbers given in Table
II. which epitomises the results of four distinct experiments. The
measurements were all made during cooling, and the higher values are
accordingly tabulated first. The first column contains the platinum
resistances, taken as convenient multiples of the resistance at 7° C.
measured afler the experiment ; and the ()ther columns gi\'e in order
the corresponding resistances of the cobalt.

Table II.

Platinum
Resistances.

CoBALi' li

E3I STANCES

Exp. r.

Exp. II.

Exp. in.

Exp. IV.

2.0

5.8047

5.799(3

5.9748

0.0361

1.8

4.5101

4.3423

4.4511

4.4580

1.0

3.1822

3.0536

3.0932

3.2210

1.4

2.2029

2.1 795

2.1111

2.2002

1.2

1.5329

1.5337

1.5050

1.0

1.0000

1.0000

1.0 000

l.OOUU

LABORATORY XÔTEy.

291

If we p.s:>mne thîit the change« in the platinum re>«i«tance follow
the same law n^i in the earlier experiment with the oil, the rise of tem-
perature which will just (l(juble the resistance is about 680° C; and the
interval from 1 to 1.2 may be taken as c<jrresponding approximately to
a rise of tem})erature (jf 136°C. According to the ex])eriment in oil,
the resistance of the cobalt would ha^'e Ijeen increased in the ratitj
1.42-48 to unity by this rise of temperature. It is p.pparent then that
under the influence of the first excessive heating tlie cobalt has been
considerably altered in its pro])erties, so tliat the average temperature
cœfficient for resistance up to 150°C. has been increased by a quarter.

That the successive heatings caused a marked change in the
structure of the wire <jr strip is shown l)y the variations in the measur-
ed resistance at 7°C. These are given in Taljle III.

Table III.

When Measured.

Resistance of
Platinum wire

Resistance of
Cobalt Strii).

At the be^'inuiuu'

.8525

.09724

After 1st heating

.85028

.09135

2!id

.85028

.09354

ord

.85013

.09(374

„ 4th „

.85232

.09978

The fall in resistance alter the first heating is no doubt due to
some change in the contact resistances. It characterises both the
platinum and cobalt. !Subse(pient heatings howe\er do not change
the platinum to any great extent until the very last experiment ; but
their effect on the cobalt is very marked. After the experiments were
completed the col);dt was found to be nuich ahered l)v oxidation. It
was exceedingly brittle and broke into small pieces when it w^as beinij-

292

c. G. KXOTT.

detîiched f"r<Jin the copper rods. AMiile the oljservatioii« were being
made, it was noticed that the loiu'th experiment was much inferior in
point of regularity and steadiness to the others, a fact sufficiently ex-
plained l)y the tinal condition of the met;d.

It is not sin"])rising, tlien, that there is considerable divergence
Ijetween the \alues of the tem])erature coefficients a« obtained from the
earlier experiment in oil and the later series in the charcoal furnace.

What is surprising is, that in spite of the great alteration in
structure going on in the sti-ij), the general behaviour of the cobalt
as show^n in the first three experiments is essentially the same. This is
well seen from the tabulation of the rates of change themselves. These
quantities were calculated by tlie same general method of interpola-
tion as was used in calculating the resistance. They correspond to
the values of dy/dx if y/ and x are taken to represent respectively the
corresponding resistances of cobalt and ])latinum. They are given
in Table IV, the first column containing the \alues of the platinum
resistances to which the tabulated rates of change correspond.

Table IV.

Platinum

Rcsistaucu (Ar-

bitrai-y Touip'r

scalo).

Ratks uf Change of Cohaivi' Resistanck i'Ei;
uxrr Change of Platinum Resistance.

Exp. I.

E.\p. U.

Exp. III.

Exp. IV.

2

1.8

1.6

1.4

1.2

7.02
0.1'J
5.45
3.70
3.58

7.80
7.24
5.57
3.58
3.23

10.33
0.7 i
0.03
3.65

2.78

9.15
5.09
6.10
3.60

I have thought it sufficient to give the condensed numerical results
as contained in Tables I, Hand 1 \'. The indi\idual observations U])on
which these results are based are shown iiraphicallv in the diagram.

LABORATORY NOTES.

29B

; i 1 : i

' I !

!

1

®

V.Ü

-0.17-

Electric Resistance
of Cobalt

(1) C()l):ili la's. — 'r('iii])CT:itnre

(2) IM.'itiiiimi lies. — 'rein])er:itnre

(3) ('o1):ilf :ni(l IMtitinuin llesistniiccs

. - Exp. I.
X = „ II.

0- „ III.

1 O = » IV.

?

/

o

0 ^

0 /

/

6.0

-0.16-

(,

/

A-

-0 15—

— c

5.0

1

/ 1

/ è

/ .r

-0.14-

/;

/ si

4. f)

n n

1

~7'

/

S (3)

. 1

2i'

-dl

SI

/

®

CO

p
Resistances (1)

/

/

/

®

(2),/'

1

o
8.0

/

©

®

^

X

-n 11

e

0

/'

o

1

/

/

®v

°

J

/

^

® 1 °

ÜU

=§

0.10 ^

9 y^

®

r^: °"

Teni

perature

Soalefor (1) an

1(2)

0°.e 2

0» 40,^^" 6

O" 80" 100° 120" l^C 160" 180° 200' 220°

1 1 !

1 !

i

' 1 !

Platin-um Resistance (3)

1,10 1.20 1.30 1.40 1.50 1.00 1.70 l.SO 190 8.00 2,10

294 C. G. KXOTT.

Curves 1 and 2 luno already l)een mcntii^ried. Tliey sliow tlie
mareli of resistance with temperature as measured on a mercurial
centigrade thermometer. \\i Xo. 3, tlie ])latiiiiim resistances are vir-
tnally used as temperatures, and form the al)scissa?. The ordinate«
are tlie corresponding cohalt resistances. The points belonging to
the various experiments are distinguished by special mark.

It will he seen at a glance tliat in ovie ])articular cobalt behaves
very like iron and nickel. There is a ra])id increase in the steepness
of the curve at the higher temperatures. Fn iron and nickel this
rapid increase is followed at still higher temperatures l)y a distinct
decrease, the curves bending so as to present a concavity toward the
temperature (or platirium resistance) axis, Ta.ble IV. gives no hint
of such a tendency in cobalt. The curves all iDCCome steeper with
rise of temperature, if we except the distinctly irregular indications of
Experiment TA'.

It Avill be seen from Table TV. that Experiments T. and H. are in
fiir agreement throughout ; arid that all four experiments point to the
existence of a critical temperatm-e, at which the resistance begins to
increase rapidly with rise of temperature. This critical temperature
is al)out the stage 1.5, which corresponds approximately to 350° C.
Tlie same conclusion may l)e drawn from Table IE and expressed in
these terms. Between the temperatures 400° and 700° C. the resist-
ance of a collait strip increases on the average at a rate nearly twice as
great as the average rate of increase loetween 0° and 300° C

2. The Thermoelectric Positions of
Cobalt and Bismuth.

So fir as I know, the only satisfictc^ry determination of the
position of the Cobalt line on the thermoelectric diagram was made by

LABORATORY NOTES. 295

Professor Tnit's students in the l^hysicnl T/il)ornt(~)rv of Edinhuro-li
University some fifteen ^^ears ao;o. The position of Ihe C^ohalt line, so
I'onnd, was o'iven alonof with the positions of certain alloys in a ])a])er
1)V Professor J. Gordon ^lacGreg'or and myself ])ul)Iished in ihe
Transactions of the Poyal Society of Edinhnr£;-li, A'ol. XXVIIT (1878).
The particidar specimen of Cohalt used in tliese early experiments wa«
a short rod ohtained hy electrol^^tic deposition. The noteworthy ficts
reo'archn«^'' its thermoelectric line were that it lav helow nickel on the
diagram, and that its inclination to tlie lead line was mnch u'reater
than the inclinations of the iron and nickel lines.

As a Lahoratorv exercise I gave to ]\Ir. Sawada, one of onr stu-
dents of physics, the task of studyin!^ tlie thermoelectric properties of
the sheet cohalt descrihed in tlie precedinii; note. The plan ado])tcd
was to f'irm a nudtiple arc of Palladium and Pismuth and ])v ])roper
adjustment of tlie resistances in these l)ranclies to ohtain an inter-
mediate line which should cut thrcnig-h the cohalt line at tempérai ures
within easy reach.

Such an iiitermediate line passes tlu'ougli the neutral point of the
component metals. It di\ides the region hetween Iheir lines so that
any transversal is cut into ])(^rtions which are directly as the resistances
in the Ijranches of the nndti|)le arc. Thus hy varying the ratio of the
resistances in tliese hranches we may sweep through the region hetween
the two coi'responding diagram lines, interpolating so to speak any
iiitermediate line suitahle lor our pur])ose. The extreme accuracy with
which we can measure elc'tric resistance enahles us to fix the position
of this intermediate line a.s a.'-curately as the positions of the component
lines are known.

The low jiosition of cohalt on the diagram very much circum-
scrihed the choice of metals for the multiple arc. lîisnuith had to he
one of them, as it alone was known to he heh^w cohalt. The other

296 C. Cx. KNOTT.

meta] fixed ii])oii was ralladiiiin. a snhstnnce convenient in every
way. Its (liauTani line is strai^'lit u]) to liii^ii tein])eratnres ; aiul its
character does not perceptil)ly clumge even after severe heating-s. Un-
forlunatclv, however, the necessity of nsin,o- ])ismnth ]iniit('(I tlie
investigation t(^ moderate temperatnres.

The l)isninth was l)roken np into small pieces, which were packed
tightly into the l)ore of a si]>hon sliaped glass tube. Gentle heating
in 11 Ihmsen flame sufficed to melt the metal, which ran together and
solidified (^n cooling into a fiirly unifnan rod. The junction wires
were fused into the ends of the hisnmth rod.

As tinallv set up, the a])paratus consisted of a triple Cohalt-
Bismuth-Palh'.dium junrtion dippirig in oil. This formed the "hot
junction." liesi;4ari<'e boxes were included in the p:dla(ham and
bisnuith bi-andies. because of the magnitude of the thermoelectro-
motive i\mv:' between these three metals and cojiper, great prec:nition
was necessary in keeping tlie various cold junctions at the same tem-
perature.

The palladium branch always contained 100 ohms resistance.
The resistarir-c of the l)i:anuth Irninch varied from infinity to 200,
lower values c:u"rvirig the intermedia.te line too fu" below the coltalt
line. For each of tlie seven selected ratios of resistances, a, careful
series of thermoelectric observations was made. A delicate high
resistance galvanometer was used ; and the temperatures were mea-
sured bv a mei-(au'ial thermometer. The electromotive forces between
the cobalt and eadi intca'inediate "equivalent metal" were in this
way measured directlv. From these the thermoelectric powers at
chosen ttanpci'aiures could be calcul-ated. lîut one of tliese " e(pii-
valent metals" was ])alladiiim itself, when tlie re:;istance in the bisnuith
branch ^vas made intiiiite. Siil)tracting 'dl llie oilier tluaanoeleetric
powers from this one, we obtained the thermoelectric powers lietween

LABORATORY NOTES.

297

palladimn aiid llic other equivaJciit iiicta].<. The Ya]iie.> of the
therinoelectric powers "were eaJculated for 0° C aiid 100'" C and are
given ill the followiiii:- Table. The symho] Pi staudo for ijioinutli, Co
for cobalt, aiid Pd for palladiuiii, Tlie variou;> " etjiiivaleiit iiietafs "
are repre.seuled by the ;<yiiil)ol Pd Pi „ where tlie ijiindier n represent«
the ratio of the reniistaiice in Um l)i.;niiilh ])raueh to the refdstanee in
the pa]Ia(b"iini branch. Tdui;^ Pd JJio means that, iince the palladium
ahvay.r; contained 100 ohms resistance, the l)is]nuth contained in this
case 200 ohms. The electromotive forces are measm-ed in microvolts.

Thermoelectric Powers referred to Palla.dium.

Metal.

'J'lieruioelectric Power
at O'C. at 10(/0.

NeiiLial Point
with Cobalt.

Co

7.00

17.ol

Pd Bi,3

5.98

(3.4(3

-10°.4C.

Pd Big

9.;38

OM

+ 2P.5

Pd l^i.

] 1.45

\ 4.(i0

74.1

Pd Bi,

17.44

17.4i

101 .4

Pd Bi,,

21.73

22. 1 o

1 18 .0

Pd Bi,

20.10

20.55

224 .0

Bi

8(3.0

88.8

The numljers in the last row liave heeu calcidated from the
numbers in all the six Pd Pi rcjws. Tor if p is the thermoelectric
power between Pd and Pi and p,i the same Ijetween l*d and Pd Pi,j we
know that

P - Ih

p,

or

p = {ll + 1) pn

Thus from Ihe six sets of values corresponding- to p^ we obtain the
following values for p at ifC. and lOO^'C.

298

C. G. KNOTT.

n + 1

1\

Pm

14

83.7

90.4

9

81-. 4

89.0

6

80. 7

88.1

5

87.2

87 2

4

80.9

88.0

3

87.3

88.7

Means

80.Ü ± .8

88.8 + .7

'riii.< t-able i.^ uljviour^ly an indication of the accuracy of the
experiment.

And now, referrin.^; everything to the Lead line, and expre.^>ing
the thermoelectric power in the form

P

^ = Ä + Bt
dt

we ohtain for tlie coetHcientiJ .4 and P» tlie foUowinu' V'due:s.

A B X 10'

Lead 0 0

Palladium - 0.18 - 3.55

Cobalt -13.18 -13. 9

Bismuth -92. 2 - 6. 4

According to the nundjer.s deduced by Fleeming Jenkin from
Mattlde.sen'y experiments, Ijismuth lie^ four times further from lead than
does cobalt. Here we have it seven times. Professor Tait's electroly-
tically deposited C(3balt lies 4 J times further from lead than does
p:d Indium. Here we have it a little over tAvo times. According to
Lecquerel's nundjers given at the end of the English translation of
Mascîut and Jcndjert's FJcciriciiij and ^Liguelism, the rati(3 at 50° C.
of the thermoelectric powers of palladium and bismuth relatively to
lead is as 7 : 40. Here we have 1:16.

LABORATOEY NOTES. 299

These di.scre});iricies nve not .snrpri.siiio-. We know how varia1)Ie
are tlie thermoelectric properties of .stähle alloy.s^^ intended t(_) ha^•e tlie
.same composition, and h(3w a very sh<^-ht eliange in com[)0.sition may
he accompanied hy a very large change in thermoelectric position. The
present experiments mnst therefore he judged of •dt(_)gether on their
own merits. A simple comparison shows us that Professor Tuit's c()l)alt
will fit in to the regi(3n hetween lead and l^i.smatli ^•ery much as
Matthiessen's cohaJt fits in io his own .series. Thus tlie cohalt inves-
tigated liere seems to ditter from tliese other specimens in much tlie
aame wa3\ The new col)alt indeed hes so high in tlie diagnim that
its line is higher than the line of Ta it'. s nickel, for which A = — 21.S.

This unexpected result was at once tested. A rougli experiment
was made witli the couple nickel-cohalt and a neutral point was oh-
tained a.t a temperature below 100° C. This cohalt line therefore, at
ordinary temperatures of the air, is ahove nickel ; l)ut because of its
greater downward inclination 'j:et:^ below it at temperatures above

100° c.

As regard.-; the inclination <jf the coljalt line, the present result
agrees as well with the earlier result as ci^uld ]'easonal)ly Ije expected
with two quite dillerent specimen;' of the metal. I'hus, expressed in
tlie same units, the thermoelectric power ;/f Professor Tait's coljalt is
given by the formula.

p=—26:S — i).\ Mi /
Avhereas for the present specimen

2j=-13.f8-U.lo86 t
\\'ith tlie exception of the .sharp upward Ijend in ni'^kcL ihi.- gives the
greatest inclination yet ohtained for a thermoelectric line. It would

1) See tlie pnpei- by MacGre^'or aud myself already referred to, also my paper ou T/«'
Electrical Proi)crtu'^ of Uydrogcnisicd Falladinia (Traus. \l. S. E., Vol. XXXIII., 18SGJ— abstract
in this Journal, Vol. I.

300 C. G. KNOTT.

be interesting to establish Ijy direct ex})erinient that the Thomson
Effect is exception;! ]ly laru'c in cobalt.

The downward trend and cDinparatively large inclination (^f the
bismnth line are also worthy of note. Uecanse of the ])osition of the
line as a whc^le, lying far below tlie lines of all other metals, this
large inclination does not greatly influence the electromotive fjrces, so
that with bismuth C(juples the electromoti^'e force is very approxi-
mately proportional to the temperature. This fact of course prevents
us from making a very accurate determination of the coefficient 15,
which in the present experiments has a large probable error. Its
mean value is a little greater than the ^•alue indicated in liattelli's
direct measurement of the Thomson Effect in liismuth^'.

liio'hi has shown-^ that the electric resistance of lUsmuth is altered
in a strcjiig magnetic field. To find if any thermoelectric change
accompanied magnetisation in nickel, a bismuth palladium cou])le was
set up IjetAveen the poles of a p(3werful electromagnet. Xo effect
whatever was obtained, although the arrangement (slightly modified)
was sensitive enough to show with great ease tlie thermomagnetic
effect discovered by v. Ettingshausen and Xernst^\

1) See Wied. Beiblätter, Vol. XI, 1887.

2) See Wied. Beiblätter, Vol. XIII, 1884.

3) See Wied. Annaleu, Vol. XXIX, 1886.

Diffraction Phenomena produced by an Aperture
on a Curved Surface.

By

H. Nagaoka.

In ordinnry prohlems on diffraction of liglit produced liy nper-
tnres of varions shapes, the diffracting- apertures are snp]:)Osed to lie on
a plane. The rm^re _£i'eneral prol:)Jeni of diffi-action prodiKXMl 1)y aper-
tures on a known _o-eonietrical surface lias noi, so fir, beeri touched.
It has been my object to fill in this gap, althc^ugh the expression for
the intensity of diffracted liglit is integrable oidy in a few ]X!rticuI:u'
cases.

In the following, I give a general expression ibr the intensitv of
light diffracted by nn aperture on a known surficc, ])otli inr Frauu-
lioi'cr's and Fresnel's diffi'action ])lienonu'iia. 'i'he expression is tlieii
a])})lied to find tlie (b'stribution of light after its ])assage tliroiigh a
small slit cut ]X'r])endi*Md;ir to the generating line of a right circiifir
cylinder.

Expression for the Intensity of the Diffracted Light. '•

Let L be a source of light, and A B an apertinx* on a known
geometrical surfice. Llie ray of liglit pro])agnted irt^ii L i- dilfracteil
by the aperture A B, and the diffraction plicnonK'na thu- ])i-oduced
may be seen either projected on a screen at 1) (^Fig". 2 ), (^r o1)ser\('(l by

* Tn the deduction of the expression for the intensity, T follow F. Xeuniann's method.

302

H. NAGAOKA.

Fin-. 1.

Fig. 2.

means of n telesoope placer! at T, and s(^ fc^ciissed that the ohserver
sees a, distant ])()int ]J (Fio-. 1) ; in other words, D is the so-caDed
diffraction ])oint.

In order to find the a'eneral expression for tlie intensity of light
after it is diffracted hy an a])ertnre on a cnrved siirfjice, T shall assnme
tliat the diffracting apertm-e is very small compared with its distance
from the source of light, and from the ])oint at which the intensity of
diffracted light is ciHisidered. Conserjuently the amplitude of N'ihraiion
of the hght coming from different ])oints c^f the aperture will not ^ ary
at the ]ioint considered.

I shall first discuss Fraunhofer's (telescopic) diffraction pheno-
mena. Referring to Fig. 1, let the vihration at L he represented hy
eus 2r ^r ; tlien, at any ])oint P on tlie diffracting a])ertiu'e, it will
he proportional to

No\Y considering the ray in the directi(on DP, tlie vihration at anv \)ohi\
C in the line FT due to tlic small element da at /' is propoi-tional t<»

ox CEKTAIX Dli'FKACTJOX l'UEXOMEXA. 303

/ t LP I'C \ ., _

\r / /, /

Describe a sphere with D as centre, and jtassiu^- throuuii C ; tlieii the
time taken ]jy the ray U) go from the s|)]iericaJ surface to the eye will
he constant, provided D be siiüicientJy distant. I.et (his constant time
be denoted by " ; the vibration at T is thus ])roportional to

an cos [ — TT, — ^ — ^- J 2 -.

\ similar expressitjn holds for the light ])ropagated from every element
of tlie aperture, so that the total effect at T will be given l)y the
inteii'ral

(1

I <((7 COS I rn — ^ — — — : — ) 2 /T,

./ \ i / / /

where the integ'raiion extends (jver the whole aperture.
Taking any point 0 near the :iperture, we may write

PC = DC - DP,

= DC + {DO - DP) - DO,
LP = LO - {LO - LP).

Denoting the c(jnst;int distances LO, DU hy U and IÎ' res[)ec-
tively, let LO - LP -- JPi, and DO - DP = JPi!

Introducing these syndjuls in the expressions for PC and LP,
we find

PC - DC - R -{- JR\
LP = R - JR.

Substituting tliese in (1), we get for the vibration at T the integral

,os r , n ~ - DC R- W JR - JR' \ ^
(2) J da COS {-^^ . ^ + . )2..

Since r, DC, R — R' are all constant, we can put

t - T DC R -- R'
-T J 1 ='^'

304 H. XAGAOKA.

and the above expression for the vibration becomes

( 3 ) Ida cos ( b -I ^ J 2-.

The intensity of Hght at T is, therefore, given Ijy the expression

or more simply by

( I ) 1= MocUfda e ' ^ (-J^- ^«')

Wlien the diffraction point is sitnated on the (_)ther side of the
surface from the source of light, and the phenomenon is seen projected
on a screen at I), ^ve must slightly modify the expression for the in-
tensity of light.

Troceeding in exactly the same way as Ijefbre, the vibration at
I) due to a small element da at P ("Fig. 2j, will he proportional to

, / t LO PV\ ,,
d(T cos {^Y^ J —\ 2-,

which can be written

/t _ LP , LO-LP OD-PD\^_

\T ~r^ I + I )-

da cos

Puttinii; as before

LO - LP == JR,
OD - CD = JW,

we get

/t LP PD\,, /, . Jli + JP'\,

Conse(piently, the total effect at I) is gi\'en by
I da cos la-] ^ j 2-.

ON CERTAIN DIFFRACTION PHENOMENA. 305

Thu.s, the intensity (A' diffracted light at D is given by

or uioYd briefly by

(II) I = MocPfdn c ' ^ (^^^ + ^^')-

Tlie abo\(' expressiijn gives tlie intensity of diffracted light for
Fresnel's diffraction phenoinena.

To evaluate the integrals given in (I) and (II), assinne 0 as the
origin oï three rectangular c<3-ordinate axes x, ij, z. Let the coordi-
nates (jf the points L, 1>, i^ referred to these axes be denoted thus : —

L : a, h, c,
D : a , b', c,

P : X, y, z,

and let the ecjuation of the surface referred t(.) the same axes be

F (x, y, iij —const.
Thus, we have

LO = ^ a^ + h- + c' = II,

OD = ^ a'+ 6"-+ c- = W,

LP ^ V (^» - ^f + (^ - Vf+ (^- - =2)' = 1^ - ^^^^

PD = ^ (rt'- xf^- {b'- yy'-{- {c'- y.)- = W - âli\

Expanding the expressions for LP and PD Ijy means of the binomial
theorem, we have

LP ^ E j^ - -^yfj^iax -Vby ■]r cz)--\- ^, + ,

T„ ax + b'y + c'z 1 , , ,, , ,., ./;-+?/-+ 12'" .

PD = K ^^-jf^ ^j^a^«'^+^^+^'^^)''+ ^-^4^^ — +■

or

306 U. NAGAOKA.

âR = ^ + -^-^ {ax + by + r.^)- .,7^

.j^, a'x + b'y-^c z , 1 , / , ?' , / n- à:^+ (/'+ f-'^

-i)

Let the direction cosines «jf OL be }c, u., v, ;in(l those oï OD be
;/, /^', y' ; then ( 4 ) becomes

i')

J22 = (;r a: + //. y + i^z) + -:^-rr {" x + n y + •. z)-— j' — -,

^ it 1 i\

1 / / / /so ^" + ?/ + •'^'

M/ï'= (;/a; + u-'y + v's) + -^-jr, {nx + //.'^ + v'a)- —

2ii;'

In Fraunhofer's diitraction phenomena, // and B! are supposed to l^e
very large compared with a;, y z, so tli:it we can neglect the terms
containing B, <3r Pi in the denominator. 'Huis

âli - âR = {)( - ;/) X + {/J. - //) y + 'c^ - v') ?:.

Writing

0 -

- ^') = n,

the expression for the intensity of the dittVacted Jiglit becomes
(T) I = Mod-.fdae'^'-'"-'"^-""^

where the integration extends o\er the wh(jle aperture.

In Fresnel's diffraction phenomena, we can no Jonu'er neuiect the

terms -5- and -^. Thus tlie expression f(.)r AR + âR' ljec(3mes very

complicated. It is, however, somewhat simphfied Ijy taking 0 in the
Une LD as shown in Fig. 2. Thereby ;/= — )(, //= — ti, >= — v,
because OL and OD are in one Une. Tlius

ON CERTAIN DTPFEACTION PHENOMENA. 807

AR + AR = ^ (;. .r + fUJ + i^zy-- {X'-+ j/ + ..2)^^_L_ ^ _I_\

Introdncino- fin's v.-iliic in (TT), we o'et for the iutensity of h'oht ;it D
( 11') / = MocP. fda .^^[(^^' ^ "'-^ + ''^^' - (^' -^ y^ + ■^^)](-zr + -7p)^

where the inteo-rntion extends over the whole npertnre.

Thus the pro1)leiii of the ditiVnctioii of lio-ht produced hy nn
n]ierture on n curved suriace is reduced to the inte^Tntion of expres-
sions (I') nnd (TT') for Fraunhoter's and Fresnel's diffraction phenomena
respectively.

Fraunhofer's Diffraclion Phenomena produced by a narrow
Slit on a cylindrical Surface.

Let us now discuss Fraunhofer's diffraction phenomena produced
hy a narrow slit cut on a right circular cylinder and perpendicular to
the generating line of the cylinder.

Tn order to calculate the intensity of liglit for different positions
of the telescope, dro]i a perpendicular on the axis of the cylinder
from the centre of the slit. Assume the centre as the orio-in of
co-ordinate axes. Fet the x axis he parallel to the axis of the cylinder,
and the z axis perpendicular thereto, hoth drawn through the centre
of the slit.

The axes heing thus fixed, we have, 1)y ( T'), to find the integral

/•

^^^^i(U + viy + nz)

where the integration extends over the whole ajierture, and /, vi,
n are determined ])y tlie (hrections of the incident light and of the
<)T)servirig telesco])e referred to the rectangular axes ahove specified,

308

H. XAGAOKA.

and by the wnve leno-th of light employed in tlie oh^^ervntion. In
addition to this, there is the equation of condition

ii/2 + «2 _ 2 az -=0

expressing the faet that tlie aperture lies on a cylinder of radius a.

In actual calculation, it is more convenient to use polar co-
ordinates. In the right circular section of the cylinder, assume polar
co-ordinates with the pole on the axis, and take

y — a sin /? , z = a {1— cos ß),

Then da = a dx da.

Thus fß^'-""y-^'"'\h = ae'^'jdx e'Y^//.'"('" "■" ^ ^ " '-'' ^^■

where 2/> denotes the hreachli of the slit.
The integral

idx e =

Ü

2 sin Ih

I
It thus remains to find the integral

taken between proper limits.

Introduce an auxihary angle â', such that

a m — 5 si)i (Y , a n = c cos />'

where c = a ^ni- + n'.

ox CERTAIN DIFFRACTION TUEXOMEXA. 309

riicii a {in sill 'V — n cos â) = ç cos [r) + I'J') = ç cos c,

where ç stands i'or ij + />'.

Thus fdn e'"^'" "'" ^ - " ''' ^' ^ fdif e' ' '"' '^•

'J'lie Jiiiiit.s of integral ion with respect to (f are fljiind ir«Mu ij' and
the known limils witli respect to ü.

The dithcuJl y of the })ro1j]eni hes simply in fin<linu- the inteu-ral

I shall hencef(.)i-th put J ~-= K + iL, where

K — I cos (-f COS (f) d(f,

L =- /sill (ç cos (f) dç.

Evaluation of the Integral '^ = fif c ^ ''' f^

There are various ways of evaluating tlie ahove integral. The
siin}»lest way would l)e lo lind a differential e(jualion which is satisfied
hy J, and hy this means to expand it in a scries proceedinuf accordin"'
to ascending powci*« of ^.

îSincc every integral of tlie form fdire^'""^ hctween known
limits can be decomposed into a smn of two separate integrals of the

/•a ■ .- , ,

form id(r c ^ "'" '^ , I sliall (jnly consider

•/o

*ultinu'

cos ç — U, cos (J. — c,
J = - / 4zz=rr- du.

■h ^\ - i(^

310 U. NAGAOKA.

Ditfcrciitiatiug with respect to ç, wc have

_ J ] (I J i .s cos (<• E) !^ sin (r ç)

Ç d Ç Ç Ç

wliere .s staiids for ^l — c^ = sm a.

Tliu.s tlie diitereiitial c'ljiiations ^^atisßc(l l)y A' ami L are res[)ect!vely

d' K 1 d, K ,, s sill, fc c)

a Ç Ç a Ç Ç

1 d' Tj 1 (Z L ^ s cos (f ci
and -^^^ + T -. \- L = _'

To find the expression for K a.nd L, assume a series proeeediii_ij,'
aceonhuy to ascendin,^' powers of |. Diffrreiitiating and pro[)er]y
chootfiiiL!' tlie eoiistauts, we easily Hud that

+ ^ Vi-

J' / r" (•■' c- 1 \

+

irs II + 1 /■ ,,2 II ..5 II - 2

]

where n stands for — (cs + a).

»)X CERTAIX DTFFRAL'TIoX I'lIEXOiMEXA. 311

It is to he remarked th:it when a = r, 7v beconios eqii:il to - J°(c),
wliore .7° denotes Ijessel'.s function of tlie Hrst kind witli index 0.
Tliiis the .'d)ove expression for Jv reduces to

K

_ / Ç- Ç' ç*' ç*^ ' \

Tlie expression within the hracket is the well-known form for -7°(ç).

Tt is easily seen that the above two series f )r 7v and L convero-e
ra]>idly so lon^' as ^ is small ; hut when ç becomes lari^-e, it wonld be
advaiitau'cons to em])loy other expres.^ions fn* K and L.

The usual process of calrulatini^- / e ^ ^"* "^ dç is to ex|)and c'' ^"' ^

in a Fourier series, and integrate eac-h term of tlie series sejiaralcly.
Thus

1

• A) ^

1

E(piatinr;- the real and imau-imu'V ])arts to K and Tj respectively,
we ha^•e

(c) A' = o..!" (,-) + 2 V (- 1)" r-- (ç

2 «

The f(^rm li'iven above is not r.a])idlv C(^nveru'ent. The values of
J"(ç) can l)e easily c:d(adate(l from the values of J°(^) and 'P{^) g'iven
in the tables of Hansen and Meissel u]) to certain values of the argai-
ment ^. \\\\\ for hig-hcr values of ç, we should have to calculate -/^"(c)
and 'J\:). ^loreover, when n exceeds ç, the value of '/"(ç) deduced
successively from J°(^) and -/^(ç) becomes inaccnrate, and we are thns
compelled to undertake the calculation se])arat('ly. Tliese considéra-

312

H. XAGAOKA.

lions îiinkc tlic foniiiihic jiist Lj'ivcri less r-onvcuioiit for (■.•ilciihilion
tliîiii llic i(<i'iniil:i(' ü'i\('ii lu'low.

As :ili(':i(lv iiUMilioiicd, tlic Inrin of tlio intOL'Tal /sliows \]\:\t wlioii
llie liniirs lie iVoiii o to -, il Ik'couk,'.-! ('(|ii;il lo - J °(ç). Tiiiis J
iiirliidc:-! IV'ssel's i'iinctioii oi'ilic llr.u kind willi index 0 :im :i |):ii'1 i('id:ir
case. l>v a s])f'(:'i:d transiocnialion, J can Ix' inadc to d('|)('nd on
J° ()iz) as will now 1k' shown.
Tuttinii' Il ^ cos if , wc \v.\\v

nu
— I dii.

J

Expanding'

V^ I — \C-

:zir m a r Olivier series.

1

1 "^

^ [ — li-
]\Inlti])l_vinii' tin's l)y ^' ^ ", and intea'i'atin^-

After a simple reduction, "we liave

/:

ju - —

?' ./?«/ 1

Ivjiiatino- the real and imaginary ])arts of both sides of tlie

equation

(0 /:^^*,, = -Lj

V

z h - ~ sin ^ M >

1

Ç- — n- --

2 r ro.s ç ?/, V

^nJ° [n ~) sin [n - ii]

Ç- — n- --

I

ox OEUTAIX DIFFl.'.VC'rioX I'llEXoMEXA. -U.')

/ ,.x rf<in (: ii) , 1 l' ros ^ II . . ^^ J°{ii z) cos n z u

+ 2 - sin (. .) 2^ .._,,.-. |-

Tlieise two cxj^vssioiis ( (• ) nnd ( /') nrc v(\vva\ to — /v •iinl — Tj

Tims 7v and L are iiiadt' to dcju'iid oii J°(», r), wliidi can 1)0
('al<adated o\\(\' loi" all ; tlic rest involvini;- siiiiplc aritliiiU'(i<-al and
triti'oiK/iiU'trical calculations.

The ex])rossions ( ^' ) and ( /') above dcdncod for K and Jj ro<|uiro
special consideration M'licn ^ is a multiple of -, since both then contain
terins of the form ---.

Let us suppose that | ^^ m - ^^ y. Then the expressions for K
and L assume f)llo\vini'" f nans.

wl lere

1 I sin zu. ^ _ . . "^\^J [n-^cosin-ii) ^ , . . ^ J {n-)cos{nzu)

. I -_ \--' sin C II. > —. :,—T, \--ZSUlZUy ;;; t, 7,

'1 V^ z V -- — n--- ^, c- — n- ~-

1 ' mrl

-, ^ '"■C^^nJ''{n■ z)sin(nz ii) ^ ,^ ^ ^nJ°xn -)sin{n-n.)

— 1 - cos C ?( > X., .,- ., ~ COS iC 11.) > i-.r ^-9~ .,

j^' _ J. ^ sin (? u) cos (y ii.) — y cos {: u) sin (y u)
"?=>' Ç- - r- •

^ 11 cos z V. . „^ . i-^ 'T in-)cos(7i-n.] ... . -P5 J (/?.-) ros ;*-'

ij = — i-l :: + 2 C COS C Ii > —, 7, — 7, |--"COSC« > —, —r, — 7,

2 l_ Ç ^ Z-— n- -- ' ^j Ç-— w- --

, -, ... ,r,l7°(72n-) st/i (/i;r?«") , -, . , .^^ J° (nz) sin iiz'
+ 2z sin':ii)y — ^ — ~, — -^^ \- 2 :: suiji z iny -. 7^—7,

+ 2 7/J°0//-)T

?)14

H. NAGAÜKA.

where

/^i, u., (]cuo\\u<j^ the limits of intei^Tntion witli resper-t to //.

Ev:iln:iiiii£;' tliese two iiidetenniiiatc forin.-; K' •nul L', we find

K

2 ;- n. + sin (2 ;- ii)

L' =

sin y (?fi + 71.,) sin y {il., — iii)

Thn.-! tlu' expressions nhove deduced cnn 1)e employed fnv cnleulnting'
A' ;iiid L ïifV nil values of ç.

I may here remark, tliouo-h it has nothing' to dn with theqnestion
of ditfra''tion, that a m<^re gvneral integ'ral of the firm

/'', i ? ens (f . 2 V
d(f e ^ sui (f,

the limits lying' between - and —z can he made to depend on ./"(nr),
hy exactly the same process as al)ove given. Tn fu-t, r^essefs function
of tlie first kind with index v can he expressed hy means of the
f )llowing f )rmula,

J\.2.ö...{-2v-l) yi? sinç 2çsinç^ n + l J^jn::) \

ox CERTAIN DIFFUACTIOX PHENOMEXA.

315

For coiivciiieiice of calculation, the following- values of J°(/tr), for
successive values of n, have ])eeu calculated aud tabulated.

n

J°(«r)

log J°(«-)

n

J\n-)

log .r{>i-)

1

- 0. 3U4242

(-) 1.483219

26

+ 0.062329

( + ) 2. 791691

2

+ 0.220277

( + ) 1.342969

: 27

-0.061168

(-)â. 786523

3

-0. 181212

(-) 1.258186

28

+ 0.060069

( + )î^. 778650

4

+ 0. 157507

( + ) I. 197300

1 29

-0.059027

(-) 2. 771051

5

-0.141182

(-) r. 149779

30

+ 0.058038

( + ) 2. 763710

0

+ 0. 120064

( + ) T. 110804

31

-0.057096

(-) 2. 756608

7

-0.110609

(_) T. 077765

32

+ 0.056199

( + ) 2. 749732

8

+ 0. 111068

( + ) 1.049093

33

-0.055343

(-) 2. 713066

9

-0. 105625

(-)T. 023768

31

+ 0.054525

( + ) 2. 736599

10

+ 0. 100251

(+) 1.001089^

35

-0.053743

(-)2. 730320

11

-0.095621

(-) 2. 98U555

36

+ 0.052993

( + ) 2. 724216

12

+ 0. 091579

( + ) 2.961796

37

-0.052273

(-) 2. 718280

13

-0.088010

(-) 2.944530

38

+ 0.051582

( + ) 2. 712501

11

+ 0.034827

( + ) 2.928535

39

:

-0. 0509] 8

(-)î^. 706872

15

-0. 081967

(-) 2. 913638

40

+ 0.050279

^ + ) 2. 701386

16

+ 0.079378

( + ) 2.899697

41

-0. 049663

(-) 2. 696035

17

-0.077019

(-) 2. 886597

42

+ 0.049070

( + ) 2.690812 ,

18

+ 0.071859

( + ) 2.874243

43

-0.048497

(-) 2. 685712

19

-0.072871

(-) 2.862554

44

+ 0. 017944

( + ) 2 680729

20

+ 0.071033

( + ) 2.851462

45

-0.047109

(-) 2. 675858

21

-0. or)9328

(-) 2.810910

46

+ 0.016892

( + ) 2.671094

22

+ 0. 06)7740

( + ) 2.830846

47

-0.046391

(-) 2.666432

23

-0.066-257

(-) 2.821228

48

+ 0.015906

( + ) 2.661868

24

+ 0.064863

(+) 2. 812017

49

-0.015136 j

(-) 2.657398

25

-0. 063560

(-) 2.803183

5j

+ 0.011980

( + ) 2. 653018

ol6 Ü. XAGAOKA.

Jieturiiiiii,'' io our [H'oblciii on FrüUiihofer's ditfrnction phcnoinena,
we iz'tit for the exiR'ession of the iuteiisitv

I — 4 a- — ß — (A-+ L-)

With a hoinoo'eneouy tsonrce of Jiu'ht, the iiiteiisity always vanishes
whenever I h is a nuiJtipJe of -. The fringes arising i'roni the term
sill' lb are exactly the same as those given hy the ])lane slit. When
the surface on which the slit is cut is cylindrical, the additional factor
K- + L'' enters into the expression for the intensity of the ditfracted
li^dit. This factor has maxima and minimu f )r ditierent positions of
tlic telescope, and moreover depends on the length of the slit. Thus,
wlien the Hmits (jf integration lie from 0 to -, K -■= - J°{^) and L = 0,
and there would be places of darkness for .such positi(His of the teles-
C(3pe as are determined by the values of ç corresponding to the roots;
uf /■>(?).

For a great numi)er of ecpiidistant slits, the expression for the
intensity would be tlie same as that for ordinary grating, multiplied
by the factor K- + L'\

The case which calls f)r sjjccial attention is when the ray
is normally incident, and tlie telescope turned so as always to
lie in the plane xij. Tlicn ;c = 0, //. := 0, v = 1, and // = 0. Thus
I = ^^4^ sin. (0, where w is the angle made by the axis of the telescope
with z axis. The places of darkness are given by

n /

Sm (I) =- — r • — r-

2 6

The maxinra and minima arising from the term K- + L" must be
separately determined f )r the ])arlicular slit in question.

ox C'EKTAIX DIFFKACTTOX PHEXÛMEXA.

317

Fresnel's Diffraction Phenomena produced by a Slit
on a Cylindrical Surface.

Fiff. 3.

Let A B he n section of
tlie slit, cut by ;i plane passino-
throiio'h the sonrce of lig-lit L,
and the point D at which the
illumination is recpiired. I shall
-X suppose that the point D is not
YQvy far from tlie line joining-
any ])oint on the slit with
the source of light. Also, the
problem will be still further

simplified if the plane L A B is made to contain the axis of the

cylinder.

For calculating the intensity of the ditfracted Hg-ht, assume tlie

point 0 wliere L I) meets the cylindric:d surfice as tlie origin of

coordinates. Let the x axis be parallel to the axis of tlie cylinder,

and y perpendicular to the plane LAB.

In this case,
where d is very small,

a = 0,

Thus

H = cos LOA = Ö,

and

1 nearly

{}( or- + /j.y 4- i.z)-= 2 Ö x s + :2^

neglecting d' u]i wards.

Recurring to formuhi (II).

A 11 + A R= Q ÖXZ- {x'Jr ?/)] (^ + ^).

'>1-'^ H. XACJAOKA.

Tlu'VcCorc, l)_v lorimil;i (11').
( T ) / = ModJ a fh .' 1 [- '^ •- - (-^ + ?/')](!+ -}p)-

Sinrr ''A .r, üiid ,: nro nil very .-;mn11, ^\c rrin writo

wliovo c stiiiids for -4- ( -7— + ,,/ Y

'l\'ikiiiu' ])()l:!r coordiiKites in llio riulif circuLir ^ediDH (.>(' llie
(■yliii(l(i- willi the ])n\r on tlio axis, we in:iv writ«'

d(T = a dx d(f,

y = a sin c, z -- a {] —cos ç).

Trdrodnrina' Ihoso expressions in (1), we o-ot Cor 1lie intensily oC Ilic
difiV:ieled rny

( 2 ) Z = Mod.' a fj'dx dif e' ^ ^'' + ""' •''""' *^\ 1 + / 2 rr /> ,- ,r (1 - cos ^')]

Tn intcgrnting tlie above expression with resper-t to rr, we mnsl (]is-
tingnish two eases according as D lies williin or witliout Ilic
geometrica I shadow.

Let 0/1=^5, and A B = h ; tlien the integratimi willi resj^ect
to X ninst extend, vvlien D lies within tlie geoinetrical shadow, from

ß to ß + h.

When D is onlside tlie geometrical shadow, tlie limits ot" integratioii
must he trom

— ß to h — ß.

The iritegration with respect t(T tr mnst extend oyer the wliolo
lenath oi' the slit.

ON CERTAIN DIFl'KACTiON PHENOIVIENA. oU)

I shall tirsl pci-l(jrm llie intégrât ion with rcspccl to ^^.
We can write

/* i 6 a' s/«"- (p , /■ / ^ (1 - t-us 2 (f ) , I / ^' /• - / ^' füÄ '2(f> , ,-, ,

/e ^ay^ = je - dç = —p— c - /c ^ a (I if).

The intégral thus ohtained corresponds to /, Avhicli was already
investifTîited in connection with Fraunhofer's (htîVaction ])henoniena.
I sh:d], therefore, write ibr simplicity

(3) p^''''"'''^dir =.K + iU

Again

/e ^ " ''"' *^ cos (f d(f = — J -\ cos {a^T~ sin (f)- d {ii ^yi~siii (f)

+ i I sill {a -y/ Ç sin if)- d {a y/ 5 sin <.ç) I.

l')ii( /'^'^ ((t a/1~ -s-i/i cr)-(^ (a^/ysmcr) are derivahle in terms of Frcsnel's
J sin

integrals, for whicli the series oljtained by Knochenhaucr, <iilbcrt,
Cauchy, or Lommel can be used for calculation. 1 shall, ihereibre,

})Ut

( i ) e ^ cos (f dif ^ I + I 1

Xext [)eriorming the integration with respect to x, we liave to iind

le ^ '*" dx and c ^ ""^ x dx.

The ürst is an ordinary Fresnel integral ; and can, therlbre, be
written

( 5 ) p ^ ""' dx = C + i S,

wliere

320 H. NAGAOKA.

C= —j^lcos {^T xy- d (\/ ? x), S = lain (^T x)- d {^T x).

The second iuteural is intei'TaljIe : thus,

= j' + i a.

Introducing the expressions ( 3 ) (4) (5) (0) in ( 1' ), we lind inr
the intensity,

/ = Mod- a (C + i S) {K +i L) + i2 a '/ : {y +i^)\ [K + i L) - f / '+ 1 2') [ .

In hndin,u" Mod.'-, we can neglect tlie terms in\<)Iving />"'.
J'husj we get for the expression of the intensity

(7) I ^ir^[C- + S-}{K'+ L-) + iafl^\C[Pr-Qr7) + S{(Jy + Pn)l'l^.

where

r - A' 1' - L 1 \

(J = K [K - /') + 7. (L - 2').

Tlie expression f<jr the intensity of light dittVacted hv a A\l on a
circular cylinder ditfers from that for the ])lane slit hy the intro(hiction
of the fictor K- + L-, and a small athlitional term multiplied Ijy ''K
lîoth K and L remain constant proyided the distances of the slit from
the source of hght and the point at Avhich the intensity is recpiired do
not change. If we observe the fringes in a plane parallel to the axis
of the cylinder, /v and L will remain scnsihly constant. Aeglecting
the term multiplied hy '''', the positions of maxima and minima wiJl
be the same :is those pnxbiccd l)y ])lane slit of the same l)readth.

If the observer approaches or recedes from tlie sht, the intensity
of light at u point directly opposite the slit will differ from that of
the plane slit, for the intensity is affected Ijy the fictor K- + L'\ wliich
is ijo lonu'er constant.

ox CEUTAIX DIFFKAC'I'ION PHENOMENA.

821

Ob.servation shews that the small additional term is of very small
eÜ'ect. Calculating the minima of C' + S' by means of Knochenhauer's
series, I find that the agreement of caleiilation with observation is
quite close, except when the point considered lies outside the
geometrical shadow.

In order to test the result of calculation with observation, the
following experiments were made with a slit of 90^ aperture, cut on
a right circuit) r cylinder of 5.0 nun. i-adius. Sunlight was admitted
into a darkened room. After passing tlirough a small vertical slit,
and a lens, it was analysed by a prism. The spectrum thus formed
was projected on the slit of a spectrometer. The slit, howcNer, was
cl(jsed by thick paper, and only a small liole was pierced, throng] i
which light was passed to the slit under examination. The spectriun.
was so distinctly formed, that one could easily make the light c<jrres-
ponding to any one of the principal Frauidiofer's line illuminate the
slit. The following oljservations were made for the positions of zero
intensity of the fringes formed by the slit.

Width of the slit l>^ = 0.5745 mm.
Wave leno'th of liHit X = 0.000-1861 mm.

Obser\red Angle of
Deviation.

Calculated Angle of
Deviation.

Obs. — Cale.

1st

Miu.

U4.3

93.3

+ I'.'u

2nd

"

191.9

186.6

+ 5.8

3rd

283.6

279.9

+ 3.4

4tli

377.8

373.2

+ 4.6

5 th

473.8

466.6

+ 7.2

eth

566.9

559.8

+ 7.1

7th

652.7

653.1

- 0.4

9 9

H. XAGAüKA.

I 11 i)l)S(,'r\ iiiLi' I'ix'sirTs (litfractiiju phcuoincii;!, tlic ()pli(';il Ijcncli
was used. The iiitcr\:ils Ijctwccu the fringes were measured by iiieau«
of ;i iiiierometer. The iollowing table _ii'ives the observed uuiidjers.

R^= 324.0, E'^= 285.4 mm.'
26 = 0.347; ;i = O.00048() mm.

Obs. Distance.

Calcul. Distance.

Obs. — Gale.

1st Miu.

1.4o mm.

1.00 mm.

-0.17

2 lid „

:3.18

3.19

-0.01

ord „

4.70

4.80

- 0.04

4th „

0.39

G.40

-0.01

5tli „

8.0o

8.00

+ 0.03

(iih „

0.58

9.59

-0.01

7th „

11.22

11.19

+ 0.03

Sih „

12.81

12.79

+ 0.02

9th „

14.45

14.39

+ 0.00

10 th „

1 i;.(m;

15.99

+ 0.07

1 1 th „

17.01

17.59

+ U.U2

12 th „

19.21

19.19

+ 0.U5

Effect of Magnetization on the Permanent
Twist of Nickel Wire.

H. Nagaoka.

PI. XXXVIII.

Professor Wiedein;iiin, in a course oï experiments on tlie mutual
relation l)2t\veen torsion ;in<l magnetization, found that there was a
reciprocal relation between the two. lie found that wliereas torsion
changed the magnetization of ir(jn, magnetization, on the other hand,
clianged the torsion. To establish the relation between the two, he
made a series of experiments, whicli seemed to indicate main^ other
intimate relations between the two. Experiments relating to tlie change
of twist by longitudinal magnetization have been, so far as I am aw-uv,
tried only with iron and steel wires. The curious effect of torsion on the
magnetization of nickel has induced me to try experiments in the same
line, and find if there also exist simiLu' reciprocal relations Ijetween
magnetization and torsi(3n in nickel wires. Want of a])paratus did
not allow me to try experiments on the effect of magnetization on
nickel wires under different conditions of twist. The present ])aper is
confined oidy to the discussion of the effect of magnetization on the
permanent twist of ru'ckel wires.

The apparatus used for twisting the wire and measuring the
effect due to magnetization was essenti:dly different from tliat (jf
Professor Wiedemann. I employed an arrangement made on the s;nne
])laii as that used ])y Professor F. Kohlrausch* in his ex])erimenis on
the torsional elasti«; after-effect of wires. Fig. 1 shows flie front view

* Poojo'. Ann. 128.

324

H. NAGAOKA.

of the appnratiis. On a firm stand furnished witli three levelling
screws (III), two stout pillars (pp) were erected. A cross bar of wood
(hh) was fixed to these pillars. At the middle point of the cross piece,
a torsion circle (t) Avas attached, wdth an arrangement for fixing the
wire. Below this stood a magnetizing coil (c) on an auxiliary stand.
To keep the wire twisted, two stout rods (rr) were raised vertically
from a thick brass plate of circular shape, which was screwed to the
stand (s). These rods were fastened to alidades, which were movable

Fig. 1.

^

^

m.

m

^T

^T

p

Fig. 2.

D

0

'W

about an axis at the centre of the plate. Thus the rods could be fixed
in any desired position, and made to catch the cross attached to the
lower end of the wire. The cross was made of tw(^ rc^ds at right
ano-les to each other. The A-ertical rod had an arrano-ement for hold-
ing a small plane mirror (7h). The horizontal rod was capable of
sliding in the vertical, «and could be clamped firmly to it by means of
a screw. A vane (v) was attached to the lower end of the vertical

EFFECT OF MAGNETIZATION ON THE PERMANENT TWIST OF NICKEL. 325

rod. It dipped into a vessel filled with water, which served to stop
the torsional oscillation of the wire, when the twist was released. The
torsion circle had a stout rod (a) for vertical axis. This was capable
of up and down motion by means of ratch work, and could be clamp-
ed by a screw. The lower end of the axis was cut, and made to bite
the upper extremity of the wire. The wire is shown in Fig. 2. Two
small pieces of thick brass plate were attached to the extremities of the
wire. The upper end was placed between the terminal cleft of the
axis, and clamped by a screwing nut, while the lower end was similar-
ly caught at the upper end of the cross, and fixed by a screw which
went through a hole in the plate, as shown in the figure.

In front of the mirror was placed a circular scale divided into
half millimetres. The radius was 85.8 cm., so that one scale division
corresponded, when seen by reflection, to one minute of arc. The
scale was illuminated by gas jets, and the reflected image was observed
by means of a telescope.

The magnetizing coil was 30 cms. long, ;ind gave a field of 36.7
C. G. S. units by passage of a current of one amj)ere. In addition to
the magnetizing coil, a small coil was inserted within the solenoid.
Through this coil, a steady current was mainttdned to compensate the
vertical component of the terrestrial magnetic force. The magnetizing
current was generally olitained from Bunsen cells, and made to vary
continuously by placing a liquid slide in the circuit. It was
measured by a Thomson graded galvanometer. The difterent parts of
the apparatus being as described above, the experiment was conducted
in the followins^ manner.

A carefully anneided nickel wire was fixed witliin the S(3lenoid, its
upper end being screwed to the axis of the torsion circle, and its lower
end to the cross as before mentioned. While the wire was being set
in position, the magnetizing force was zero within the solenoid, the

ÖJ,(1

26 H. NAGAOKA.

vertical component of the terrestrial magnetic field being neutralised
by a current in the dmall coil placed within the main coil. The reading
of the tor.sion circle was then taken, and the two vertical rods (rr) were
so placed, that they just touched the horizontal r(jd (d) of the cross on
its opposite sides. The wire was now twisted by turning the torsion
circle, and held in the twisted condition for ,S(.)nie time. The reading
of the scale was noted. The circle was then turned in the opposite
direction so ;is to make the cross free of the vertical rods. When the
residual twist was small, the torsi<jn circle was brought back complete-
ly to its original position, and the small amount of residual twist was
given by the difference in the initi;d and final scale readings. For
l-irge residual twists, however, turning the torsion circle back to its
origin;d position would have thrown the reflected hcmn off the sc;de.
Accordingly the torsion circle was turned back tlu'ongh a convenient
and known angle, so that the wire hung free, and the reflected l^eam
remained on the scale.

In the preliminary course of experiments, after the wire was
freed from torsional oscillations, the elastic ufter-efFect was measured
by simultaneously noting the successive scale readings and the corres-
ponding times after the release. Wlien the untwisting due to the after-
effect had become very slow, the magnetizing fjrce was applied, and
the corresponding scale readings noted.

Before entering' into the in\ estin'ation of the effect of magnetiza-
tion on the permanent twist, it was desirable to have some knowledge
of the torsional after-effect of nickel. In every experiment, the wire
was twisted and left for some time in this state. On releasing the
torsion, torsional oscillations ensued. After its cessation, the wire
continued gradually to untwist in virtue of the elastic after-effect. It
was then necessary to know when the untwisting due to the after
efïect should cease, for otherwise the untwistings due to magnetization

EFFECT OF MA<JN'ETLZ VTIOX OX THE PEUMAXEXT TWlST OF XlOlvEL. 327

;ind to the elastic after-effect would l)e mixed together.

As a general rule, the after-effect for the same angle of twist is
smaller as the wire becomes thicker. For this reason, Professor
Wiedemann used tolerably thick iron wires. Althouo-li much cer-
tainty is gained as to the effect due to magnetization only Ijy usinn-
thick wires, yet there is the great disadvantage that the amount of
untwisting is ^ ery small. A\'itli nickel wires, the elastic after-effect is
very small, and we can use tliin wires without iiicurring tlie risk of
mixing the effect due to magnetization and tliat due to elastic after-
effect.

A nickel wire 0.51 mm. thick and 27 cms. l<jng was kept twisted
through 60° for an hour, the longitudinal ])ull acting on the wire
being the weight of the cross before mentioiied. When released from
torsion, the wire had a permanent twist of 2° 38'. On the cessation of
the torsional oscillations, the f(j] lowing deiiections with simultaneous
readings of the chronometer were taken.

Time Torsion

3'' 19'".0 p. m. (Uth \]n-\] 1889) 2° 38'.()

ol'.G ('remperature)

o7'.4 0".5

37'.2

37'.0
a. m. (12th) 3(r.2 (1)^2)

The readings show that the after-effect in nickel is very small.
The wire above tested would have been untwisted through a few minutes
mtjre, if we had waited for some weeks or months. J^oading the wire,
however, increases the after-effect, 1)ut when C(jmpared with the after-
effect in iron under similar circumstances, it is very small. It is un-
necessary to give the result of numerous similar experiments. Suffice to
say, they all lead t(j the same conclusion. The precautions which must

20.5

55

22.0

35

24.0

'5

36.0

55

42.0

a. n

328 H. NAGAOKA.

be taken in discriminating the untwistings due to magnetization and
to after-effect in nickel is greatly lessened as compared with the pre-
cautions necessary in the case of iron. If sufficient care be taken to
wait till the after-effect l^ecomes very small, we may use thin nickel
wires in the investigation of the effect of magnetization on torsion.
Generally I waited an hour after the cessation of torsional oscillation,
but if the wire was loaded, it was left for a night.

The thicknesses of the wires used in the present investigation varied
from 0.34 to 0.72 mm. Most of the experiments were tried with
the thinnest, for with it the effects were greatest. The wire was
always carefully annealed by means of a Bunsen flame. It so happen-
ed, that when the twist was very large, the wire once used assumed
a spiral aspect as was observed by Himstedt.* Such wires were rejected,
and other wires cut from the same specimen were used instead.

The experiment was first tried with the wires aljove mentioned,
when the permanent twist was very small, and the wire was subjected
to weak longitudinal stress. The following gives the readings of
untwistino- due to man-nctization when the load was the weio^ht of the
cross only.

* Wied. Ann. 17. pg. 712.

EFFECT OF MAGNETIZATION ON THE PERMANENT IWIST OF NICKEL.329

r^

0.17

r:=

0.24

r ^

0.35

Perm. Twist = 1'.6

Perm. Twist — 7'.2

Perm. Twist = 3'.3

Ô

r

S^

7

»

-

7.1

9'.3

6.0

4'.7

4.4

2'.9

9.8

17M

10.3

1 2'.3

6.3

5'.8

13.2

25'.6

13.2

16'.4

8.2

9'.1

15.9

30'-0

16.7

20'.6

10.3

12'.4

20.7

37'.3

24.1

25'. 9

13.4

15'.9

25.2

43'.0

35.4

29'.8

16.0

IS'.l

31.3

48'.0

42.2

31'.3

18.5

] 0'.5

40.7

52'.G

54.1

32'.9

25.8

23'.0

52.0

55'.5

67.2

34'.0

38.4

26'.7

65.0

57'. 6

157.8

34'.6

53.2

28'.1

82.9

59M

<)6.7

28'.8

112.7

60M

...

103.0

29'. 7

184.3

60'. 7

...

184.3

29'.9

The above table shows how the iiiïtwi.stiiig' proceeds as the
streno^th of the field is increased. AVith the increase of the mao-netizino-
force, untwisting becomes greater and greater, until at a certain point,
the ratio of the untwisting to the corresponding magnetizing f^rce
reaches a maximum ; in other words, the curve of untwisting has a
Wendepunkt. After this, the untwisting takes place very slowly, so
that ultimatel}^ the curve (Fig. I, PI. XXXYIII) becomes almost
straight. Up to ^ = 180, the curve doesjnot reach a maximum.

In comparing the curves obtained with different wires, we easily see
that the untwisting is greater for tlie thinner wire. Fc^r it will he noticed
in the experiments first given that the permanent twist is greater for
the thicker wire. Nevertheless, even with sucli liandicapping, it is tlie
thinner wire which has the greater untwisting as shown at a glance on
the curves.

330

H. NAGAOKA.

If after the magnetizing field has attained a certain value, it lie grad-
ually decreased, the wire again twists back. Tlie twisting produced hy
the removal of the magnetizing force is, however, far smaller than the
untwisting produced by the increase of tlie magnetizing force. Con-
sequently, the return curve goes above the other, as is shown hy the
dotted line in curve (1) Fig. I. This fact can be briefly expressed by
saying that there is magnetic after-effect in the twisting which becomes
conspicuous l)y the removal of the magnetizing force. So long as the
îimount of permanent twist remains very small, the curve showing the
torsional etfect of a continuously changing magnetie force reseml)les the
ordinary curve of magnetic hysteresis.

The al)ove remark does riot hold wlieri the permanent twist
exceeds a certain limit. The decrease of twist with increase of mag-
netization soon reaches a maximum. After this, the wire begins to
twist in spite of the increase (^f magnetizing force. The amoimt oi'
twisting of course varies witli the ])ermanent set of the wire as well as
with the amount of pulling stress. The following tal)]e gives the
amount of chano'e of twist with the wire of 0.17 mm. radius.

Perm. T

wist 6'. 7

Perm. Twist 10'.6

Perm. T

wist scr

4.7

r

>>

7

>>

r

11 '.5

2.7

7'.0

6.7

26'.5

7.U

18'.2

•3.9

1 5'.8

1 ().0

38'.6

10.0

24'. 7

8.8

19'.G

12.3

45'.5

13..'3

26'.8

12.8

23'.0

15.1

52'.ü

17.0

20'. 9

17.1

25'.8

17.7

54'.9

18.2

2()'.9

20.5

26'.8

23.1

61'.0

27.1»

26'. 2

:}2.0

28'.1

:îl.6

64'.2

42.2

25'.1

1.7.0

26'. 9

5U.1

GO '.9

G2.3

24'.7

72.«i

22'.(»

62.8

56'. 1

125

24'.2

12.4

11 '.5

88
157

4G'.2
19'.0

EFFECT OF MAGNETIZATION ON THE PERMANENT TWIST OF NICKEL.

331

These reading« are plotted in curves (1) (2) (3) Fig. 2. respec-
tively. Curve (4) is plotted from an experiment made with a wire of
the same thickness, with a permanent twist of 1621°. These curves
shoAv that tlie untwisting on the first application (jf the magnetizing
foi'ce is very large. AVhen the twist is small the untwisting im-
mediately becomes very small, and the wire 1)egins to twist, liut the
further increase of the magnetizing force is of very little effect. The
curve, shortly after the maxinum is attained, liecomes nearly ])arallel
to the line of no twisting.

This appearance is confined to those cases where the ]3ermanent
twist is small. With a residual torsion of 10°.6 in the same wire, the
curve acquires quite n différent appearance. The I'ate of increase of
untwisting with the increase cjf magnetizing force liecomes less, so that
the untwisting gradually approaches the maximum. Thereafter the
twisting takes place gi'adually and steadily. On removing the mag-
netizing force, there is at first untwisting Avhich reaches a maximum
in a magnetizing Held less than that corresponding to the maximum
untwisting on the first a])plication of the magnetizing force. The
wire then again begins to twist, Imt on the complete removal of the
magnetizing force, the Avire remains untwisted relatively to its first
position. The most striking difference between the curves in Fig. 1
and those of Fig. 2 is that the latter has a maximiun ])()int and the
former has none. This maximum which seems to be closely connected
with the amount of residual torsion occurs in weak magnetizing fields
when the twist is small, l)ut as the twist is increased, it occurs in
stronger fields.

When the permanent twist is very large, the features of the curve
do not change essentially. Curves (3) (4) Fig. 2. show the state of
things for the twists of 861° and 1621° respectively. From these it
will be seen that the untwisting does not increase ])roportionally with

332 H. NAGAOKA.

the periuînieiit twist. On the coiitrary, the luitwisthig for the twist
of 861° is greater than that for tlie twist of 1021°. The course of the
curve, after passing the ur.txiiuuni becomes steeper with tlie larger per-
inauent twist as the coui|)arison of (1) (2) with (o) (4) will show.
Thus, when the twist is large, and the magnetizing force sufticiently
great, the curve may he expected to cut the line of no twisting.

Another difference in the curves of torsion ol)t: lined f )r different
permanent twists consists in the course of the ciu'\'e on the removal of
the m;ignetizing f^rce. In curve (2), we find tliat the "off"' returns
below the " on " curve, while in curve (3), it returns above it. In
the former there is hysteresis or lagging, in the latter priming or
negative hysteresis. This distinctive feature in the curves obtained
for different twists also varies with the thickness of the wire.

It is unnecessary to give numerical details f )r the various experi-
ments made with different wires and with different twists. The
characteristics above descrilied are illustrated in the curves of Fig. 3,
which gives the results f)r nickel wires of diameters 0.5, 0.4, 0.7 mm.
For these also the untwisting re-iches a maximum f )r a comparatively
low field, and a hoisting liegins to set in, and continues to the highest
field used.

The f )llowing experiment shows that this twisting may proceed
so fir as to result in a final condition of twistednrss relatively to the
original condition of the wire. The wire, 0.34 nriu. thick and 30 cms.
long was twisted through eight complete revolutions of the torsion
circle, and then released. It thus acquired a ])ermanent twist of
2548°. The magnetizing current was derived from a shunt dyn-niio.
The current strength was ndjusted l)y the li([uid slide bef )re described.

EFFECT OF MAGXETIZATIOX OX THE PEEMAXEXT TWIST OF XICKEL. 333

Field.

Untwisting.

14.3

+ 20'.0

3G.0

49'.0

68.0

53'.0

81). 4

51'. 0

152.0

34'.0

200

24'. 0

245

12'.3

308

+ CO

361

- 4'.0

432

-lO'.O

ïlie application of the iiiaii'netiziiiijf force .showed at tir.st an un-
twisting, which reached a niaxiniinu in a field strength of ahont 65
C. G. S. units. The wire then began to twist. In a field ofal)<)ut
335 units, it came back to the condition in which it was after release
before the magnetizing force was applied. Thereafter the wire continued
steadily to twist with the increase of magnetizing force, so that when
§=432, the wire became twisted 10' from its initial position of
ec[uilibrium. Thus a nickel wire with large permanent twist can he
twisted by applying sufQciently great magnetizing force. As the
course of the curve after passing the maximum is less steep in thick
than in thin wires, still stnniger magnetizing fields will l)e necessary
to twist the former.

The next set of experiments has to do with nickel wires under
longitudinal stress. The only change in the process of experimenting
consisted in loadirig the wire. Tlie vane was detached from the lower
end of the cross and a short hook placed in its stead. A pan of
weights hung from this hcx^k, and was completely immersed in the
water.

334

H. NAGAOKA.

Different experiments were tried with wires oï various thicknesses,
and with different amounts of twist. lS(3me of the results are shown
plotted in Figs. (5) and (6). In all of these the untwisting by mag-
netization becomes greater l)y ]oa(hng. When the permanent twist is
small, the curve representing change of torsion reaches a maximum
quite aln-uptly. The course of the curve immediately after passing the
maximum is quite steep for some time, but after the magnetizing force
attains a certain value, the return twist l)ecomes very small. More-
over, there is hysteresis on the gradual removal (3f the magnetizing
force. An inspection of the figures will be of more service than mere
verbal description.

With large twists, the features of the curve of torsion do not

greatly differ from those obtained with the unloaded wire. The chief

change wrought by the loading is that after the maximum untwisting

has been passed, the curve goes down more steeply than in the case

when there is no load. This evidently suggests the possibility that

the curve for the loaded wire will cut the line of ntj untwisting in

mao-netizinof fields smaller tlian those needed to effect the same for un-
es o

loaded wires. And this I f(3und to be the case, as shown by the readings
on the following page, which were made on a wire of 0.17 mm.
diameter under a load of 342 grm. Aveight.

The former readings are shown plotted in curve (2) Fig. 4, and
the first part of the latter in (4) Fig. 6, and the readings in strong
fields in curve (3) Fig. 4. The comparison of these two curves with
(1) sliows that with the loaded wire the initial position is reached at
smaller mao-netizintr fields than with the iud(xaded wire. Moreover,
there is hysteresis when the permanent twist is moderate, but priming
when the twist becomes large.

Finally the effect of transverse magnetization on the permanent
twist was investig-ated. The \v\re beinir treated as before described,

EFFECT OF MAGXETIZATEOX OX THE PERMAXEXT TWIST OF NICKEL.

335

Perm. Twist 95°.

Perm. Twist 583'.

Field

Untwisting

Field

Untwisting

5.8
10.8
14.6
19.4
24.6
28.6
33.5
38.8
44.1
52 2
61.5
72.6
98.1
125.7
182.1

16'.1
26'.2
32'.7
36'.8

4r.o

42'.6
42'.8
4r.6
40'.0
36'.4
32'.0
27'.0
16'.9
9'.5

-r.3

6.1

7.5

10.9

14.1
17.4
24.5
39.4
50.3
54.4
66.7
85.6
98.6

164.2

191.9

271

328

8'.0
15'.0
23'.5

33'.0
39'.4
53'.6
73',0
76'.2
77'.2
7ô'.0
69'.5
62'.0
25'.0
17'.0
- 4'.0
-12'.2

was placed between two Hat coil.-s, through which niagiietizing currents
of various strength were passed. The wire, liowever, did iKjt show the
least sign of being affected by transverse magnetizati<jn, although the
apparatus was capable of measuring 0/1 of deflection.

The next point of inquiry was a comparison of these effects with
those produced by twisting magnetized nickel wires. The apparatus
used for examining the latter was similar to that used by Professor
Wiedemann, in his investigation on tlie effect of twist on magnetiza-
tion, and described in his ' Elektricität ' Ikl. 3. A nickel wire 1 mm.
thick and 30 cm. long had two pieces of stout brass wire soldered at
the ends. The wire was placed magnetic east and west, and carefully

336

H. XAGAOKA.

niineîik'il in this positùju. It was then slid into a magnetizing coi],
and its extremities firmly clamped to the twisting apparatus. The
magnetizing current was gradually increased l)y means of the liquid
slide, and then slowly remoyed. Thereupon the deflection of the
magnetometer mirror with corresponding angle of twist was read.
The following table giyes the changes produced on the permanent
masfnetism in arbitrary scale unit, the amount of permanent maof-
netism being proportional to tlie numl)er of scale diyisions when the
twist is zero.

Twist

(I)

(11)

(III)

( IV )

(V)

0

589

455

183

83

41

5°

549

410

161

74

27

10^

550

885

159

78

20

15=

554

378

173

88

19

20^

558

384

194

102

24

25°

560

393

212

119

33

;30°

559

402

224

131

40

35°

554

405

233

141

47

40'^

551

405

238

149

52

45°

404

238

153

50°

540

403

239

155

62

60°

529

397

237

I5,s

68

70°

235

1 61)

71

80°

...

160

90°

161

74

120°

1()()

75

180°

76

270°

74

EFFECT OF MAGNETIZATION ON THE PERMANENT TWIST OF NICKEL. 337

The examination of the ab(3ve talkie shows that the first effect of
tAvist is always to decrease the magnetism of the wire. This decrease
soon ceases, and an increase sets in as the twist becomes larger. AVhen
the perm:ment magnetism is large, the increase is small, and the
oricrinal value of tlie mau'netic moment is not recovered. On theotlier
hand, for small values of permanent magnetism, the increase is consi-
deral)le; and as the twisting continues the wire acquires a greater magne-
tic moment than it had originally. When the wire is further twist-
ed, the magnetic moment reaches a maximum, and begins to decrease.
The maximum comes earlier for greater values of permanent magnetism.
The maximum increase ivi weakly magnetized wire occurs at tolerably
larf>'e twists, as an examinati(3n of the above tal)le will sliow. In
addition to this, the range of change in permanent magnetism l)y
twisting does not increase, but rather seems to diminish with the
auKHint of permanent magnetism, for moderate angles of twist.

The experiments hitherto described show close relations between
the effects produced by twisting the permanently magnetized wire, and
those produced by magnetizing the permanently twisted wire. The
relation between the two can be most clearly represented by collecting
the results in the f)l lowing parallel statements.

1 The permanent magnetism I. The permanent twist of nickel
of nickel wire is at first dimini- wire is at first diminished by
shed l)y twisting. magnetization.

2. With Zan/t' ])ermanent mag- II. With swzaZ/ permanent twist,
netism, the decrease incre;ises with the untwisting increases with the
increased twist. strength of magnetizati(jn.

3. Unless the permanent mag- III. Unless the permanent twist
netism is very large, tlie decrease is very small, the untwisting pro-
produced by twisting reaches a duced 1)y magnetization reaches a
maxhnum. Further twisting in- maximum. The twisting produced

338 H. NAGAOKA.

creases the m;!<rnetisin, so th;ît it by farther increase of nni^'netiza-
hecomes greater than its original tion is so large, that the wire ac-
value. (juires greater twist than it ori-

ginally had.

It a|)])ears from the readings given above for the changes in
permanent magnetism, tliat there is a tendency to a decrease ugain
setting in at the higher twists. This suggests that there may be un-
twisting in very strong fields, after the wire has been for some time
twisting under tlie influence of m;ignetization. The current at my
dis])Osa] did not allow me to try ex{)eriments with fields much c)ver
400. Up to that limit, the twisting ccjntinued. It still remains un-
decided if further increase of magnetizing force gives a maximum
twisting, corresponding to the maximum value of ])ei'manent magne-
tism obtained by twisting.

When the subject is viewed from tlie theory of rotating molecular
magnets, we fall into difficulties which cannot be easily explained.
Professor Wiedemann in coordinating the nnitual relatii^ns between
twist and magnetization of iron and steel wires, assumes that the mole-
cules are subject to disturl)ances in trying to point tlieir poles in the
directicm of magnetization. Drawing an analogy from the effect of
mechanical disturbance ap])lied to the twisted wire, he concludes that
the disturbance caused by magnetization must untwist the iron or steel
wires. This easily ex]>lains the efiect of inagnetization on the per-
manently twisted iron wire. It seems (juite ])ro])able that a, similar
explariation can be applied to the untwisting observed in nickel wires.
The effect of magnetization, however, is not so sim])le in nickel as in
iron. It seems very difficult to explain the luaximum untwisting ob-
served in nickel. Moreover the disturbance caused in molecular group-
ings is not limited to longitudinal magnetization only. Transverse
magnetization must likewise produce similar changes am<mg the

EFFECT OF MAO>'ET]ZATTOX OF TUE FETîMAXENT TWTST OF XTOKEL. .^39

molooiilos. Thus tlio pormnnont twist would ho afl'ectod by trnnsvorsnl
:is well :is 1)\- loiigitii(liii;il iiiuLjiioti/.titioii. In my cxjM'i'iincnts, lr;iiis-
\(Ts:il iiingTiC'tizutioii 1)\- 1l;it coil Ik-mI no sensible eft'eet.

riie ('liaiiii'e oi jn'rinnncnt niii^'nctisni by twistiiio- is nmre eoni-
plex in nickel tlinn in iron. Tf we Imve to explain the nitixiinnni
(Iccrense in ])erin:inent in:iL!'n*'tisni on \\'ie(l('ni:inn's rlicorx-, we nnisi
nssiiine th:it tlie iiiekcl niolccnles n^tîitc only tbrono-]) ;i eert;iin niiiilc
l)y twisting, l)nt lieyoiid th;it niigle, tliey move ])nek townrds the
orininnl jiosition. 'I'liis we lime no i-iglit to nssuine. It seenis bojieless
to find :iny ex])lan:ition of tlie ^':lrio^s relati(^ns ])etweeii twist and
mauTieiization in tiTins of* a really satisi'aetorv theory of i-olaling
nioleenles.

On Certain Thermoelectric Effects of
Stress in Iron.

By

C. G. KnoU, D. Sc, F.R.S.E.

Professor of Physics, Imperial University.
And

S. Kimura, Ri^akushi.

Since the (li.^covery mrifle by Tlioni.^^on thnt the thermoelectric
propertie.:! of wire:; of certain metal.^ were altered l)y tension, the sub-
ject has lieen studied cxperimentnlly by various scientific men. Of
these we may mention more particularly Le Ivoux, von Tunzelman,
Colin, and Ewino;-. Tlie work done by Cohn and Ewing is of special
im])ortance ; and the latter's investigation for iron is the most com-
plete that lias been carried out. Eeference will be made to their
results hereafter. It is sufficient at present to point out one respect
in wliich the work of these experimenters lacks completeness. In all,
the method (^f exjX'riment consisted in studying the effects of stress
upon the thermoelectric properties of a wire, whose junctions with tlie
other es.^ential wire of the circuit were kept at steady temperatures.
The variations of stress were, in the best experiments, carried through
a cycle ; and at different successive stages tlie thermoelectric current
was measured on a suitable galvanometer. The observed changes in
the electromotive force might be due to either of two quite different
effects ; and the experimental methods adopted could give no criterioi^

342

C. G. KNOTT AND S. KIMURA..

by which to rlrnw the correct conclusion. The nature of tlie prol:)]em
is most sini])ly expressed in terms of the Jnng-nnge of the thermoelec-
tric (liag'rnm. In this diagram tlie thermoelectric relations of the
ditfererit metals are represented l)y liries (usnally straigdit) in such a
manricr that the electromotive force existing in any circnit of two
metals is cqnnl to the area included between the appropriate metal
lines arid the two lines drawn perpendicular to the temperature axis
and throngli the j^oints representing the temperatures of the two junc-
tions. The (pie;!tion pr(^pounded aliove is tlicn this. AYhat chp.nge
does stress applied to a given metal produce upon tlie jjosition of the
line in the tliermoelectric diagram ? Does it translate it as a Avliole u\)
or down ; d<X's it rotate it as a whole about some definite point ; or
does it effect a cond)ination of these so that the line is deformed as
well as shifted ? In other words does stress cliange the Thomson
.Effect in a wii-e, or does it sim])ly change the l\'ltier Effect with
reference to an unaffected second wire ?

Now it is (piite clear that the only way to answer this problem is
to arrange an apparatus in which the electromotive forces due to
(lifcrnit differences of temperature can l)e measured siinnlfaneouslij on
a wire under given conditions of stress. This could be accomplished
only by having tlie gradient of temperature along the wire both
steady and gradual. Junctions could tlien be made at several points
along tlie Avire ; and the electromotive forces due to the several cir-
cuits so olitainable could be easily measured and compared, once the
temperatures were steady. The sim])lest way of realizing these condi-
tions seemed to be to stretch the wire inside a metal tube, and then to
heat the metal tube as in the Forbes Experiment on the conduction of
heat along bars.

For ease in manipulation the tube, which was of iron, was made
in two semi-cylindrical parts. The upper part or lid fitted accurately

ON CEUTAIN THERiMOELECTiilC EFFECTS OP STKEri.S IN IRON. 343

upon the lower part whieli rested lioriz(jiit'i]ly on oliar[) edged ;uip[)ort«.
The lower part wa.s somewhat longer than the up[)er ])art, the extra
length being a solid cylindrieal piece of iron which during the ex[>eri-
nient was sustained at a IjriLZ'ht red heat in a charcoal furnace. To
render the fitting secure a ridge cut out longitudinaJIy on each plane
surface of the semi-cylindrical lid fitted into a groove cut out on the
opposing surface of the lower part. At suitable intervals along this
lower part small radial notches were cut. These became IkjIcs when
the lid was set in positit^n, and through them wdres were led from the
interior of the tube. The wire to be used was stretched alonij' the
axial line of the tube ; and it and all the various junction wires were
arranged and adjusted l)efore the lid was laid in position. Each junc-
tion was a junction of three wires — ( 1 ) the axial wire to be tested,
( 2 ) a tliin wire of the same material, ( 3 ) a thin wire of some (jther
metal. The tw(j last formed what we shall call the Thermometric
Circuit. Its indications served to measure the temperature of the
junction. The circuit, formed Ijy the axial wire and the thin wire of
the same materia], was the essential element in the experiment. We
shall call it the ThernKjeiectric Circuit.

The tension was ap[)lied by means of a screw at the extremity of
the wire, Avhich projected some distance from the (jpen end of the
tube ; and was measured on a spring dynamometer set in line. To
])revent currents of air circulating in the tube, the open cold end was
plugged with cotton wo(j], and the side IkjIcs, tlnvjugh which the thin
wires came, were filled up with asbestos. The liot end of the tube was
closed naturally by the Acrtical face of the solid cylindrical portion
îdready menti<jned. The end of the wire was clamped to this face.

The current were measured on a hiu'li resistance double coiled
galvanometer, which was carefully gauged after every single day's
experiment.

â44

C. G. KXOTT AND S. KIMUIIA.

The g-eiiei'al pliiii of (jx[)uriiiieiitiiig w.u oiiiiple eiKJUg-h. After
the tulje und coniaiiied wire h'id attained a ;iteidy coiidici(jii a.i regard«
temperature, a «eriei of reading.^ of the différent thermoelectrie and
thermouietrie enrrent« were taken a« rapidly a.; poy.-sible, with a .suf-
ficient nunilier of repetition« of the same to yield a good mean for
each individual circuit. Thi.s operation wa« carried through for a
series of ascending and descending value« of tension. The small
value of the current in the thermoelectric circuit as compared with
that in the thermometric required that only a small shunted portion
of the latter should be taken through the galvanometer. This neces-
sitated a somewhat complicated arrangement of resistances and com-
mutators, which hoAvever it is unnecessary to describe.

In the earlier experiments the thermometric circuits were of
copper andiron, and the thermoelectric of iron and iron. J5y using
copper and iron, we expected to be able to get good measurements of
the true temperature values of the junctions ; and this because of the
existence of a neutral point at an easily attainable temperature. It
was found, however, that the uncertainties of reduction from the
parabolic temperature scale of experiment to the linear scale of
accepted use far (outweighed the advantag(3s of having an (_)b-ierved
neutral point as a guide. Accordingly a.fter many experiments had
been made the Copper Iron thermometric jiUiction was aljandoned in
favour of a Gierman- silver Ir()n junction. As is well known, the elec-
tromotive force of this pair of metals varies in ;in approximately linear
manner with temperature up to a dull red heat. The graphical com-
parison of the thermoelectric with the thermometric currents will not
in tliis case differ greatly in appearance from wliat would be the case if an
accurate aljsolute scale of temperature were used instead. Ultimately,
of course, the thermometric readings were reduced to the ordinary
temperature scale by calculation frcjm the results of direct ex[)eriment.

ON CERTAIN THERMOELECTRIC EFFECTS OF STRESS IN IRON. 345

It wu.s deteriniucd to experinient tir.st with injii wire. Previous
worker« liad all Ibiind that the thernioelectric effects of stress were
much more prouounced in this metal than in others. It seemed
natural therefore to begin with it. Sh(3uld the experiments prove
promising', it was intended to pursue the en(|uiry in regard to copper,
nickel, platinum, etc. A few experiments were indeed tried with
copper and nickel wires ; but in the latter its viscosity under the
influence of sustained stress pn^duced a gradual decay in the value of
the stress, applied as it was by a tightened screw. It was obvious
that a steady stress could be applied only l)y means of a load acting
by its weight ; and ïov this the apparatus was not readily adjustable.

Several modifications in the mode of experimenting were repeated-
ly tried before resiüts (jf a satisfactory character were obtained. In cer-
tain experiments with the iron Avire, theruK^electric changes of very
small amount were (jbtained by simply varying the tension without
having established the temperature gradient. This thermoelectric
éUect increased with the tension. The direction of the current was
opposite to the direction of all the currents obtained when the gradient
of temperature existed ah^ng the wire. In (jther W()rds, the current
was such as miu'lit have resulted from a slight heating of the wire
where it was gripped by the dynamometer clamp. The probable
explanation of this effect is that the part of the stretched wire which
lay outside the tube was a little warmer than the part inside the tube.
ISuch a slight gradient of temperature might easily ensue under the
influence of the air as it grew warmer with the advance of dav, the
more massive tube changin£>' more slowly in tem])erature. If this is
the true explanation, the elfect will have no existence in the real
experiment, in which a steady temperature gradient is to be sustained.
In any case, however, these initial currents, as they might be termed,
were much smaller than the currents sul^secpiently oljtained.

346 C. G. KNOTT AND S. KIMÜKA.

After the be.st iiietliod of experinieutiiig had been by long trial
decided ii[)on, the character oï the experiment td part of the rewearch ^y•dü
in it.self very tediou.s ; and «ince month« of preliminary and other-
wiöc futile labour had already been .spent it :-eemed be;^t to po.^tpone a
.continuation of the experiment« till «ome future date. So far there
have been no opportunities for renewing the attack, other work fully
enL'TOs.sini>' our time.

Vie are now pre[)ared to di«cu8.s the result« of the fi nal set of
experiment« with iron wire.

The dimen«ion« of the tube bar were a« follow« :

Total length of l)ar 102 cm.

,, ,, tubular part 90 ,,

External diameter (jf ,, ,, 4.4 ,,

Internal ,, „ „ „ 2.2 „

The diameter of the iron wire u«ed was 1.2 mm. It projected about
a foot beyond the c<)ld (jpen eiid of tlie tul^e and wa« attached to a
spring dynamometer measuring p(/unds- weight. The dynamometer
wa« fixed to a «crew working in a lixed nut ; and Ijy thi« means the
tension could be increased or diminished as desired.

In the final set of experiments each applied stress acted for at
least one wIkjIc day before the thermoelectric observation« w^ere
begun. The wire w;i« left ïnv this interval at the «nxlinary tempera-
ture of the air.

The solid end of tlie cylinder wa« then heated to brio'ht redness
in a cliarcoal furnace ; arid after 2 hour«' heating the temperature
gradient l)ecame fdrly steady, as indicate«! l)y the thermometric cur-
rents on the iz'alvanometer.

o

There were five pairs (jf jiuictions, ten in all — five thermoelectric
and five thermometric. The [xjsitions of these juncti<jns along the

ON CERTAIN THERMOELECTRIC EFFECTS OF STRESS IN IRON. 347

iron Avire were so nrrancfed tlint the temperatures of two successive
posirions difl'ered l)y 40°-60° ('. Tlie ]):iir:-; of junotious were distin-
guished hv iiumlter, Xo. T. Ijcinu' tlie hottest niid Xo. V. the coldest.

The o])serv:itioiis were in:i(h' in tlie f )llowinü' order. First, tlie
five thei'nionietric curi'ents wei'e measured in ra])id succession i'roui T.
to y., eacli ciu'i'ent Ix'ing" measured first in tlie one and then in the
other direction throuu'h the galvanometei'. [This was an invariable
rule in the measurement of ;dl currents, the total ran^'e from the
direct to the reverse reading- <i'ivinu- twice the true deflection.] Then
followed a simi]:u' set of reading's of tlie fi\'e thermoelectric currents ;
then a secoiid set of tlie thcrmoinctiâc ; and so on until 4 sets of tlie
thermometric currents witli o iiiterpofsted sets of the thermoelectric
currents had l)een comjtleted. Kxactly similar sets of ohservations
were made for the series of tensions re])resenting t(^tal loads of 0, 5,
10, 15, 20, 24 and 0 pounds. Ke(hice(l to kilou'ram-weiu-ht per square
millimetre, these tensions are a little L'Tcatcr than 0, 2, 4, G, (S, 9*6,
and 0 respectively.

In Tal)le T, the ohservations are given in full ior the üve ])airs of
junctions ; T stan(hng for the thermometric junctions wliose indica-
tions form an arbitrary tem])erature scale, and E f n- the thermoelectric
junctions whose e]ectr(^motive forces are the real sul)jects of investiga-
ti(^<n. Tlie tensi(^n, the temperature of the cold junctions, and the
factor fn- reducing tlie deflections to electromagnetic units of electro-
motive fn'ce, are given in the space to the left of the tal)ulated numhers.

348

C. G. KNOTT AND S. KIMURA.

Table I.

Scale. rendings of the Galvanometer-deflectioriS of the E. ]\I. F.
of Iron-Copper junctions (T), and of Iron-Iron
junctions (E) in the live Tlaces.

Tension 0

I

IT

III

TV

V

T

E

T

E

T

E

T

E

T

E

170

134

100.7

75.15

50.55

Cold tpuip. l^^°.0~l''

140.75

109.5

87.3

09

49.2

Galv. Factor 147.5

174.5

138

104.15

78

53.(5

14th Oct. 1889

174
173.5

141
111

137.7
137.5

111
111

104.2
104

89.75
90

78.5
78.4

72.25

72.75

54.25
54.5

52.2
53.5

158,5

123.4

91.5

07.5

4 k 5

Tensio 1 5

130.75

101.3

79

01.7

42.2

Cold temp 14°.4-15'.5

159.9

125.5

94.15

70.5

47.5

Galv. Factor 143

137.25

103.25

81.75

05.3

•10.5

Tension applied

100.7

120.75

95.75

72.25

49.3

11.30 a. m. Oct. 14tli

139.75

105.25

83.75

08

48.75

to 1 p. ni Oct 15th

103.75

129.25

97.8

73.7

50.5

100.9

125.75

94.5

70.5

47.95

Tension 10

142.2

105.0

84.9

07.95

40.5

Cold temp. IG'-IT'.S

101.25

120.75

95.7

72.15

49.05

Galv. Factor 143.3

140.5

104.5

84.5

08.95

47.8

Tension applied

159.2

125.5

94.85

71.5

49 0

7 a. m Oct. 17th to

130.5

101.95

82.0

08.2

47.9

9 a. m Oct. 18th

155.5

122.75

93.25

70.8

49.25

103.25

120.55

93

08

44

Tension 15

144.1

105

81.5

02.25

39.7

Cold temp. 13' 8-15'.2

172.2

1.34.75

10075

75

50.25

Galv. Factor 143 05

143.75

107.05

85.3

08.5

44.75

Tension applied

171.5

135.5

102.5

77.15

52.9

11 a. m. Oct. 18th to

145.15

108.3

80.5

70.4

47.2

9 a.m. Oct. 19th

174.5

138.5

105.25

79.1

54.5

ON CEETAIX THERMOELECTRIC EFFECTS OF STRESS IX IRON. 340

Table I. (Continued)

I

II

III

TV

V

T

E

T

E

T

E

T

E

T

E

170.8

134.5

101.5

76.85

52.8

Tension 20

150.25

114.5

92.75

76.75

54.5

Cold tomp. 17'.3-18 .5

171.75

135.35

102.5

77.45

53.7

Galv. Factor 14:î.5

152

116.5

94.75

79.65

56.75

Tension applied

171.2

135

102.5

77.5

53.75

9 a. m. Oct. 2lst to

151.75

116.75

95 25

80.4

57.95

10 a. ui. Oct. 22ud

169.25

134

102

77

53.45

1G1.G5

126

93.1

69

4.->.5

Tension 25-22.5

145.75

107.5

82.95

64

41.7

Cold temp. 14'.8-1GM

171 j

133.75

100.15

74.7

50.4

Galv. Factor ?

148.5

112

89

71

48

Tontion applied

173.5

136.75

103

77.25

52.75

10 a. ui. Oct. 23rd to

14G.05

111.7

90.5

74.3

51.8

2 p. m. Oct. 2-ith

1GS.7

133.3

102

76.65

52.75

164.1

127.95

95.3

71.25

46.25

Tension 0

143

108.7

8.>.5

63.3

43

Cold temp. 1G°3-17'.0

171

1.34.5

101

76.25

50.7

Galv. Factor 138.7

144.95

111.75

89

67.55

49.05

Tension released

173.7

137.5

104

79.05

53.15

11 a. m. Oct. 25th to

145

112.25

90.3

69.5

50.8

10 a. m. Oct. 31st

173

137.25

104.5

79.55

54.5

1

It may now be n.^sunied tliat the mean <ifanv set of four iinml)eiN
for T will correspond to the mean of the .set of three iiiimhers loi- rlie
J'] of the same jiinotion-])nir. Tlie.se means are then to 1)e rcihiccd to
tem])eratnre in the one ca.se, to electromotive force in the otiier. For
tlic re(hictioii to temj^eratnre independent exjX'rimcnts were made to
determine the constanrs of tlie ir(-)n u'erman-siKei- circuits ; ;ind in llie
final reduction full account was taken of the slii^-jit \ai'i:irions in the
\alue of the cold junction tem])erature. The residts of the reduction
are emliodied in Tahlell., all the cold juncti(^n t(Mn])eratui'es heino-
reduced to lo°.ô C.

350

C. G. KNOTT AND S. KIMÜRA.

Table II.

E. M. F. between stretched and unaffected Iron wires, at
various Tensions and Temperatures.

Ten-

Cold

Hot

E.M.F. IN

Ten-

Cold

Hot

E.M.F. IN

sion

Temp.

Temp.

MiCROVOLT.S

sion

Temp.

Temp.

Microvolts

0

13°.:.

2G7°.9

5G2

6

1 3°. 5

269°.4

589

2ir.7

437

2ir.o

436

159°.4

352

157°.5

344

119°.7

282

160°.6

274

82^0

205

77^5

179

2

13^5

2.55°.0

562

8

1 3°.5

272M

617

199^8

421

215^3

473

149°.6

333

163M

385

iir.5

265

123^3

322

74^7

187

85^2

230

4

1 .3°.5

254^2

570

9ß

13°.5

2 78°. 9

624

199".8

424

218°.7

468

150°.9

343

164^2

366

1 13".4

279

122°.2

29(>

78M

193

86°. 7

2(»0

()

1 3°.5

281^0
22^.3

16 6°. 7

^25^9

84^3

605
4<)6
371
281
203

The heading's of the cohunns sufficiently ex])lain themselves.
Tlie lensii^ns are expressed iu kilourani-weio-lit ])er S(piare iiiilli-
nietre. The hiii'liest tension atlained corrcsjiouds to :i |():id of 11.3

ox CER'l'AIX THEKMUELEOTRIC EFFEOl'Ö OF ,ST1{L:SÖ IN IliOX. o51

kilo^. actiiiii' aloiin' the wire. It will Ije iiotired that tliere i.s u wlixrht
diminution in this liinliest tension as the experiment progressed,
doubtless due to the yielding of the highly heated part of the wire.
This yielding occurred at all the tensi(jns if tlie experiment were be""un
soon alter the tension was applied. For this reason, eacli new tension
was allowed to act at least for a wlujie day before the thermoelectric
experiment was l)egun. Also just Ijefore the taking of the observa-
tions tiie (hnamometer was carefully looked to, and the tension was
raised to the desired value if any slight fall had occurred. Of course,
once the experiment itself was entered upon tlie wire was not touched
until the wliole series (jf observations had been completed. To go to
higher teusioiis than those here recorded was not practicable because of
the diminished tenacity of the wire at its blattest parts. Not a few
experiments were spoiled by the Ijreaking of the wire at or near the
highest tensi(Hi attempted.

For each tension we have determinations of electromotive fn'ces
at five different temperatures. Some oï the results are shown
in Figure T., IMate XXXIX. To prevent confusion of figure, only
three are shown — the initial and final ïov no tension, and the fifth for
tension 8. Of pai-ticular interest in the manner in which the initial
and final curves cut each other at a temi)erature of about 150° or 160°
C. In interpreting this result, we must know the thermoelectric rela-
tion of the two kinds of iron used in forming the junctions. In tlie
language of the thermoelectric diagram, in which tbe german-silver line
lies below the iron line, the iron forming the small wires had its line
also below the line of the iron that was or was to be strained. In
other words, tlie current always fli^wed fr(;m the unaffected wire to
the strained or to be strained through the hot junction. Xow from
Fig. I., ^ve see that the effect (jf the stress is to increase the currents
for all temperatures. The wire imder the stress 8 has therefore the

352 CG. KXOTT AND S. KIMUKA.

same relation to the iiiistrained wire wliich this latter has to the small
unaftected wire. The stress, s(i to speak, displaces the liue upwards
on the diagram. The eurreiit is acc<3rdiiJLi'ly from the unstrained to
the strained ir(jn through the liot junction. On the stress bein,u'
renKJN'ed, the wire is left permanently strained, or, as we shall for
brevity call it, after-strained. And we sec that for temperatures below
lo5°± the current is from the afterstrained to the unstrained throuii'h
the hot junction ; but that al)ove 155° tlie current passes in the other
direction, 'iliis Avould mean that tlie dia^'ram lines for the unstrained
and after-strained wires intersect each other indicating a neutral tem-
perature at a tem])erature of 85° oi- tliereab(juts. The directions of the
currents as «iven above show that the diai»ram line for the after strain-
ed wire is inclined at a less angle to the lead line. Hence the (nega-
tive) Thoms(jn Eifect in this particular iron wire is numerically
decreased after the application and withdrawal of longitudinal tension.
Curves, representative of all tlie experiments whose results are
given in Table II., were carefully drawn by free hand on a large scale ;
and from these the electronKjtive forces corresponding to particular
temperatures were picked «nit. A more pretensious process of inter-
polation could hardly ha\e ])een more accurate under the circum-
stances ; for the curves, though smooth, liaNc all a distinctly sinuous
form, which it would Ik; ditHcult if not impossible to repfesent by an
equation of degree lower than the Iburth. The electromotive forces
corres])(jnding to convenient temperatures, picked out as just described
by ins])ection of the curves, will be fbimd tabulated in Table 111. ; and
in Tahle W. the result (jf subtracting each number in the zero tension
coliunn from all tlie others in the same row is shown :

ox CEItTAlN THERMOELECTllIO EFFECTS OF STRESS IN IKON. o53

Table III.

E. M. F. hrtuceii the stretched and uiiatteeted Iron AVires,
at cliosen Temperatures and at \ari(jiis Tensi(jns.

Hot Tkmi'.

Tension

' 0

Tension

2

Tension

4

Tension

6

Tension
8

Tension
9.6

Tension
0

UM/

2J2

242

249

240

27()

248

231

]-2(f

28:3

282

294

282

81(3

294

270

1 oU°

388

•sss

842

838

800

842

885

180'

887

884

392

881

410

891

396

20U°

419

428

425

416

445

424

430

230°

470

494

500

478

506

491

485

250^

515

548

557

581

556

514

529

Table IV.

E. M. I'\ helween the unstretelied and stretdied Iron ^\''ire.-
at chosen leniperatures and at various Tensions.

Hot Tejii'.

Tension

0

Tension

2

Tension

4

Tension

6

Tension

8

Tension

9.6

Tension
0

100"

0

0

7

_2

28

1

-11

120°

u

-I

11

-1

bo

11

-18

i5œ

0

— 5

4

-5

28

4

-8

180°

0

— 8

5

-6

23

4

+ 9

200"

0

+ 4

6

— 3

26

5

11

230°

0

24

30

+ 8

36

21

15

250"

0

33

42

16

41

29

14

354

0. G. KXOTT AND .s. KIMÜKA.

In the last Table we see, almost at a glance, the pr();n'i'e.s.s of thiii"'.s
as the tension increased. The graphs of Figure 11. are obtained by plot-
ting the electromotive Ibrces c<)iTesponding to one temperature in
terms of tensions. These sIkjuRI correspond in general features to tlie
cm'ves obtained by Colin and Ewing. In a very general they do s«) ;
but they are nuich more irregular. This perliaps is not surprisinu" if
Ave bear in mind the fact that each graph is made up out of as many
ditt'erent days' experiments as there are points. If we leave out of
consideration the experiment for tensicm (>, the remaining points on
each graph arrange themselves in a fairly regular manner. There does
not, however, seem to be any sufficient reason for omitting this experi-
ment. Tor the peculiar deviations of aJI the p<jints belonging to it
cannot l)e easily explained as due to any errors in reduction either to
temperature or to electrom<jti\ e force. The same pecidiarity appears if
we use the unreduced thermometric readings in drawing the curves.
Un the other hand, the galvanometer constant was almost exactly the
same day aftci' day (as may be seen from Table I), excepting for tlie
two last series (jf ex])eriments at the liighest tension and the ünal zero.

In draAving our conclusions we nuist liowever Ijear in mind the
smallness of tlie cjuantities tal)ula(ed iu Table \\ . The probable errors
of (jl)ser\ ation are of tlie order of tlie smaller cjuantities given in that
Table; so that it would be out oi' the (juesti<ni to attach anv impor-
tance to N'alues less than 5.

j\e\'ertheless, we ai'e able t<) recoi^'nize in the graphs fiu-ured a
certain ordered succession of changes ; and there can be no doubt as to
the signilicance of the values for the aftei'-strained Avire. Here Ave
have a result apparently ncAV to the subject ; we are not aAvai-e that the
possibility of such an effect has eyeu been hinted at Ijy })revious
workers. We have already expressed the nature of this result bv
saying that the Thomson Effect in an iron Avire imdergoes a permanent

ON CERTAIN THEßMOELECTRIC EFFECTS OF STRESS IN IRON. 355

change after the l(Higitudiiial tenswii lias been appHed and removed.
Tf e is tlie electromoth e force between tlie after-strained and unstrain-
ed wire, reckoned ]io^iti\'e wben the current flows from the after-strain-
ed to the unstrjiiiied through the lu^t junction, Ave ma\^ represent
the vahies in tlie l;ist cobuini of Table \\ . by tlio linear expression

e ^ + 34 - 0.21 t
where t is tlie tem])er;itui-e in Centigrade degrees, and tlie unit is 1
microvoh. 'I'he de\iati()ii of tliis straight line from the curve drawn
tln-ough the points is well within the errors of ob.servation. It would
Ije unsafe to sttach any importance to the suggestion of t^vo vertices in
the tabulated numbers, indicating two neutral points, one al)ove and
one below 160° C. (6' = 0).

Thomson, ( ohn, Ewing, and other investigators have worked
witli temperatures linver 1han the highest we used ; so that it is not
possible to make a thorough compai-ison between these earlier results
and ours. AVhere a satisfactory comparison can 1)e made there is
(■om])lete agreement. For example in Ewing's tirst set of experiinents,
tlie after-strained wire came out poî^itive to the unstrained wire with
\\\e. hot junction at 100°C. Our residt is c =- -|- 13.

In his later series of experiments Pn^fessor Ewing was concerned
wliolly with the thermoeleotric lieluniour of iron wii-e under the com-
bined influence of stress and magnetization. He kept his hot junction
at a temperature of 160°(' ; and it will be noticed that the after-strained
wire comes out ^legalivc to tlie unstrained wire. Since however no
observation is recc^i-ded for an uinuagnetized wii'c, and since Professor
Jiving hiîiiself seems disposed to regard this negatiN-e character as due
to the magnetization, it is im])ossil)le to make a satisfictorv comparisoii.
The \alues of the eleetromotive forces given l)\- him are of the s:nne
oi-dei' of (juantity as that just given.

Oui- experiments indicate a maximum current as occurring about

356 0. G. KNOTT AND S. KIMURA.

the tensi(iii ol'S, whioli corresponds to :i \on(\ of between 9 nnd 10 kilos
— a resnJt in fair agreement witli some of IVofessor Ewing's.

The 2'eneral conehision that may be deduced is that the effect of
tension on the tliermoelectric position (^f an iron wn-e is a complex
function of the temperature. Not only does the line on the ther-
moelectric diagram suffer dis])lacement u]) or down hui it also suffers
rotation. In other words the Peltier Effect and Thomson Effect are
both changed.

These results can only be regarded as preliminary. They are
sufficient to show that the method is workable, and they have a dis-
tinct value in themselves. Tt would be advisal)le to re])eat and extend
the experiments with a much more massi\e irt^n tulx' tlian that liere
used. A sniîdler gradient of tem])eratu]'e would l)e therein' obtained,
and it would not be necessary to keep tlie one end of the wire at a very
high temperature. JW such a modific:ition, mucli highei- tensions
might 1)e applied.

On some Cretaceous Fossils from Shikoku.

By
Malajiro Yokoyama.

With Plate XL.

The Cretaceous F<n-iii:i1i<)n iii Shikoku occurs in several places.
These occurrences, however, are re^trictcl to two zones, lying one on
each side of, and paridlel to, the central zone of crystalline schists,
wliicli traverses tlie island from EXE to W8W along its longitudinal
axis. The Cretaceous strata in tlie northern zone directly overlie
these schists, and f )rni a long narrow helt along the whole nortliern
coast of the island, interrupted only here and there hy alhivial flats.
]>e3^ond Shikoku, they continue on the e:ist over the southern portion
of the island of Awaji to the Katsuragi ^[ouiitains in Kii, while on
the west, vanishiiig partly uiider the sea arid partly under the voJcanic
rocks of Kyfishfi, they seem to reappear on the islands of Amakusa.
In the southern zorie, they are not so conti rmous. They rather fill up
trough-like depressions in the Palaeozoic rocks, together with some
other memhers of the ^lesozoic Grou]). Tliese depressions are known
as the Katsnragaim Basin, the MonohiUjuwa Badu^ tlie Bijöschi Basin,
and the Sakawa Basin. Ihit here also the zonal distriljution of the
Cretaceous rocks is quite evident, as these basins all lie in one straight
line parallel to the longitudin;d axis of the island.

The northern zone is essentially composed of alternatiiig layers of

sandstone and shale, for whicli complex Dr. lfarad:d' proposed the

/

1) T. ilarada. Die Japanhchen Inseln. — Eine topo(jv(t2)hisi-h.-tjt'oh>(ji^che Uehersiclit. I Lief.
Pullislted by the Imperial Geological Survey of Japan, 1890.

358

M. YOKOYAMA.

name oï I zumi- Sandstone^ from tlie prédominance of a certain greenish-
grey liard sandstone, locally kn(5wn under the name of Izumi-stone.

Fossils from this sandstone are very few. l>esides a laro*e so-
called Fucoid which occurs at several places in Sanuki, we know only
a Hdicoccras described below, and some fragments of a large Haiuitea-
like Ammonite found by ^Ir. Suzuki at Okuzure in Awaji. Harada,'^
however, mentions also some F(^raminifer:i, bivahes, and conifers as
occurring in this sandstone.

The Mesozoic Basin of the Kalsuragawd occu])ies tlie upper part of
the ri\er of the same name in Awa. Tt wns freoloofically inves-
tigated in 1883 by Mr. Y. Kikufhi, to wliom we owe tlie first dis-
covery of the Cretaceous formation in Shikoku. Here it consists of
sandstones and conglomerates, superposing tlie Jurassic plaiit-bearing
series. The sandstone is hard, tine-UTained, and when fresh oTeenish-
grey in coLnu', and has nearly the same appeai'ance as the Tzumi-stone,
while on weathering it assumes a vellowish tint. It contains shells in
great profusion, wliich hmvever belong to a very few species, aud are
mostly found as casts. They are —

Trigonia jwciUifonnis,
Trifjonia Kikueliiana,
Trigonia rot undo fa .

Mr. Kikuclii also found a fragment of an evoliite as well as of
a spirally rolled Ammonite.

The Monohegawa Basin is in Tosn. Its geological nnture is not
well known. We possess only a block of sandstone like that of tlie
Katsuragawa, (piite filled with casts of Trigonia pocilliforriiis.

The Bgöschi Basin is not far from the above, and occupies the
southern portion of Xagaoka-gori, Tosa. Here the Cretaeecnis forma-

1) Loc, cit., p. 107.

ox SOME CRETACEOUS FOSSILS FKOM SHIKOKU. 359

tioii seems to consist solely of sandstone which is us usual frey to
greyish-green, fine-grained and hard. It c(3ntains Trigonia pocilli-
foiviis and Tr. KikucJiiana in tolerable abundance. Besides, it yields
remains of many ether Lamellibranchs, some Gasteropods and Echi-
noids, whose preservation, however, is very imperfect. The rock at
Okuminodani directly overlies the Upper Jurassic C'idaris- Limestone.

Lastly, the Sakaiva Basin is situated in Takaoka-g<)ri of the same
province, alj(jut 40 Km to the west of Kyöseki. Wliat is kncjwn of it
we ow^e to the investigations of Messrs Xaumann^' and Xasa,--' the
latter of wdiom planned the geological map"^^ of the district.

The Cretaceous Formation of Saknwa is wholly composed of
sandstone, wdiich is quite similar to that of Hyôseki. On the south of
the town of Sakawa, it lies partly on the Cidaris-Limcstone, and
partly on a series of shales and sandstone, w'hich at Yoshida-Yashiki
yields some plants.^^ Xear Ochi, however, it seems to overlie directly
the Triassic sandstone of the district.

Besides Alectryonia, Lucina, Xucula, ISolen, Khynclionella and a
Scaphites-like Ammonite, Trigonia pociUiforniis Tr. Kikucliiana, and
Tr. rutundata were also obtained from the above sandstone.

From what has been said above, it will be seen that tlie number
of fossil species in the Shikoku Cretaceous is rather small ; and these,
moreover, are so imperfectly preserved that the majority of them are
indeterminable. On this account, I can describe only four species
in this paper, 'iliese four, however, are very important, as some of
them not only show the undoubted Cretaceous age of the strata con-
taining them, but at the same time, they give lis the probability that

1) Naumann u. Xeumajr. Zur Geoloiyie ii. Paläoiitohtjie von Jap'm. Dcnks. d. math.-natûrw.
Classe d. K. Akad. d. Wissens., JVien, BX. LVll, 1890.

2) T. Nasa. lîejMrt. of Geol. Sure, of Saliawa)iiu]\i, Tosa^ 1SS5 (MS).

3) Given in Harada's Japanischen Inseln, 1. c.

•i) Nithorst considers these plants as Upper Jurassic. Vide Beitr. i. Mesoz. Flora Jaj^ans.
Denks. d. Mith.-Nat. CI. d. K. Akid. d. Wissen.;. Wien, Bd. LVII, 1890.

360 M. YOKOYAMÂ.

at ]e;i8t the Irigoiiia-Samlstonc i.s to be considered as conteiiiporaiieolts
with the Gauko-Ceiioinuiiiaii Formation of Hokkaido (Ezo). Ah-eady
in my paper entitled " Versteinerungen aus der japanischen Kreide," ^'
I have mentioned the occurrence of a scabrous Trigonia, alUed to Tr.
alifonnis Park., in the Cretaceous of Kagahara Avhich I considered as
probal)]y belonging to tlie^ame epoch as that of Hokkaido. It is this
same Trigonia, 7V. jiocillifonitis as I call it, which is so profusely
found in the soutliern Z()ne of Shiki^ku, playing so to say the role of
the leading fus. 'il of the Sliik()ku Cretaceous. The above view is
moreover justified l)y the fact tliat Mr. Jimbo has recently discovered
the same form of Trigonia occurring together with Ammonites in
the Cretaceous of Hokkaido. Whether the Izumi-Saiuhtone is also to
be referred to the same age is at present unsettled, as it has not yet
given any characteristic fossils.

The two species of glabrous Trigoniœ also described l)elow are
pakeontologically very interesting. They are forms which, like some
Liassic specie-^, exhiljit a great external resemblance to tlie Triassic
genus Myophovia. Tr'ajouia Kikucltiana, whose only ally among the
Trigoniie is Tr. LuKioiicnsis Bum. of the Lias of En^-land and France,
reminds one strongly of S(jme forms of Myophoria glalnie, e.g. M. lœviga-
ta Alb. The otlier s]}ecies, Tr. rotandala, has no kindred form among
the TrigoniiL' hitherto descriljcd ; on the other hand, it has several
corresponding ones among the glabrous ^lyophoria', such as M. plcheja
Gich., M. orb:c:tI([ris Gold/., M . vuliuida Alb. In fict, this recurrence
of Myophoria-like Trigonite in the Japanese Cretaceous seems to con-
firm the view generally entertained by palaeontologists, that there is a
close relationship between these two genera.

1). Palseoatoj;iMpliica, Bd. XXXVI, 18Ü0.

ox SOME CRETACEOUS FOSSILS FKOM SHIKOKU. oÜ 1

Description of the Species.

Trigonia pocilliformis H.sp.

VI XL, Fig. la, 11), 2, 3.

Tr'ninida sj>. Yokoyama, Vei-steincruugeu ans tier Japaiiisclieu Kreide, p. 19'J.

Shell subcrescentic, very iiiequilateial, iuflated auteriorlv, at-
tenuated, iiarroAved, aiid tiattened powteiiorly. 1 leaks autero-iiiesial
tuuchiuu', pointed, much incurved and also recur\ed. The anterior
side of the valve is somewhat pr«xluced, and its margin is stron^'ly
convex, gradually passing int() the convex ventral margin which is
raised up posteriorly without any marked excavatitjn. The dorsid
margin commences at the small ligamental a[)erture behind the beak,
and descends posteriorly with a slight concavity to meet the truncated
siphonal margin nearly at a right angle. The escutcheon is lengthen-
ed, ovato-lanceolate when the valves are closed, broadest at about y., the
distance from the beak, and concave for about -/-^ the length from the
same point, beyond which it liattens. It is transversely or somewhat
obliquely costellated : the costellai are sim])le and smooth, beinf-
coarser, more elevated and distant in the p<jsterior than in the anterior
])(jrti(3n of the escutcheon. They are also sliglitlv curved out towards
the posterior side, and somewhat oblicpie, with the marginal ends
directed anteriorly. The number of these costella' is probably 15-18,
but those situated near the beak p.re so fiint that thev are hardly
visible. The area begins near the beak as a slight ridge which
gradually widens posteriorly and l^ecomes broadest at the siphonal end,
where it attains about 7^ the total heiglit of the shell, and forms at the
same time its posteriijr boi-der. It is for the greater part of its length
rendered bipartite Ijy a groove which runs a little above its median

•H 6 2 M. YoKOYAMA.

line and parallel to it ; each of the two somewhat iiiie({ual halves thus
formed is moderately convex, and marked by fine transverse plications,
some of which can become ^■ery coarse. The remaining- pcjrtion of the
valve is ornamented with coarse, elevated, slightly flexuous, crenated
ribs wh<3se number exivctly corresponds to that at' the costella} of the
escutcheon, being, so to speak, the continuations of the same, although
interrupted in their course by the intervention of the area between.
Tiiey arise at the border of the ;a-ea as narrow crenulated ridges, aiid
diverij'e in every direction, liettin"' hio-her and l)roader as thev ai)-
proach the palliai border, into which they pass over withcjut any marked
curvature. The interspaces are smooth. The palliai border is render-
ed dentate by these ribs.

The internal characters of the shell are iiot well known.

The younger specimens of this shell are a little sliorter, and the
ribs nujre straight and less in number (Fig. 2).

T have already compared this species, in the work above cited, with
Ti'ujnuia alifvriuis Parli. It is nearer t<j the varietv called altcnuata
by J^ycett (A Monograph (jf the British Fossil Trigonia,''' Xo. o, p. 117,
pi. XXA , Fig. 6) than to its typical form. v>t\\\ there are marked
ditferences between the two. The most striking lies in the ribs which,
in the Eiii>lish form, are not onlv more numerous, but also describe
concentric curves in the anterior portion of the shell, whereas in
the fbipanese, although somewliat fiexuous in themselves, they all
piss over straight to the palliai border without making any distinct
curvature, lîesides, in the former, the marginal ends of the costelliii
of the escutcheon are directed posteriorly instead of anteriorly.

A species called Tri'j mia Fi>rbesii Lycrtl (1. c. p. 122) from
^'erdachellum in Iridia seems to show similarity in the course

1) I'aheontofinijjhical Societi/, Vol. A' A' T'A', isaiied for 1875.

ON SOME CRETACEOUS FOSSILS FROM SHIKOKU. 368

of the rihs to the Jnpanese. I'ut it diifers in hnviiig n shorter shell
and a broad costellated area.

Trifjonia pociUiformis occiirs sometimes in great aliiindanre, filling'
tlie whole rock. Tt is, however, mostly preserved as casts, and even
wlien tlie shell itself is loiuid, this is so firmly attaclied to the stone that
it is impossible to isolate it without breaking it to pieces. Further-
more, these casts are often so deformed that it is difficult to get
specimens on whicli we could found a good diagnosis. The above
figures'^ were taken from gv])sum pressings of an external cast of a
young as Avell as of a full growri specimen, wlii^'h was considered as
nearly perfect in slia])e.

This species is one of the characteristic fossils of the Japanese
Cretaceous, being met with almost whereever the Cretaceous fossils
are found. In Shikoku it is to lie found at the following ])l:ices :

Tanno in the Kntsuragawn l>asin ; Söy:nna and Okumin(^d:nii in
the Kyoseki liasin ; llagino in the Monobegawa lîasin (Kamigor',
Tosa) ; Sendachino and Mirano nenr Ochi, arid Yamanokanii (Xagano)
near Sakawa, botli in the Sakawa IVasin ; Obama, Vokoliata-mura,
Agawagori, Tosa.

Outside of Shikoku, it occurs in the Sanchü P)asin. and in
Hokkaido.

Trigonia Kikuchiana n.i^p.
V\. XI., Fig. 4, Ô, 6.

Shell ovately trigonal, obli(pie, very convex, lîeaks antero-
mesial, ])rominent, incurved, and very slightly recur\ed. Anterior
margin convex, gradually passing into a less convex ventral margin

1) The teeth which would be more or less visible in the dorsal as w<;ll as iu the poïtcrior
view of this and of the following species are not shown in onr figures, as these ligures were
f^ll drawn after gypsum pressings of external casts.

304 M. YOKOYAMA.

wliich ]30steriorly meets with the nearly straight, obhqnely ascending,
si])hona] margin almost at a riglit angle, the corner being rounded.
Hinge-margin obliqnelv sleeping on the ])osterior side, and going o\ er
to the siplional margin without forming aiiv marked angle at the point
of junction. Area and escutclieon not distinctly separated, forming
one, more or less flat, surface whirli is slightly depressed along its
median liiie. Tlie other |>orti<ui of the shell, which makes an anHe of
about 90^ with tlie areal surf ice, is marked otf from the latter l)v a
rounded edge, running from the beak to tlie postero-ventral corner
and slopes to the anterior and ventral margins with a sliglit convexity.
The entire surface of the shell is smooth, except near tlie lieak where
n few coarse shallow coiicentric sulci are mostly f^urid.

The shell seems to have been moderately thick. The median
depression of the posterior surfice is more marked in the adult tlian
in the younger specimens. The interrial casts show two ftrong,
transversely striated, diveroino* teeth.

Among the Cretaceous Trigonia? there is none Avhich can be
compjired to this species. lUit in the Lias there is one, Trigomn
Liiifjoitcnsis Dum. (T.ycett, Monogr. i^f lîrit. Foss. Trigonia, Xo. 8,
p. 98, pi. X.Xn, Fig. 1-4), which shows a close affinity to it. The
latter, however, has a little In-oader shell and tlie posterior side
distinctly separated into area and escutcheon by a sharp ridge.

Trigonia KiJcnchiana, like Tr. Lingoîicnsis, is one of those forms of
Trigonia which externidly exhibit a i;reat resemblance to the older
genus MijopJioria.

It or-curs almost always as casts, and also often much distorted.
The internal mould drawn on the ])late Iras the back acciden-
tally depressed. Fig. 4 is from the largest specimen we got.
This was a liroken one, but has beeii restored iri our figure. Its
shape somewhat differs from that of Fig. 5, especially in its anterior

ox SOME CRETACEOUS FOSSILS FROM SHIKOKU. 365

margin. ]jut this difference is probably clue to the mode of preser-
vation.

Very frequent at Tanno ; also occurs at Söyama near Ryöseki,
and Yamanokami (iSTagano) near Sakawa.

Trigonia rotunciata n. sp.
PI. XL, Fig. 7, 8, 9.

Shell subor])icular, slightly broiider than high, somewhat in-
equilateral, convex. Beaks approximate, a little pushed anteriorly,
prominent, pointed, and incurved, lioth the anterior and posterior
margins convex, gradually passing into a less convex ventral margin.
Hinge- margin also arched. The escutcheon is not clearly separated
from the area, there being only a trace of a broad and flat ridge
between, running from the beak to the upper end of the posterior
margin, which makes the area slightly depressed along its median
line. The other portion of the shell is moderately convex, and sepa-
rated from the area by an olrtuse edge, and making with the latter an
angle of about 120°. The entire surface of the shell is smooth, if we
except a few coarse, shallow concentric sulci near the beak, and coarse,
concentric rugce which sometimes appear on the posterior side near
the ventral maroin.

Tn appearance of area and escutcheon this species is very similar
to the preceding one.

Among the f^rms of Trigonia hitherto described, there is none
which shows any relation to it. Among the Myophorise, however,
there are several corresponding forms which have been mentioned
before.

Like the two foregoing species, Trigonia rotundata occurs mostly

366 M. YOKOYAMA.

as casts, one of which is figured on tlie plate. It shows two strong,
striated, divero'ino; teeth.

Quite as numerous as Trig niia Kikucliiana at Tanno ; occurs also
near Sakawa at Ninomiya, Yamanukami, and Sendachino.

Helicoeeras sp.
PL XL, Fig. 10, 10a.

A fragment of the body-whorl of a snail-like Ammonite, elliptical
in section, somewhat higher than broad, and with the body-chamber
occup3àng about one half of the entire volution. The external
sculpture consists in fine, rounded, transverse ribs, slightly undulatory
in tlieir course, and weakest on the umbilical side of the whorl, where
some of them even disappear. Their number is about 50 in one
circuit. The sutures on the external side of the whorl are indistinct,
lîut as flir as they are seen in our specimen, they are deeply and much
incised, with saddles and lobes bipartite ; the siphuncle seems to lie
on the outer side, so tliat our fragment is that of a Helicoeeras which
is, at least, closely akin to Helicoeeras indicum, Stol. (Cret. Cephalo-
poda of Southern India, p. 184, pi. 86, Fig. 1-2). But as the
specimen is imperfect, its exact specific determination is not possible.

Helicoeeras indicum occurs in the Arrialoor Group of India.
. Our specimen was found in a very fine-grained shaly sandstone
of Oumi, Ouchigori, Prov. Sanuki.

pig^ 2. — Comparison Diagrams of the East-West

Component Instruments,

Pit Instr.

Jour. Se. Coll. Vol. IV. PI. XXXV.

Swf. Instr.

Fig. 3. — Earthquake of January 15th, ISST.
Easl-Wcsl Component.

Jour. So. Coll. Vol, IV. PI. XXXVI.

(Pit)

\

( Surface)

Fig. 5.— Earthquake of April 29lh, 18SS. ■""'■ «"■ ""'I- *'<"■ 'f. ''/. XXXVII.

East-West Component,

(Pit)

Fig. 7.— Earthquake of February 18th, 1880.
East-West Component.

PLATE XXXVIII.

Fig. ]. Fig. 2.

( ] ) r=.0.17, r=r.7, w = 25. (1) r=0.17, T= (f. 7, w = 2b.

(2) r = 0.24, T = 7°.2, w = 25. (2) r^0A7, T= Kf.Q, w = 2ô.

(;3) r=0.:3G, T=-^.S, w = 25. (3) r = 0.17, r= SUl", ?ü = 25.

(i) r=0.17, r=l(32r, i<; = 25.

Fig. 3. Fig. 4.

(1 ) r=0.25, T= 626°, w = 2-). (1) r-0.l7, r = 2548^ iv= 25.

(2) r = 0.20, r=l'269^ m; = 25. (2) rr^0.\7,T= 0^, iv = U2.

(;]) r-n.a5, 7-1349°, iü = 2b. (3) r=0.17, r= 583', w = 342.

Fig. 5. Fig. 6.

(1) r = 0.36, r=: 4°.3, «0 = 218. (1) r=0.17, ^=18^5, «; = 156.

(2) r=0.20, T= 22.°, iü = 218. (2) r=0.17, r= 584°, mj = 156.

(3) r = 0.20, r=16^7, ?ü-411. (3) r = 0.17, r= 260°, it'-342.

(4) /•= 0.20, r= 968°, t<;:-218. (4) /• = 0.17, 2'- 583°, w = 342.

r gives radius in mm.,

T „ permanent twist in degrees,

tu „ longitudinal stress in grm. weight.

Jour. Sc, Coll. Vol. IV. PI. XXXVIII.

—

/!>.

Ï.

60'

■■■(3)

ro'

(1)

^

^^

(1)

_5»

/

^

/^

>

F^

.i".

60'

/

/

/

S^

X

4«r

/

(

Î)/

/

'S

\

N

50'

30

//

N

1

,---1

--A

(2)

f

N

N

\"-

40' ;

/

^

(3)

\

N

\-..

1

ao

//

^/

60

80

100

>&

1

/

\^

N;

30'

/'

*o

/

r^-

^^

(1)

\

\: XS)

10

/

/

■•■■'/

7

:::^

\

K.

20'

//

20

/

»

^

(2)

w

\'

0

/

1

"^

~-»

,5

1

20

40

60

80

100

120

140

160

56

-7'^

.- ■■**

(1)

"^■

^-^

(2)

F^.

3.

/

^

k

•.,

i

/

'Jr...

N

0

1

\

\V

!

i

/

___

cu

-^

7 \

/jy.U

' ■ •■'

4or

/

r''

X

i

L.A

\\

\'\

— -

/

CS)

-^

1

*

■\

>>

1

—

S(S

/

-^

.____

1

\

X\

'

f /

/

11

A
\

\

\

3tf

y

;|

\

(

,)x

Y-

\

■ cl

/

§

\\

(2J

\

N

r—

i

10'

h\

20

iO

60

80

100

\

■■•.
\

S) ^

^^

■+.,

1/

1

\

\

\ •

0

. (4)

0

50

100

i;>o

"^ 200

—

'HO

30U

^

SäU

40«

i^--

TO

/

■• ■ ■,

\

(1)

-^^

*

>

- ---

-

60

'

Mtf.

5.

,/

''

^^^-

.......

'^h

^

^

(4)

TO'
"50^

50'

(

0,

'

'%

%•

ß.

\

X-.,

(3)

-~. \

^\

eö

/'

"^

f

/

\

\

^\

/m

k

(3)

/

/

\

30

1

/■*

k

\V,

f^^

i

'

/ /

/

K

bti

a

)

^

rrrr.

\ ! ~-*-^4ÄL

--__

/

%

1

^\^

40

20'

i^-i

>_:.

"TD'

»c-

-^-

! ' — ■

■ /

A

^

^

1

."^3).,

N

^

SO'

irf

1

1_

•<5.

^

—

0

/

'-1

^^~i~

(1)

■'-

1
1

20'

/

#

Ifl

1
1

' L_ — j — .

1 1

20

40

60

80

100

120

jfj

10

i/'

' 1
i 1

^

0

30

40

60

so

100

120

140

1
1
1

1

Jour. Sc. Coll. Vol. IV. PI. XXXIX.

-GOOOO—

/

n

Relation between. Electro-

D

motive Force and Tem-
perature, at Various
Tensions.
rii.-trniiie(l t<j strnineil
rliroiiU-h Hot 'luiictioii.

/w

-50000—

y

/

/

!

S

Afinnn

^

X

Fig.

T.

^ A

<y

</

X

-^

. Initial Zero Ti'nsir)ii.
Q Teasiou 8
O Fiufil Zero Teiisiou.

^

/.

//

g

20000

0

/

i
i
1

!

1
i

1

SO' 100" 120' 140° 160° 180° L'"'(t =

Temperature in Centi»-rade Decrees.

-'40° 'ICO" ^80' 300

■lA 25 27 21) 33 35

Thousandth parts of Electromotive Force.

:j'.t 41 41! 47 4!» 51

0

Fig.

IL

8

! \

\

1

i X

r

\

^^

N

\

C,

111(1°

J

y

y^'

}

-200°—

X' ot

/

4

)\

\

M

lau

\

\
\
\

1 180

\i

\

■>

I 1

; /

/ /

/

/

/

/

/i

/

'

/

\

i\

\ ;

\
\

( ';

/

PLATE XL.

Plate XL.

Fig. la, Ih. Trigonia pocillif<n'inis. A full grown specimen.

,,2. ,, „ A young specimen.

,,3. ,, ,, Cast, somewhat distorted.

,, 4. Trigonia Ivikucliiana. A full grown specimen,

,, bab. ,, ,, A somewhat smaller specimen. 5h,

seen from the posterior side, showing the indistinct separa-
tion of area and escutcheon.

„ 6. Trigonia Kikuchian-a. Cast, accidentally depressed on the
back.

,, 7. Trigonia rotundata. Ivight valve.

„ Sal). ,, „ Left valve of a full grown specimen

partly restored, h, seen from the posterior side.

,, 9. Trisfonia rotundata. Cast.

,, 10. Helicoceras sp.

,, 10(/. ,, ,, Transverse section.

Yoloijainft . Ci-('la<y'(>iis Fossils.

Jour. Sc. Coll. Vol. IV. PI. XL.

/■

3 ^

la

/

-afl^T^ "r-

#'^*

]0a

'^%^V

t 4 vf -^ '

yo

fe'-Ä

lllï

;?