Skip to main content

Full text of "The Journal of the College of Science, Imperial University of Tokyo, Japan = Tokyo Teikoku Daigaku kiyo. Rika"

See other formats


^ ^ IS ^ -P ,f 5 



■i- & » 

THE 



JOURNAL 



OF THE 



COLLEGE OE SCIENCE, 

IMPERIAL UNIVERSITY OF TOKYO, 

VOL. XIII. 



M -^^ -^ m i^ ^ ^\i 'ii 
PUBLISHED BY THE UNIVEPtiSlTY. 

TOKYO, JAPAN. 

1900—1901. 

MEIJl XXXIII — XXXIV. 



COJTEiNTS. 



Pt. I. Publis >eâ Ju- e 2nd, 1900. 

Notes on the Geology of i_ie j mdent Isles of Taiwan. 
By B. Koto, Ph. D. B'qoh-h' ski, Professor of Geology, 
Science College, Imperi.i " t dty, Tokyo. {With Plates 
I-V) 1 

Change of Volume and of I Iron, Steel, and Kickel 
Ovoid s by Magnetiza'' i I^agaoka, Rigakuhakushi, 
Professor of Applied Ma ' ^ and K. Honda, Rigakushi, 
Post-graduate in Physios (J> U'- Plates VI & VII) 57 

Combined Effect of Longitucinal and Circular Magnetizations 
on the Dimensions of Iron, rsteel and Nickel Tubes. 
By K. Honda, Rigakushi, Post-graduate in Physics. {With 
Plates VIII & IX) 77 

Studien über die Anpassangsfahigkeit einiger Infusorien an 
concentrirte Losungen. Von A. Yasuda, Rigakushi, Pro- 
fessor der Naturgeschichte an der Zweiten Hochschule ^u 
Sendai. {Hierzu Tafel X- XII)... 101 

Ueber die Wachsthumsbescbleuni^'iing einiger Algen und 
Pilze durch chemische Reizdc Von N. Ono, Rigakushi. 
{Hierzu Tafel X-XIII) 141 

Pt. IL Publ, hed July 25th, 1900. 

Ammonium Amidosulphite B" "F Divers and M. Ogawa. 

Iiiil)erial University, Toi •* .. 187 

Products of heating Ammoniiun Sulphites, Thio sulphate, 
and Trithionate. By E. L'ivehs and M. Ogawa, Imperial 
University, Tokyo ... .. 201 

Pottassium Nitrito-hydroxymidosulphates and the Non- 
existence of Dihyd]ozylamine Derivatives. By E. 
DivEus, M.n., D. Sc, F. K. S , Eiueritus Prof., and T. Haga, 
D. Se, F. G. S., Professor, T »kyo 'Imperial University 211 



•^»r,..f 



Identification and Constitution of Fremy's Sulphazotised 
Salts of Pottasium, his Sulphazate, Sulphazite, etc. By 
E. Divers, M.D., D. Sc, F. R. S., EiiK^ritus Prof., and T. Haga, 
D. Sc, F. C. S., Professor, Tokyo Imperial University 225 

On a Specimen of a Gigantic Hydroid, Branchiocerianthus 
imperator Allmati), found in the Sagami Sea. By M. 
MiYAJiMA, Rigakushi, Science College, Imperial University, 
To^ {With Plates XIV & XV) 235 

Mutual Relations between Torsion and Magnetization in Iron 
and Nickel Wires. By H. Nagaoka, Rigakuhakushi, Pro- 
fessor of Applied mathematics, and K. Honda, Rigakushi, 
Post-graduate in Physics. {With Plates X^VÎ) 263 

The Interaction between Sulphites and Nitrities. By E. 
Divers, M. D., D. Sc, F. R. S., Emeriti s Prof., and T. Haga, 
DTSc. F. C. S., Professor, Tokyo Impenal University 281 

Pt. III. Published Dec. 28th, 1900. 

Contributions to the Morphology of Cyclostomata. II. — The 
Development of Pronephros and Segmental Duct in Petromyzon. 
By S. Haïta, Professor in the College of Peers, Tokyo. {With 
Plates XVII-XXI) 311 

Beiträge zur Wachstumsgeschichte der Bambusgewachse. 

Von K. Shibata, Rigakushi. {Mit Tafeln XXII- XXIV) ... 427 

Decomposition of Hydroxyamidosulphates by Copper 
Sulphate. By E. Divers, M. D., D. Sc, F. R. S., Emeritus 
Prof., and T. Haga, D. Sc, F. C. S., Professor, Tokyo Imperial 
University 497 

Pt. IV. Published Oct. 5th, 1901. 

Observations on the Development, Structure and Metamor- 
phosis of Actinotrocha. Iwaji Ikeda, Rigakushi. {With 
Plates XXV-XXX) 508 



PKINTED AT THE -'TOKYO TSUKIJI TYPE FOUNDRY.' 



1.T3J 



Publishing Committee. 
^^ 

Prof. K. IVlitSUkuri, Ph. D., BigakuhakusM, Director of the Obliege 

(ex officio). 
Prof. B. Koto, Ph. D., Rigakiihakushi. 
Prof. T. Haga, RigakuhakmU. 
Prof. S. Watasé, Ph. D., Rigah'hakusM. 



AU communications relating to tliis ^'ournal should be addressed to the 
Director of ihc, CoLege of Science. 



CORRIGENDA. 



Page 



4, 
6, 
18, 
21, 
28, 
29, 
42, 
45, 
5?>, 
57, 



12th line, 
the last line, 
18th line, 
the last line, 
9ih line, 
13th line, 
2nd line, 
14th line, 
24th line, 
20th line, 



for 
for 
for 
for 
for 
for 



Gi-ô-tô 

Cholnecky 

there lies 

basite 

crystals 

Büking 



read 
read 
read 
read 
read 
read 



Gio-ô-tô- 

Cholnoky. 

the relics. 

bastite. 

crystal. 

Bucking 



„ Explanation Pts. I & II, 



delete the word ' macroscopically.' 

for ' leached ' read ' leached or percolated. 

for pyrites read pyrite. 

for griinual read granular. 

for octant read quadrant. 



jDeiwecn tne, geologically neglected, south-east coast of China 
and Taiwan, the expanse of sea is studded with a great number 
of ishinds, collectively called tlie Hoko or Pescadores Group. 
It consists of islands^ islets and rocks, great or small, altogether 
numbering ö7, besides countless hidden rocks under the water. 
The waterway on the continental side of the Pescadores is the 
shallow Fokien Strait, only a hundred miles wide, and on the 
Taiw^an side, is the still narrower Hoko Channel, — the only pas- 
sages which allow free communication to the waters of the de- 



Publis hing Com mittee. 



Notes on the Geology of the Dependent 
Isles of Taiwan. 



By 



B. Koto, Ph. D. RigahiJiakushi, 
Professor of Geology, Science College, Iiiijierial University, Tokyo. 



With Plaies I-V. 



THE HO KO GROUP (PBSCAHOIiBS). 

I. Introductory. 

Between the, geologically neglected, south-east coast of China 
and Taiwan, the expanse of sea is studded with a great numher 
of islands, collectively called the Hoko or Pescadores Group. 
It consists of islands^ islets and rocks, great or small, altogether 
numbering ô7, besides countless hidden rocks under the water. 
The waterway on the continental side of the Pescadores is the 
shallow^ Fokien Strait, only a hundred miles wide, and on the 
Taiwan side, is the still narrower Hoko Channel, — the only pas- 
sages which allow free communication to the waters of the de- 



KOTO : XOTES ON THE GEOLOGY 




pressions of the North and South China Seas. Tlie region 
is alternately suhjected to strong ebbs and floods through the 
influence of the branch currents of 
the swift Kuro-shiivo from the 
north and south, creating foamy 
and turbulent waves, in conjunctin 
with the steadily blowing heavy 
north-easters, — the dread of coast- 
ing navigators for ship- wrecks and 
other deplorable accidents. 

I have not yet had oppor- 
tunity to learn by my own inspec- 
tion the geology of the Pescadores 
Group ; but through the kind- 

FiG. 1. — Index inaj) of Taiwan to sliow 
neSS of jMeSSrS. Y. Saito and the position of the islands dosmbed. 

T. Tada, I have obtained about forty specimens of rocks, which 
no doul)t fairly represent the types that build up the crust of the 
islands. In anticipation of a fuller report by Prof. Yokoyama, 
who has made the islands the subject of his special study, I 
may give here brief notes on the descriptions of rocks and the 
inference drawn as to the probable geologic structure of this 
interesting volcanic group. 

The islands are, broadly speaking, distributed within an 
elliptical space. On the north of the Tropic of Cancer lie main- 
ly the larger islands which are arranged after the manner of 
Santorin. They resemble the latter not merely in general out- 
lines, but they owe their very existence to tlie same cause ; both 
are of volcanic origin. These Santorin-like islands are Gio-o, 
Hoko, Hakusha, and Chii-don, the latter three fuse together, 
especially duiing low tide, into one mass with the intervening 



OF THE DEPENDENT ISLES OF TAIWAX. 3 

coral-reefs wliicli stretch from one island to the other, making the 
shape very much like Thera. The single island of (rio-o, then, 
corresponds in shajoe and position to that of Tlierasia. Here, how- 
ever, we look in vain for the active centre of Kaimenis of Santorin. 
Taking into account the general distriljution of the above-men- 
tioned islands, and also the bathometrical condition, which the 
chart, Plate IV plainly shows, it is likely that they form an 
independent centre of effusion, in contrast to the Southern group 
(the Rover group), from which this Northern is separated by 
the Hover Channel, though both sit upon the eastern end of the 
so-called Formosa Bank, which stretches out hither from the 
coast of Fokien. The same type of topography seems to prevail 
throughout the whole group. It is simple, monotonous, flat- 
topped and low ; the highest prominence scarcely exceeds '16 m. 
(located at the south-west point of Gio-O), and the land can 
only be recognised from the sea witbin few miles. The islands con- 
sequently are wanting in wind-protected harbours, being constantly 
exposed to the north-east stiff gales during full three-quarters 
of a year. The land surface is bare, desolate and barren, being 
entirely destitute of green covering, due, it is said, mainly to 
the savage violence of the wind, against which even hardy 
shrubs can not maintain their footing. 

The rain-fall, which the south winds occasionally brings 
thither during the summer season, is soaked up as soon 
as it falls on the craggy ground ; and there are scarcely any 
rivulets that properly deserve the name. The erosive actions 
of running water thus become totally suspended, and valleys 
and dales are scarcely to be seen in the interior, but only the 
butte-like table-land capped with the Basalt-sheet. The deflation 
alone is instrumental in modelling the topography, and here we 



KOÏO : NOTES ON THE GEOLOGY 



have a qiiasi-deseii, and not an oasis, amidst tlic green island- 
world of South-eastern Asin. 



Forty of Mr. Tada's specimens of rocks, on wliicli I base 
my petrographical descriptions in the present paper, were collect- 
ed from the following islands : — 

1) Hoko island, the largest of the whole group. 

2) Haku-sha-to,'^ b'^"o ^loi'th to the foregoing. 
o) Tmpai-sho. 

4) Clio-sho, the eastern neighbour of Hakusha-to. 
Ô) Kippai (Bird Island of English Admirality chart), the 
northernmost of the whole group. 

6) Gi-o-to (Fisher Island), west of Ilôko-tù. 

7) Hattô-slio, lying farther to the south of the main group. 
In addition to these, I have received lately a few specimens 

collected by Mr. Y. 8aito. 

1) The worilg 'to' and 's/w' recurs frequently in tlie geographical name of Taiwan, the 
former signifying an island, the hitler an islet or rock. 



OF THE DEPENDENT ISLES OF TAIWAN. 5 

II. Stratigraphical Characteristics. 

HÔKO ISLAND. 

Hôko or Tai-saii-slio^^ is the largest among the forty-seven 
islands of the HAko group, having an area of 02.7 square kilo- 
metres. Its general outline is k-shapeel, curving in at three 
points in the coves, Fiikibir^ Giu-bo-ken/'' and KAtei.^' The re- 
lief is simple, low and flat-topped, the maximal elevation being 
Mount Tai-bu,^^' located nearly at the centre, Avilh a height of 
only 48 m. The coast is clilTy, interrupted often by sandy flats 
fringed with coral reefs. 

Idr. Y. Saito has geologically reconnoltered the principal 
islands of the group during last winter, and has kindly placed at 
my disposal the written account of his observations, which lam 
here following in hs main points. 

The island is essentially composed of the Tciilanj Basalts, 
of which three different flows, poured out after long intervals, 
are well marked by the intervening tulaccous sedimentaries of a 
considerable thickness. The topmost fliow caps the surface of 
butte-like elevations, or makes the flows of extensive ' mesas,' the 
surface being covered with its eluvial products— a line, ferrugi- 
nous loam which gradually passes downwards into a blocky 
loam and then the massive lava. The flow is rather thin, and 
characteristically columnar. It is frequently wanting in some 
parts of the island. 

In the irregularly formed strip of land— the Fiikibi-Jiri"' 

n TaUan-Hho (^f.Oj«!«), signifying 'great mountain islet,' is by no means literally true, 

thouuli uniioiibtedly it is the largest of llie whole Pescauorcs. 

-) \)iM'Ji ^!i ^"-tS-^ -i) Vila 3) :J^iitm C) 



6 KOTO : NOTES ox THE GEOLOGY 

tongue, which projects out from >Sei-shi-an^^ towards the citadel 
of Baku, thus enclosing within it a safe harbour, — we see the second 
slteet of flow, beautifull}^ exposed along the steep declivity all round 
the shore under the uppermost lava-liow, from which it is 
separated by a thin bed of tuffite. This is a most extensive and 
strong sheet, aggregating about 1() m. In its upper portion, the 
lava is porous, whitish, and much decomposed, Avhile the lower 
portion is fresh and compact. It is the one which we usually 
see along the sea-shore on whose tra})pean üoor the rollers 
break and recoil in tumultuous waves. 

The third is the lowest, consequently the oldest flow visible 
in the Pescadores, and frequently forms the floor of the coast, when 
the second sheet, already referred to, makes its appearance higher 
up the precipice. It is likewise doleritic and porous as in the 
above flow, and this Basalt is well seen at the environs of .liri, 
already referred to, where it is underlaid by a meagre lignite- 
bearing bed. It rarely happens to come to the surface not because 
of its absence but that it is hidden under the level of sea. 

Tertiary strata, often accompanied by lignite seams, occur 
inserted between the iirst and second flows, and also below the 
third sheet. An undeterminable cast of gastcropod together with 
an Area were secured by Saitû from the corresponding bed at 
Bun (Lun) point in the Island of Gio-ô. The sure proofs of their 
being of the Tertiary age are not at hand ; but from the analogy 
of the occurrences of Basalts in the neighbouring regions, I 
conjecture the sediraeutaries, here referred to, to be of later Terti- 
ar}^ According to Cholneckv"', two volcanic lines are said to be 



1) mm- 

-) ' Vurlilufiger Eeiiclite über meine Forsclumgsreibe in China.' Petcrmanns JfiUh. 45, 
1S99, S. 8. 



OF THE DEPENDENT ISLES OF TAIWAN. 7 

distinguished in Eastern Asia ; the one has served for the well- 
ing out of an enormous quantity of Basalt in later Tertiary age, the 
other has given rise to the chains of modern (Andesitic) volcanoes. 
In the north of the Chang-pei-shan, in Korea, he announced 
recently the discovery of an extensive Basaltic mesa more than 
G0,000 square km., which extends from Mukden through Kirin to 
Ninguta, forming the water-shed of the Sungari Hiver and the 
Tumen-kiang. I have been informed, verbally by j\Ir. Nishiwada, 
of the occurrence, outside of Manchuria, of a trappean plateau, 
of small extent, along the eastern water-shed of the Korean 
Peninsula, and the island of Quelpart ; and Venukoff' ^ cites a 
number of localities where Basalts make their appearance on the 
plateau of Mongolia. Furthermore, the Basalts occur sj)oradically 
in Liau-tuug, and 8hang-tung as far down as Nanking, ap- 
proximately in a straight line, and v. Eichthofen"^ brings the 
line in connection with the tectonic movement which has created 
the ' great plain ' of China, and he assigns the age of this crustal 
movement to the Tertiary period. The Basalts of the Pescadores 
seem to me to be included in this petrographical province of 
Eastern Asia. 

Since the beginning of the Diluvial epoch, a subaerial condition 
has prevailed over HAko, as well as in all the islands of the whole 
group, and erosion and disintegration have been at work, thereby 
carrying off the greater j^art of the uppermost flow, and gradual!}^ 
diminishing the area of the islands, and finally reducing them to 
ruins, as we see at present. Consequently, no record is left of 
the deposit representing this period, unless we take for it the 



1) 'Les Eoclies basaltiques de la ^Mongolie,' Bulldin de la Société be'çje de Géologie, etc., 
tome II, p. 441. 

2) 'Shantung und seine Eingangspfort Kiautschou,' 1S9S, S. GG. 



8 KOTO : NOTES ON THE GEOLOGY 

thin superficial covering of ferruginous loam wliicli is in part 
at least the product of decay of Recent epoch, thongh a certain 
portion may have been defladed away and lost during dust-storms. 
Along the shore free from escarpment, white sandy beaches 
stretch from one point to another. Tliey are the Alluvial cleposiis, 
into whose composition enters a special element which we are not 
accustomed to see in our own coast. Xearly all round the island, 
coral reefs grow npon the Basaltic shelf, and the detritns derived 
from them is driven up to form low sand dunes, leaving behind 
them, if the coast-line is deeply indented, as it is in many 
places, muddy shallows filled with the residual clay of decomposed 
Basalt. 



Such is the general outline of the geology of the Island, 
and of the rest of the group as well. 

Looking more into the details, we find that at Bako'^ 
Point, on which is situated the town of the same name, the second 
flow extends in a great sheet, covering all but a few points of eleva- 
tion which are capped with there lies of the young columnar lava, 
being separated from it by a blue rock. The last is a fuller s 
earth, which is a bluish-grey, dull, compact mass of greasy 
lustre, splitting, when dry, into angular clods with sub-conchoidal 
fracture. It adheres to the tongue, and falls readily to a muddy 
state on placing in water, and is not plastic. Under the micro- 
scope, the whole mass consists of brownish, double- refracting 
particles, and seems to have been derived from the decomposition 
of a Basaltic glass. It crops out for a short distance, and on shore 
a poor bed of Tertiary lignite occurs associated with it. 



OF THE DEPENDENT ISLES OF TAIWAX. 



9 



The same state of things prevails throughout the tract 
southward as far as Sei-shi-an", the surface being covered with 
thick ferruginous loam mixed up with Basaltic fragments, and 
the upper and middle flows coming in direct contact, distinguish- 




Fig. 2. — Isolated erosion Lill Slia-bö-san, iienr Jiri, sliowiuj two uppei' flows with 
iutirbediled sediuiontaiios.* 



able only in the difference of structures. At the last-mentioned 
locality, a ' haul-over ' of base-levelled middle flow, masked with 
coral sand, separates the tongue of land Jiri-\ on which stands 
a Basaltic, hat-shaped Bha-bü-sau"\ 47 m. high (Fig. 2). 

A good section may be seen along the shore, west of Jiri, 
as is shown in Fig. 3. The columnar, upper (No. 1), and doleritic, 
jjorous middle flows (Xo. 2), aggregating about 6 m., cap the 
cliff, 20 feet high. That the two flows are separated by long 
time intervals can be clearly shown elsewhere (Fig. 2) by a bed 
inserted between them. I may cite the case of a lignite bed at 
Baku, occurring in company with fuller's earth. Another instance 
mav be given of it just east of Jiri, where an ash l)ed makes 
its appearance. This ash bed is a fine, greyish-white, pulverent 



*A11 the figures in the following wood-cuts, not otherwise mentioned, are originally 
sketched hv Y. Saito. 



10 



KOTO : XOTES OX THE GEOLOGY 




Banded fclsimr 
sands and clays. 

No. 3. Basalt. 

Sandy c-lay with 
lignite. 

Fi(.-. 3.— Section expcsed at the. west coast 
of Jiii, Hôko. 




So. 1. Basalt. 



Basalt. 



Felspar sand witli 
linionite r.olulcs. 



oartli, wholly consisting of the microscopic particles of plagio- 
clase, a few fragments of pleochroic Jiypersthene, and little 
magnetite, but no glass splinters are seen. It reminds me of 

the felspar sand that cover the 
flat and form the ground of 
Pampanga, north of Manila^\ 
After this short digression, I re- 
turn to the former subject. Xow, 
a yellowish-brown, loose sandy 
bed, -3 m. thick, comes below the 
middle flow, locally with liraonitic 
nodules (Fig. o) . This is succeed- 
ed by another complex bed, 3 to 
4 m. thick, made up of multi- 
farious alternations of clays and 
sands, all retaining the original 
horizontal position. Then comes 
the third sheet of porous lava 
of variable thickness, underlaid by a lignite bed, the last one 
can be only seen at low tide. The whole seems to me to be 
one complex bed belonging to later Teritary ; and this profile 
serves as a type of the stratigraphical order of the island. After 
passing over the second ' haul-over ' to the Fûkibi point {Plate V), 
opposite to Baku, nothing but the two upper flows is exposed. 

A table island, named Ko-sei-sliO"',off tlie coast of Jiri, 
already referred to, is an erosion relic of the Basaltic mesa, 
surely connected in former times with the main island of Hoko. 
The adjoining wood- cut shows clearly the geological structure 

1) B. Koto, 'Geologic Strnctuie of the Malayan Arcliipehigo.' Tliis Jouinal, Vol. XI-, 
p. 113. 

2) 



OF THE DEPEXDEXT ISLES OF TAlWAX. 11 

and the general view, as seen from Jiri, exhibiting the two 
upper flows, mainly hidden by debris cones. This ishmd served 
for tlie Chinese in former times for the strateiiic base airainst 




Fig. 4. — A view of the Isle of Ko-soi-sl)o, an erosion relie of Basaltic mesa, 
as seen fi'om the coast of Jiri. 



the Hollanders and Koku-sen-ya (Koxinga), in maintaining the 
sovereignty over her supposed vassal domain of Taiwan. 

Starting again from Sei-shi-au, already referred to, and 
going round the south coast along the points of Kan-on-san" 
and Kô-kaku^\ Basaltic cliffs with underlvino- sandv bed, and 
sandy coves repeatedly occur as far as A-kau"'. At Sa-kan", a 
little south of the last-mentioned locality, fuller's earth similar 
to that of Bako, is said to occur according to Tada and Ishii. 
Upon the walls of the cliff at the recesses of the coves arc found, 
attached, according to Saito, apparently recent shells, telling the fact 
that at no geologically remote period, probably Diluvial, a negative 
shifting of sea-level has taken place in this tract. We are, 
however, not informed of the height of the former level, as 
compared with the present ; but at any rate it is of paramount 
importance for us to have been acquainted with this movement 
in view of the fact that on the opposite coast, i.e. on Front 
Taiwan, there are not wanting evidences tending to prove the 
neo-ative chau2;e on the shore. 



1) mtOi 2) t56Ä 3) %^ 4) iit. 



12 KOÏÔ : NOTES ox THE GEOLOGY 

Between A-kaii and Ki-sei-ktiku'', the easternmost point of 
the island, a white sandy beach bounds the south shore. 

All along the coast from l\i-sci-kaka to PIoku-ryo'-\ coral 
reefs limit the eastern shore, and the detritals derived from them 
form the beach-flat. It is a noteworthy fact that on the north side 
the coast is very deeply indented in the norlh-south direction, and 
the lowland, partly marshy, is covered likewise with coral sand. 
I may here mention an occurrence of coal which was once con- 
sidered to be a very important natural resource of the island, 
though afterwards it turned out to be almost worthless and unworthy 
of public attention. At one of the points, called Koto or dragon 
head, that stretches out northwards, a butte of Basalt, 22 m. 
high, elevates itself from the shore, and at its northern foot 
a seam of lignite, Ö feet thick, crops out with a sandy rock be- 
tween the first and second flows, corresponding to the Area /one 
in Gio-o Island, already referred to. The exposure is meagre and 
soon disappears under the rubbish to be seen no more. This 
mineral combustible is but imperfectly incarbonized, and the 
woody structure is said to be yet w^ell preserved. 

From Sei-kei-'^ through KO-tei'*, and Sha-ko'^ as far west as 
to the oft-mentioned Bako, along the north coast, the two upper 
flows are the sole rocks that can be seen, being covered with 
an incoherent brownish, coarse and craggy earth. 

HAKU-SHA ISLAND. 

Haku-sha-to,'^^ or the white sand island is bodily connected 
with Hoko through the intervening islet of Chii-don'^, at the 



1) WilEn 2) ^if- 8) m% 4) jfig: 5) WM 6) âîJJ 



OF THE DEPENDENT ISLES OF TAIWAN. 13 

two iiarj'ow iiecks of tlie aljracled second flow of Basalt, ami 
forms a part of the geolo^i;'ical iinii, (littering from tliem only in 
that here the interstratified sedinienlaries seem to Le wantinir. 
The other features that strike the eyes of observers are firstly, 
the lowness of its relief, the highest point being Ivô-don-san'\ 
36 m. high, and secondly, a considerable development of Allnvial 
accumulation of the shells and skeletons of low organisms, hence 
the name of the island. Cliffs, however, can be seen in its northern 
shore, exposing the youngest flow with its usual columnar ëtruciuve 
at the water's edge. White sand}^ flats prevail throughout the rest 
of the lonely island, especially towards the Bay of Hoko, and 
the residual product of considerable thickness, derived from 
the Basaltic decomiposition, covers the interior. 

One thing worthy of mentioniiig is a sporadic occurrence 
of lapilli that had run aground on the east shore, probably 
from one of the Indonesian volcanoes. The j^umiceous fragments, 
worn and rounded, belong to a Hypersthene-andesite with a 
highly pleochroie, rhombic augite, and this rock either massive 
or pumiceous can be seen in no other 2'>ai"ts of the group. 

The islets, Tn^pai'-' and Chô-sho"^ or Bird Island, off the 
east coast, seem to be geologically identical, representing the 
erosion-relics of the Diluvial epoch. A luxuriant growth of coral 
reefs fringes the latter, as well as the neighbouring islets, just 
as in Haku-sha, 

KIPPAI ISLAND. 

Farther away in a northerly direction lies the islet of 
Kippai^*, which is a low Basaltic flat, covered with half- 



14 KOTO : KOTES ox THE GEOLOGY 

hardened foramiiiifer sand {PL II, Fig. 0.) of Eecent age ; frag- 
ments of corals, bivalves and serpula mixed witli otlier components. 
The foraminiferal rock consists of millions of discoidal and 
spiral, water-worn shells. Karely they have spines well-preserved. 
Viewing a section of the shell under the microscope, it is seen 
that the test consists of the tubulated proper walls of chambers, 
besides the canaliculated intermediate skeleton which forms 
spur-like marginal appendages, characteristic of Calcari/ia, and 
its external form and microscopic details agree well with C. 
Spenglei'i, Linné'\ dredged for the first time near the coast 
of Amboina at the depth of 1,42.") fathoms. This species seems 
to be cpiite as abundant in the East Indian Archipelago, as we 
find here in the Pescadores. By wear and tear of rolling waves, 
the surface of the test becomes smooth, and the presence of spines 
can be usuallv only recoonized in examinins; the structure of the 
supplementary skeleton which points to the former existence of 
some sort of prominence. 

GIO-Ô ISLAND. 

Gio-o, or Fisher Island, lies to the west of Hôko, and 
encloses with the latter the head-less Bay of Hôko, or rather 
an arm of sea. What has been said of other islands as regards 
the geology and the topography, holds true also of Gio-o, with 
the differences, that the island is really table-shaped, bounded 
on all sides by cliffs, leaving no space for .Vlluvial deposits, 
excepting the shore and fringing reefs ; and that the igneous 
sheet as well as the interbedded sedimentaries are developed to 
their full advantage, thus affording the best opportunities for geolo- 

1) Chullcn-'jor Report, ' Foraminilera.' 



OF THE DEPENDENT ISLES OF TAIWAN. 



15 




No. 1. Ea--alt. 



mdy tuflite. 



gists to get insight into the geological structure, and to study 
the stratigraphie details, of the whole Pescadores. 

The oft-mentioned three flows and interstratified tuffites as 
well as the underlying bed are likewise present, and well seen 

especially in that portion that lies 



southwards of 8hô-chi-kaku.^' 
Between the last-mentioned 
locality that is situated in the 
middle of the island and Shii- 
ba-wan,'-' good sections may be 
traced, as in fig. 5, in descending 
order. Under the superficial 
covering of the ferruginous soil 
of decomposition from Basalt 
comes the No. 1.^ Basalt- flow, 
with its usual columnar struc- 
ture, of about one foot, and 
sometimes disappears altogether. 
The third in the series 
consists of pelitic sand and 
loose sandstone, the latter being 
made up of musooviie, plagioclase, 
and Basalt-glass. Concentric 
nodules of hematite are frequent- 
ly found in them. Saito is 



No. 2. rasait. 



^ . XX >- ^^ .> 




1 el par saul. 



No. .3. Basalt. 



gaiitlv c-lav. 



Fin. ,5. — General profile as seen in the southern 
tiart of Gic-ô. 



fortunate enough to find in this complex bed casts of an Ai^ca and 
gasteropod {Turbo) in the matrix of ferruginous fels]3ar sand 
wdth a little magnetite. Judging from the cast, the shell of the 



1) ^hmn 



]G KOTO : NOTES OX TEE GEOLOGY 

Area is tliick, egg-sliaped, tlie ends of the margin obtuse-angled ; 
the margin anteriorly rounded, j^osteriorly sloping; the beak 
prominent, anteriorly inclined, widely separated and inflated ; 
coarse radial ribs more than 20 in number. Our specimen 
apparently resembles A. suhcrenata, Lischke, though in details 
they may differ, if perfect samples are taken in comparison. 

The next in the series is the porous, No, 2. sheet, under- 
laid by a fine felspar sand bed. Then the lowest, jN'o. 3. sheet 
of G-7 feet, often Agglomeratic ; and lastly, the bluish-grey 
sandy clay, consisting of clay, miiscovite, plagioclase and brownish 
opaque grains probably of Basaltic glass together with carbona- 
ceous matter. It is remarkable that muscovite is more or less 
intermixed with in all the sedimen taries. 

Before quitting Gio-o, it should be remarked that the area 
north of 8ho-clii-kaku, as well as the whole east coast is com- 
posed of the two upper flows only with or without interstratified 
beds ; while the rest of the island, as may be seen in fig. ö, are 
built up of the second and third flows, accompanied with sedi- 
nientaries, unsurpassed in complexit}'- and in thickness. 

According to Tada, the islands of the Southern Grouj) {PL 
IV.) of the Pescadores, are geologically of the same type. 
Counting southwards, they are : — Hatto,^^ with the dependent 
isle of Sho-gun-o'-^ ; the Smaller and the Larger Biü-sho"'\ so 
named cat islands from their appearance as seen from a dis- 
tance ; Tai-sho'^ and Sho-hei"' with columnar Basalt; To-kitsu''* 
and Hei-kitsu'\ likewise Basaltic ; all being encircled by coral 
reefs. 



1) A? 2) )ifip:^ 3) 3Stf| 4) ±m 5) /h^ 6) %] 



OF THE DEPENDENT ISLES OF TAIWAN. 17 



III. Petrography of the EfFusives. 

The groundwork of the Pescadores is essentially built up of 
Basalts, making extensive flows to the water's edge, and the whole 
is encircled by the fringing reefs of cornls, which, in parts raised 
above the water, connect many of the detached rocks with the shore, 
thus contributing greatly to the enlargement of the areas of the 
island. Each and every island visited by Tada and Saito, presents 
the same physiognomy, and consists of the same black rock. The 
specimens, brought back from most of the islands, and of which 
descriptions will be given in the sequel, have a certain common 
feature which stamps them as genetically identical, and their 
field relations in different areas seem to point to a common 
centre of volcanic activity. They exhibit, however, a considerable 
variation of character. Thus from the same island, T have 
specimens at one place perfectly massive and compact, at another 
vesicular and porous, and sometimes Doleritic. Colours var}^ 
from black to bluish-grey in fresh ones, and through weathering 
the Doleritic and vesicular varieties become whitish or grey, 
while the compact rocks acquire a reddish brown tinge. 

AVe are indebted to Mr. Y. Saito, for characterising the 
different flows, and for tracing their vertical as well as horizontal 
distributions in the Northern group. According to him, there 
are three distinct Basaltic flows of nearly the same distribution, 
separated by long time-intervals which are represented in inter- 
bedded sedimentary rocks. Judging from the nearly perfect 
horizontality which the beds and flows keep in all the islands, 
it seems probable that there existed a lava field or volcanic mesa 
of considerable extent. But, on account of its remote age, pro- 



18 KOTO : NOTES OX THE GEOLOGY 

bably later Tertiary, and of its insular position, waves gnawed 
the ground in time, finally reducing the once wide volcanic 
field into the ruins of islands, as we see at present. Tt is not 
easy to know the former extent, and the ancient surface 
feature, of this lava-flat ; but, generally speaking, the relief 
becomes higher as we go southwards from one to the other 
in the islands of the jS"orthern group. 8aitô recognises, as I 
have already said, three lava-flows in the Northern group, 
viz., the uppermost or youngest being of columnar, the middle 
porous and vesicular, and the lowest also partly vesicular, and 
Agglomeratic. After the comparative study of the Basaltic rocks, 
to which the eflusives exclusively belong, several important facts 
are brought to light, and now I am able to say, that the young- 
est flow^ (a, b and ? d types) contains the iddingsitized olivine, at 
least in one type, and violet brown titan-augite ; the second 
(c and ? d types) the brown augite, olivine sometimes lacking, 
being often replaced by hypersthene ; and the third (e type) the 
analcime-bearing. I w'ill record first my observations on the 
component-minerals, and then giye the special descriptio?i of rocks. 



A. Component-minerals of Basalts. 

OLIVINE. 

Olivine is rarely automorphic, but mostly xenomorphic, 
being the remains of resorption by magma. The olivine in the 
Basalts of the Hôko Islands seems to be of several varieties. 
Automorphic ones show^ vivid polarisation-colours, and alter 
usually into some red minerals. The xenomorphic type shows com- 
paratively a low degree of polarisation, and suffers deep corrosion, 



OF THE DEPENDENT ISLES OF TAIWAN. 19 

often being reduced to a mere grain, and is also traversed with 
fracturai lines, from which the mineral begins to form a ser- 
pen tinous substance. The olivines are undoubtedly the intra- 
telluric products, being sometimes enclosed by an automorphic 
augite, and large individuals are habitually surrounded by 
heaps of the crystals of augite (in Andésites, instead of it often 
hypersthene). Inclusions of gas and liquid are not rare, and 
the octohedra of magnetite are also found in the olivines. 

Zonal structure of olivine is, as is well known, of rare 
occurrence, and if it really exist, this could only be discerned 
either by measuring the optical angles at different portions, or 
by finding the altered zones in a crystal in consequence of the 
formation of the minercd rouge. The zone of the red mineral is 
not constant in position, for, it makes its appearance sometimes 
on the periphery, at other times in the interior; but, so far as 
my experience goes, the recurrent zones are never found. Tlie 
condition under which the isomorphic shells of different chemical 
compounds are formed in the olivine, seems to depend, as 
Lagorio^^ and Morozewics'-^ say, mainly on the Blasse moirkiing, 
that is, the degree of saturation of magma in certain temj)erature 
and 2^1'essure. In my slide, in which olivine has a red central 
zone (the Kippai Island specimen), magnetite is scarce, and large 
in its size and rod-shaped ; while the magnetite-rich rock (the Hoko 
specimen) has an olivine with an external red zone. Here the 
magnetite occurs in small isometric crystals and grains. 

The red mineral, that forms the periphery (PL I. Figs. 1, 
4 and 5) and the kernel {PL I. Ing. 6), differs in habit. The 



1) 'Leber die Natur der Glasbasis, sowie der Krystallisationsvorgänge im eruptiven 
Magma.' T. M. M. Bd. VIII, ISST. 

2) Jbld. Bd. XVIII, 1898. 



20 KOTO : NOTES ON THE GEOLOGY 

firsi lias a facile cleavage but brittle, and consequently becomes 
lamellar like a brittle micaceous mineral. It is probably identical 
with the so-called biotite, which we occasionally find mentioned 
in petrological literature, as being formed from an alteration of 
olivine, just like as schillerspar has been considered to be a mica, 
us an alteration-product of enstatite. Recently, Iddings^^ and 
Lawson-^ described a similar mineral and the latter author 
named it iddingsite. In the second, we fail to find such a 
distinct cleavage, and it seems to me to be the same body which 
Michel-Levy called the mineral rouge''. Now, a question suggests 
itself to me, whether the red micaceous mineral is identical with 
the mineral rouge or not ? It is true that the former confines 
itself to the margin, and in the case where the entire substance of 
olivine has been transformed to this mineral, the process of 
alteration has started from the periphery, and it not infrequently 
happened to me to find every stage of progress from the very 
beginning to the complete alteration. The latter, on the contrary, 
starts from the centre in irregular patches, and gradually attacks 
the whole body but the clear and granulated, thin margin. The 
formation of the red lamellœ begins with the development of a 
fine parting which appears like stripes, and which runs parallel 
to the vertical axis [PL I. Fig. 1) ; while cracks on the margin 
favour the olivine being changed into the red mineral in the 
centre. 

In my opinion, there may be a slight diöerence in the 

1) U. S. Geol. Surv., ' ■Münograpli ' XX., p, ;588. Iddings ideiitiiies this iiiincral to 
tlierniüphyllite, a fi)liated mineral liaving tlic composition of serpentine. 

2) ' The Geology of Carmelo Bay.' Bulletin of the Department of Geology in tlie 
University of California, Vol. I., p. ol. See also Pirsson's paj^er. Amer. Journ. Sei., XLV, 
1893, p. 381. 

3) ' La Chaîne des Buys et le Mont Lore,' Bull. Géol. Soc France, 3me Serie, XVIII, 
1890. 



OF THE DEENDENT ISLES OF TAIWAN. 21 

chemical composition of the two alteration-products, yet on the 
whole they must be practically identical. The lamellœ are 
oriented parallel to one of the pinacoids, as may be deduced from 
the position of the optic plane (in the fresh substance of olivine), 
which stands at right-angles to the easy cleavage {PL I. Figs. 1 
and Ö). Pleochroism is distinct; it is brownish-green in the 
direction of facile cleavage, but greenish-brown when at right- 
angles to it. Hence, oa or b. MiiggV, however, sa^-s that 
the absorption is stronger m the direction perpendicular to the 
' Längsrichtung ' than in that parallel to it. ZirkeP and Kosen- 
buch^^ interpret the above statement in the terms, that the rays 
vibrating parallel to c absorb far less than those parallel to a 
and b. The observers, however, seem to have examined the 
mineral rouge. My observation, therefore, accords well with that 
made by Lawson for iddingsite ; but it is not known to which 
pinacoid, 010 or 100, the lamelke are parallel, though it is 
probable that the brachypinacoid is the lamellar plane, as may 
be inferred from the fact that the elasticity perpendicular to 
the lamellae is greater {^i = b) than that parallel to the c-axis, the 
latter corresponding to the mean axis of elasticity. 

With HCl, the iddingsitic mineral becomes bleached, and 
then acquires a greenish-yellow colour, with corresponding decrease 
of pleochroism. Bearing in mind the fact of the brachypina- 
coidal lamellar cleavage, of the colour, and of the chemical com- 
position which is a hydrous non-aluminous silicate of iron, lime, 
magnesia, and soda, / am rather inclined to consider the iddingsite 
to be a mineral approaching to basite. Prof. E.osenbusch^^, 

1) Neues Jahrbuch, 1883, II, S. 205. 

2) 'Pétrographie/ Bd. I, 1893, S. 353. 

3) ' Phjsiographie,' Bd. I., 1892, S. 4G9. 

4) ' Physiogmphie,' 1892, Ed. I., S. 461. 



22 KOTO : NOTES ON THE GEOLOGY 

in sj^eaking of bastite, says ' die Umbildung scheint in 
hohem Grade durch die gleichzeitige Anwesenheit des Olivin 
und dessen Umwandlung zu Serpentine befördert zu werden.' 
Chemically speaking, there exists a close resemblance between 
iddingsite and the * crystallised diallage ' of Baste^', considering 
out of question a trace of alumina. Optical schemes differ, of 
course, in the two minerals, but I could not make out surely 
the optical orientation of iddingsite in my slides, on account 
of its extremely line lamellar structure. 

PLAGIOCLASE. 

Plagioclase has, generally speaking, crystallised out in a 
single generation of the flow period. Diflering from the Ande- 
sitic plagioclases which present various dimensions, the felspar 
of Basalt is uniform in size. It is, however, not wanting in 
large, phenocrystic crystals in some slides, which also belong 
to the products of the effusive period, slightly earlier in crys- 
tallisation than the ones in the general mass ; for, the small 
laths of plagioclase ai-e partly embraced by the phenocr3'sts, — 
a fact which also leads me suppose that the plagioclases have 
grown in a comparatively motionless magma. They show no 
signs of corrosion, so common in the olivine of the intratelluric 
origin, though the effects of tossing and fracturing of crystals 
are by no means seldom observed. 

The phenocrystic plagioclase {PI. II, Fig. 2) has a tabular 
form on M, somewhat elongated towards the vertical axis. Zonal 
structure is rare in contrast to the Andesitic felspar ; the same 

1) Hintze, 'Handbuch dei- Mineralogie,' Bd. U., S. 972. 



OF THE DEPENDENT ISLES OF TAIWAN. 23 

is the case of glass-enclosures, save that ground-mass which fills 
up the rectangular space between the lamellae, showing as if the 
larger crystals have grown out by the apposition of numerous 
flowing lamellae. Penck^^ holds the same view, as given here. 
In consequence of this lamellar composition, which is one of 
the causes of the paucity of inclusions, both terminations of the 
ledges became indented and forked, after the manner of a 
parapet {PL I, Figs. 4 and 6., PI. II, Fig. o), a characteristic com- 
mon to all the plagioclases of Basalts. It seems more rea- 
sonable to consider these monstrosities as incipient forms of groicth, 
having simultaneously many centres of crystallisation in space, 
which in later stages have groivn together to make up one in- 
dividual unth but internal complex compositions. Morphological 
and optical homogeneities are, however, frequently disturbed 
through the flowing motion and sudden cooling of the consolidat- 
ing magma. Several stages of similar kind in crystallisation 
may be frequently observed under the microscope in the forma- 
tion of artificial crystals. ' 

Symmetrical but contrary extinction takes place at the 
maximum angle of 33° — 3ö°, with reference to the suture of the 
albite-twinning, and the extinction with regard to the pericline- 
lamellae amounts to —16°, showing that the plagioclase is of a 
basic labradorite. It is easily acted on by HCl. The Baveno 
twins were once observed. 

The plagioclase in the ground-mass is lath-shaped, extremely 
slender, and polysynthetic ; termination being also a parapet-like. 
The habit of crystals is prismatic, and such a form is said to 
be elongated parallel to the a-axis. This is indeed true ; for, 

1) ' Studien ueber lockere vulcanische Answüröinge.' Zeit-fhr. d. d. geol. Gesdl. Bd. 
XXX., S. 101. Taf. v., Figs. 3, 5, and 7. 



24 KOTO : XOTES ox THE GEOLOGY 

along tlie longest extension lies the axis of greatest elasticity, 
and there are tens of thonsands of laths visible in microscope 
slides, with bnt a few tabular sections. Symmetrically opposite 
extinctions make the maximum angles of 23° to 25° with the 
suture of lamellae, Microlithic sections twinned on the albite 
type extinguish at the angles from 0° to 26°, with reference to 
the longer dimension. According to Michel-Levy, labradorite 
and albite have similar optical deportment, but as they do not 
usually come together, and as we are dealing now with a basic 
rock, the nature of the microlite should be considered to aj)proach 
that of an acidic labradorite. In a few slides, the poles of 
the laths resolve themselves into a number of prisms, and such 
fine slender needles are scattered through the whole groundmass. 



AÜGITE. 

Augite, so says Morozewicz^^ belongs to one of the * t'er- 
hangniss'vollen minerals. It does not obey Fouque and M. -Levy's 
rule of crystallisation of silicates in the reversed order of 
fusibility, nor Rosenbusch's scheme of crystallisation according 
to acidity. It is rather subjected to the influence of masses, i.e. 
a degree of saturation under certain temperature and pressure. 
Under such circumstances, augite may form crystals before 
plagioclases, and at other times, just the reverse may occur, 
while in the third case they may individualise at the same time. 
According to the priority of secretion of either of the two 
minerals, in other words, the relative idiomorphism of one to 

]) 'Experimentelle Untersuchungen] uel>er die Eiklung der Minerale im Magma.' 
Tschermak'.^ Milth., Bd. XVIII., S. 84. 



OF THE DEPENDENT ISLES OF TAIWAN. 25 

anotLer, various structures may be brought about under vary- 
ing conditions, and these we find in fact in my slides. 

As regards the forms of augite, it is sometimes idiomorphic, 
bounded by faces x P co^ cc Px , x P, and Px , the first being well 
developed, consequently the crystals become tabular ; at other 
times granular, needle-shaped, in ophitic plate, and in partial 
crystals. Prevailing colour is either violet or yellowish-brown. 
It is to be expressedly remarked that the typical Basaltic augite 
with a tinge of violet occurs only in the Pescadores, and in the 
dykes of Basalt near Taihoku and Taikokan, in Formosa. My long 
experience forced me to conclude that, in Japan proper, the 
Basalt with the violet augite is confined to the northern Kiu- 
shiu, and Chiu-goku, in Hondo, as far east as the provincial 
boundary of Tajima and Tamba. The same type of Basalt is 
also known to be wide-spread in Korea, Liau-tung, and Mongolia. 
Thus the distribution of the Basalt ivith the violet titaniferous 
augite marhs a definite area, being, so far as my knowledge goes, 
confined to the inner side of the festoon islands and the adjoiîiing 
continent in Eastern Asia, constituting the well-defined Ja'pan- 
China petrographical iwovince. Larger crystals show a zonal 
structure, coloured intensely on the periphery, and the hour- 
glass structure occurs frequently with deeply- coloured, additive 
cones in the prismatic zones, which have at the same time a 
greater angle of extinction. Pleochroism is stronger in the 
direction parallel to the c-axis. Polarization-colours are generally 
weak in comparison to those of the Andesitic augite. Twins on 
ooPö) have a suture, running just along the middle of the body 
of the crystals. Crystals often form stellar aggregates ; they are 
generally free from foreign interpositions, excepting the larger 
ones which have sometimes enclosures of glass and magnetite. 



26 KOTO : NOTES ON THE GEOLOGY 



UYPEESTHENE. 



Hypersthene takes the place of olivine in some Basalts of 
the Pescadores ; consequently the presence of one totally excludes 
that of the other, — a state of thing quite exceptional to the modern 
Japanese Andésite of a glassy, black, porphyritic type, in whicli 
both minerals api^ear always concomitantly. We have then the 
Hypersthene-Basalt, in lieu of the Basalt proper. It is a note- 
worthy fact that this stray variety of rock seems to be wide- 
spread, at least in my specimens, in the out of the way islets, 
such as Irapai, Kin-sho, and Hatto, the only exception being 
the one from Sei-kei (West Valley) in Hoko, though I could 
not find a sufficient reason accounting for the special dis- 
tribution of this hypersthene-bearing rock. 

It is usually a comparatively easy task to discriminate 
hypersthene from olivine, but in the present case some difficulty 
is experienced in making out for certain the presence of the 
former. 

In regard to the form, the (1) hypersthene is extremely 
slender, being about six times longer than broad, and, as being 
of the intratelluric origin, it lias a marginal zone deeply corroded 
and partly granulated, and has indefinite faces at the poles 
of the crystals (PI. II, Fig. 3). I observed once a morphotropic 
growth of a highly-polarising, mouoclinic pyroxene around a 
hypersthene, just as is the case in Andésites. Cleavage is 
developed along the longest extension of the crystals. In a patch 
of a coarse aggregate which appears as an endogeneous or homoge- 
neous enclosure in the finer general mass, the (2) hypersthene 
comes together with plagioclase and augite, and in this case the 
hypersthene occurs in broad plates [PL II, Fig. 4), with only a few 



OF THE DEPENDENT ISLES OF TAIWAN. 27 

traces of cleavage, but with uumerous fissures ; and has an ap- 
pearance exactly like olivine. 

The hypersthene possesses a brown colour, and its pleochroism 
is scarcely discernible. In favourable cases, the ray vibrating 
parallel to the c-axis is slightly green. Sections present a rough 
surface, owing to its having a high index, approaching to that 
of olivine; its polarisation- colour is grey. 

From the brief diagnosis, given above, of the hypersthene, 
its cleavage, colour, non-pleochroism or very weak if present, high 
index, but low magnitude of refraction, extinction-direction, and 
similar chemical composition, — these several physical properties 
afford no means of discriminating it from a fresh olivine. Olivine 
has, however, a lighter colour, and has usually but one trace of 
cleavage in a section. The hy23ersthene on the other hand pos- 
sesses the characteristic traces of prismatic cleavage, which in 
a random section gives scarcely a clue to distinguish it conosco- 
pically from monoclinic pyroxene. A basal section, once observed, 
presented a square outline, truncated little at the four corners. 

From the combined evidence of more slender section, of the 
want of decomposition-products, of indifferent behaviour towards 
common acids, of the presence of comparatively numerous 
traces of cleavage, I infer, in the Basalts, the presence of a hypers- 
thene. It is to be remembered that the prismatic sections of 
olivine show also a low colour of polarisation, exactly like that 
of a hypersthene. It seems to me that the hypersthene in 
the Hoko Basalts stands in its chemical composition near to 
that of bronzite. The want of a distinct pleochroism may be 
attributed to the same cause. Axial angles, therefore, become 
large, and the axial poles were not observed in any of the 
pinacoids by ordinary methods. 



28 iiOTÔ : notes on the geolog"ï- 

APATITE. 

Apatite occurs in the Doleritic or Anamesitic rocks iii the 
form of extremely fine needles, devoid of terminal faces, 
being colourless, and always traversed by transversal fissures. 
Its crystals sink almost to a minimum size, and are not, com- 
paratively speaking, so large as those found in the typical 
European Dolerites ; and for this reason they might be easily 
mistaken for the microlites of felspar which often resolves from 
the poles of a larger crystals in Basalts. The apatite is typi- 
cally found in the three slides only, which are in my pos- 
sesion (Kippai and Hoko), and both are magnetite (not ilmenite)- 
bearing rocks. The crystals are dark-margined, owing to the 
total reflection of light on the prismatic faces ; and sometimes a 
single brown-coloured axis entirely or partially runs through the 
crystal. A grey or light-brown variety, so often found in An- 
désites, is entirely absent, though a dark-brown crystal of an 
apatite-like mineral was once observed with strong absorj)tion 
parallel to the prismatic axis. The sure criterion of the presence 
of apatite can only be found in its hexagonal cross-section. 

ANA WIME AND NATROLITE. 

A cave-rock in the southern Gio-ô, presents an anomalous 
habit ; a slide made of it contains a colourless mineral in angular 
or polygonal interspaces between the crystals of plagioclase 
{PL II, Fig. 5). It shows no signs of any crystallographic face, 
nor cleavage, but only has a frittered appearance, being traversed 
with irregular cracks, and also being pierced through in all 
directions with the needles of apatite which is excessively rich 



OF THE DEPENDENT ISLES OF TAIWAN. 29 

in this rock. The polysomatic mineral has a smaller index of 
refraction, when compared with that of the accompanying pla- 
gioclase, as may be easily experimented upon by Becke's method. 
These colourless patches, as a rule, behave optically isotropic ; 
at times, however, faintly double-refractive, and separate into 
several optical fields. They readily dissolve in HCl, with the 
formation of the cubes of rock-salt. The same patches frequently 
resolve themselves into a radial-fibrous, somewhat brownish and 
highly double-refractive body (well seen at the margin in the 
lower, left quadrant in PL II, Fig. 5) with the positive sign along 
the axis of the needles. 

The polygonal base-like mineral, moulded upon plagioelase 
and augite, seems to be identical with what Biiking^^ calls the 
^ Basis zweiter Art,' and is allied to i\\Q pitchstone-glass of Hunter 
and Rosenbusch.'-^ Recently, this base was studied with great zeal 
by the American petrologists, Lindgren,"^ T. F. Williams,^' Kemp,^^ 
Fairbanks,*^^ Cross,'- Coleman^* and Pirsson"'; the last author 
especially paid close attention to this subject, in making care- 
ful analyses and also recalculating tlie analytical result, ob- 
tained by Hunter. From his study, Pirrson is forced to the 
conclusion that the so-called colourless base has exactly the 

1) 'Basaltische Gesteine, etc.,' Jahrb. K. K. preus?. geol. Landesanstalt. 1S80, S. 153j und 
1881, S. 606. 

2) ' Ueber Monchiquite, ein camptonitisches Ganggesteiu aus der Gefolgenschaft der 
Eleolithsyenite,' Tschermark's Min. Math. XI, 1890, S. 415. 

3) Proc. Cal. Acad. Sei., Vol. HI, 1390. 

4) Cited in Pirsson's paper. 

5) 'Trap Dikes,' Bull. 107, U. S. G. S. 1893. 

6) 'On Aualcite Diabase from San Luis Obispo County, California,' Bull. Geol. Depart. 
Univ. Cal., Vol. I. p. 27?. 

7) ' An Analcite-Basalt from Colorado,' Journ. Geol. Vol. V. p. 684. 

8) 'A new Analcite Kock from Lake Superior' Journ. Geol. Vol. VII, 1899, p. 4?2. 

9) ' The Mochiquites or Analcite Group of Igneous Eocks,' Journ. Geol., Vol. IV. 1896, 
p. 679. 



30 KOTO : NOTES ON THE GEOLOGY 

same chemical composition as tliat of analcime, and tlie physical 
properties observed give no hinderance to the assumption that 
this component actually is that miueral. He thinks the analcime 
is primary, having been formed from the magma, containing 
water and much soda, under pressure with considerable rapidity. 
From what has been stated before, I have also, to all ap- 
pearances, the primary analcime in the interspaces of the com- 
ponents in the Basalt from Gio-ô, and the radiating bundles of 
a strongly biréfringent natrolite are formed secondarily from the 
analcime through a molecular rearrangement. Both components 
make their appearance with the dodecahedric networks {PL II, 
Fig. 5) of the skeleton magnetite which occupies the other portion 
of the slides. 

THE IRON ORES. 

Both ilmenite and magnetite are present, and they usually 
belong to a single generation, and indeed the product of the 
eftusive period, as the iron ores were not found enclosed in the 
olivine of the intratelluric crystallisation. Both ores, especially 
the ilmentite, have crystallised later than plagioclase, but slightly 
prior to, or contemporaneous with, the monoclinic pyroxene. 
The ilmenite and magnetite are, under the microscope, not easy 
to be distinguished, as every petrographer will agree, if crystal 
forms are not serviceable for their diagnosis. 

The ilmenite is, however, tabular and needle-shaped in sec- 
tion in the Basalt with a strong lustre and a violet tinge, when 
seen by reflected light, on the flanks corresponding to the 
thickness of slide. The laths are slender, appearing merely as 
lines, and cross several crystals of felspar and augite, mean- 



OF THE DEPENDENT ISLES OF TAIWAN. 31 

while the substance of the ilmeuite entirely disappears when 
traversing other crystals, and comes again into view in the same 
direction as a continuation of the interrupted crystals. Unfortu- 
nately basal sections were not frequently observed, and this 
was the greal obstacle in ascertaining the presence of ilmenite 
11 microscopic analysis. The ore with above-mentioned lamellar 
habit occurs exclusively in a coarse-crystalline type of intersertal, 
or ophitic structure, irrespective of hyperstheue or olivine-bearing 
Basalt ; and this fact lends evidently a strong support to the view 
advanced by K. Hofmann,^^ that the ilmenite accumulates in the 
lower portion of lava-flows, and in that which has crystallised 
under high pressure, while the magnetite is rich in the upper 
part that has consolidated under a low pressure. Fr. Sandberger-^ 
says also that Basalts may be classified into Dolerite and Basalt 
proper, by the presence of ilmenite in the former and magnetite 
in the latter. These fruitful ideas inaugurated by both authors, 
now unfortunately passing into oblivion, deserve the careful 
attention of petrologists. 

A slide of the Basalt from the islet of Hatto was treated for 
a considerable length of time with a strong hydrochloric acid 
without any appreciable result. A large quantity of the jduI- 
verised sample of the same specimen was then digested in 
boiling HCl with the addition of tin-foil, and the solution was 
coloured slightly violet, showing the presence of titanium in the 
dissolved j)ortion of the ore. Ilmenite also occurs, according 
to VénukoÔ"^ very abundantly in the Basalts of Mongolia, and 
even transparent lamellae were found by him, just as in the Pes- 
cadores rocks. The ilmenite is fresh and leucoxene not noticed. 

1) 'Basalt von Eakony,' Zeitschr. d. d. gcol. Ges., XXIX., 1877, S. 191. 

2) Eosenbusch, 'Mikroskopische Physiographie,,' II., 3te Auflage, S. 1015. 

3) ' Les Bodies Basaltiques de la Mongolie,' Bdl. Soc. belge de Géologie, etc. T. IL, p. 443. 



32 KOTO : NOTES ON THE GEOLOGY 

In my few slides of Basalts, bearing the iddingsUised olivine, 
ilmenite seems to he ivanting, though the rocks approach to a 
Doleritic type, being replaced by magnetite. I cannot say 
positively that this rule holds true for all the iddingsite- bearing 
Basalts. 

The magnetite is, on the other hand, the prevailing ore in 
the compact Basalt, and in the Limburgitic type, in the form 
of isometric crystals and dust, occurring either in the general 
mass, or else enclosed in augite and olivine. The face of the crystals 
shows a metallic lustre with a tinge of blue by reflected light. 
The dust is sometimes peripherally altered into a blood-red 
iron-glance. A slide made of a chip from Hoko, was digested 
in HCl with the addition of KI ; and then the black ore, 
therein contained, was entirely removed, and the solution not 
coloured when tested with tin-foil, proving thus the presence of a 
pure magnetite. As it is already stated above, the magnetite- 
rich, compact type seems to make up the upper portion of the 
thick flows of the Hoko Basalts. 

In the Anamesitic type from the islet of Gio-o, we find beauti- 
ful networks of the skeleton-crystals of magnetite in a devitrified 
mesostasis within the polygonal spaces between crystals. They 
are the dodecahedric dendrites, consequently the skeletons inter- 
sect each other at the angles of 60° and 120°, and are said to 
consist of garnetohedrons. They all go into solution by treating 
with HCl. Morozewics^^ tells us that the spire and filigree-work 
of the skeleton magnetite, crystallising out of the magma rich in 
iron oxides, consist of minute octahedra, arranged rectilinearly 
in the direction of the crystallographic axes with secondary and 

1) ' Experimentelle Untersuchungen über die Bildung der Minerale im Magma/ 
T&chermak's MUthcilungen, 18, 1898, S. 90. 



OF THE DEPEXDEXT ISLES OF TAIWAN. o3 

teîtiary offshoots. This mode of growth, the octahedric dendrile, 
so called by ^Eorozewies, is well known in petrographical 
literatnre, sii'.ec the publication of Prof. Zirkel's^^ woik. On 
the other hand, it is said that the dodecahedric dendrite, as is in 
the present case, is formed out of the magma poor in iron- 
oxides. 



B- Special Description of Individual Occurrences of Basalts. 

A. THE GRANULAR TYPE. 

{PL I, F'uj^. 1 and 2.) 

As seen by the naked eye, it is greyish-black and compact, 
with the dots of olivine which is the only visible component 
of the whole mass. This type is represented by two specimens 
from Hôko, and one from Hakusha.-' Microscopically it is 
holocrystalline with the smaller phenocryst of olivine, imbedded 
in the still finer aggregate of the ground-mass. 

The fineness of the ground-mass, however, varies in different 
specimens, and even in the same slide. Some portion of the 
same slide is, therefore, extremely rich in idiomorphic angite 
to the total exclusion of felspar and olivine, but with small 
patches of brown glass. Were this portion independently devel- 
oped, it Avould be fitly called the Äugitile {Fig, 2). It is the 
local assemblage of augite within the rock, and that mineral es- 

I ) ' Die mikroskopische Beschaffenheit de'" Mineralien und Gesteine.' Leipzig, 1873, S. 244. 

2) Collected at Eyô-fcô-san (^^ÜJ); and, according to Mr. Saito, it appears in I. horizon, 

i. e., the uppermost slieet, consequently the youngest of all the lavas of the Hôko Group 

{Fl. r, fig. 1). 



34 KOTO : NOTES ox THE GEOLOGY 

pecially accumulates near the margin of the secretionary mass, 
the augite being sometimes arranged along the linear common 
base, with the free ends of the crystals toward the interior. These 
phenomena indicate that the lava' had consolidated in a quiet 
state. 

The relative proportion of augite and plagioclase is also vari- 
ous, and in the cases where the former outweighs the latter, the 
olivine increases in its quantity and comes also in the ground- 
mass, as a product of the crystallization of the effusive period ; 
and at the same time the texture of the rock becomes finer. 
If, on the other hand, the plagioclase becomes predominant over 
the augite, then, the texture gets coarser and more crystalline, 
and the distinction between phenocrysts and ground-mass is not 
then commonly well maiked. Apatite and ilmenite seem to occur 
in the latter variety only, the ilmenite is sometimes transparent 
with a deep brown colour. 

The only mineral that serves as the phenocrijsi is olivine. 
Its forms are various, owing to the various degrees of resorption. 
Most have partial crystallographic faces with dee|D indentations 
of corrosion, and a drop-like black iron-ore and feUpars were 
formed in those spaces. Sometimes the act of corrosion has ad- 
vanced so far that there remain but patches as the relics of a 
large crystal, and^the eating away oi tlie body by the magmatic 
menstruum proceeds always from the lateral pinacoids. As usual, 
the crystals of the olivine are not fresh ; but the routine of change 
is the same in all. They become fibrous and lamellar, parallel to 
one of the lateral ^linacoids, the altered portion being yellow or 
brown, according to the degrees of transformation. The mode 
of change is similar to iddingsitizalion {Figs. 1 and 2). 

The (jroimd-)nass consists, first of all, of the crystals and 



OF THE DEPEXDENT ISLES OF TAIWAX. 35 

grains of augite, all of a liolei-hvown colour, besides the grains 
of olivine, and the laths of the multiple-twinned plagioclase, the 
octahedra and dust of magnetite, and ilmenite. The texture of 
the rock is crystalline and typically granulitlc. In a coarse 
variety, the idioinorphic augite with hour-glass structure forms 
stellar aggregates, and these aggregates closely resemble ihe 
glome roporphyri tic phenocryst. 

B. THE TYPE OF THE IDDIXGSrTE-BEARIXG BASALT. 
{PI. I, Figs. 4, 5 and 6; PI, II, Fl<j. 1.) 

Megascopically this type is greyish-black Anamesitic-looking, 
and finely uniform-granular, owing to the nearly equal size and 
form of the laths of plagioclase which predominates quantitatively 
over the other components. 

The characteristic features of this group are firstly, the pre- 
sence of large phenocrysts of olivine which is more or less iddingsi- 
tized ; secondly, the majority of the augite is xeuomorphic or 
granuhn-, and of small size, and these grains are grouped together 
inteisertal! v Avith the devitrified irlass between the laths of 
plagioclase. The structure is typically intersertal. The promi- 
nent characters distinguish this group from the rest of the Basalts. 
This type is represented in my slides from the Pescadores by 
three specimens, one from Kippai, and two from Hôko, one of 
which was struck off at the locality Tai-san ;^' according to 
Y. Saito, it forms the uppermost flow there. The same may 
be said of the specimen from Kippai, since the youngest 

1) ^^^'cUl) Sç-SA î0îffl- '^Jie rock efleive;ces witli ticid. The microscope discloses the 
fact tliat tlie radinlly arranged fibies of calcite fill up the j^olygonal spaces between other 

components, showing bars which correspond to the position of crossed niçois. 



36 KOTO : NOTES ON THE GHOLOGY 

lava-flow is the only effusive that can be met with on that 
island. 

The olivine is the sole i')henocryü\ it is variable in size (the 
largest one measures even ö mm.), irregular in distribution, and 
multifarious in form, some having partial crystallograpliic faces, 
while others have none of them. The idding^itiz ition is pecu- 
liarly inherent in the olivine of this roek-group, and I refer the 
readers for further details to the topic : " component-minerals " 
p. 18 et seq. By the way, I have only to mention that the name 
iddingsite may conveniently be applied to a special transitional 
form of the alteration of an olivine which, after passing this 
stage, changes into dirty-green spherulitic fibres of an optically 
positive character. 

In the felspar-rich rocks {PL /, Fir/, (j), which are prevalent 
in the group under question, the j^lccffioclases are all approximately 
of the same size, and surpass the augite both in dimension and 
quantity; while in the augite-rich rocks {PL I, jig- 4), the 
])lagioclases are of two generations, and the larger ones behave 
porphyritically towards the minor ones. They are lath-shaped, 
and muUii)le-t\vinned, the terminations being imperfect aiid 
sheafy, and these laths are thrown together in an orderless 
plexus, which eminently characterises the structure of normal 
Basalt in contradistinction to that of Andésite. 

The avrjite is all of a single generation, consequently uni- 
form, but inferior in size to the phigioclase and olivine. Some 
are rudely idiomorphic, but by fir the most of it is granu- 
lar, occurring in groups, and filling the angular spaces left 
by the laths of phigioclase. The augite is, as usual, of a violet- 
brown colour, but in the specimen from Tai-san, it is almost 
colourless in sections. It is free from foreign inclusions, and 



OF THE DEPENDENT ISLES OF TAIWAN. 37 

the liouf-glass structure is faintly indicated in some individuals. 
In the coarse, felspar-rich .«specimen, the iron-ore is piesent only 
in small quantity [PL I, Fig. 6), but comparatively large, lamel- 
lar and flat with glittering bluish lustre on ihc perfect cleavage- 
surface. It looks rather more like ilnienite than magnetite. Stiff!, 
slender (r/;K////e- need les, sometimes \Yith a bi'own canal traversing 
the whole length, are particularly abundant, being scattered 
through the whole mass. 

In the dark fine specimens {PL I, Figs. 4 and ~)\ small 
regular crystals of inagneliie are plentiful, and in these slides, 
I found abundantly the small laths of twinned plagioclase, which 
lesolve at the ends into slightly diverging columns {PL I, Fig. o), 
and these may be easily mistaken for those of apatite, if needles are 
found detached from the waist. Optical properties aie not in- 
dependtly shown in them, on account of their extreme thinness- 
Similar bodies are noticed by H. S. Washington in the sanidine 
of some Ischian Trachytes and named by him heraunoid.^^ 
He and also Lehmann'-' attribute the splittings and ramifications 
from the main crystal to the existence of internal tensions in 
felspar, but the cause of the existence of such tensions remains 
to be solved. 

The glass together with the augile fill up the poly synthetic 
space left by the laths of j^h^gioclase. The glass-base is coloured 
bottle-green, sometimes dirty brown, and devitrified in various 
ways. It consists of polarizing scaly aggregates of vermiform, 
spherulitic, or, irregular shapes. Sometimes fascicular and 
radiating needles, which are colourless and biréfringent, are 
imbedded, in the green base as a product of devitrification. The 

1) 'On some Ischian Trachyte.' Journ. Anier. Sei., May, 1896, p. 3S0. 
2j ' Molecularphysik,' I, 1888, S. 378. 



38 KOTO : NOTES ON THE GEOLOGY 

needles may possibly of a felspathic nature and such a structure 
is termed tarioUiio by Harker/^ though in the present case those 
circular, whitish spots, called varioles, are wanting. Green, 
fresh base is here and there also found between the angular spaces. 
Thin lameHœ of rugged outlines, with violet-brown colour, 
may be frequently noticed in all of my slides, and they closely 
resemble tiiose found as interpositions in a hypersthene. I 
conjecture the minéralogie nature of these plates to be ilmenite. 



6'. THE OPHITIG TYPE. 

{Fl. II, Fl)/. 2) 

This type is represented by a single specimen from Hoko, 
and Gio-o"' respectively, and two from Haku-sha, though the 
* lie ' is not known to me exactly ; but it is highly probable that 
samples are taken from the second sheet which is separated 
usually from the uppermost columnar flow by an ash bed of a 
certain tliickness. It is a greyish-black, Anamesitic rock, with the 
brownish, lath-shaped phenocrysts of plagioclase (4 mm. length). 
This is the coarsest type of the Hoko Basalts, and is the one rich 
in plagioclase in comparison with ferro-magnesian silicates ; it 
seems to have solidified in the lower portion of the lava flow. 

Under the microscope, it shows a porphyritic, hypocrys- 
talline, diabasic structure {Fig. 2) with the ophitic plates of 
augite of considerable dimensions, enclosing the laths of plagio- 
clase which lie in all possible directions. The augite is of a 

1) "The aggregates of felspar-microlites or fibres with fan-like or sheaf-like groupings. 
They may be closely packed to make up the entire mass of a portion of the rock (Basalt)." 

Petrology for Students.' find. Edit., p. IUI and 201, Fig. 41 A. 

2) Tlie ex:ict locality being Sho-chi-kaku, ('hJfe:^) :»t tlie middle of the iskxnd. 



ON THE DEPENDENT ISLES OF TAIWAN. 39 

kind of VigJit-hrownish colour, and its plates are often multiple- 
twinned, and enclose, besides plagioclase, a number of round and 
corroded crystals of olivine which is for the most part changed 
into green, pleochroic fibres ; the iddiugsitization of the olivine was 
so far not observed. The 2)loffioclases are of two generations (i^/y. 2), 
the larger, probably intratelluric, species has fissures (?ee Fig. 2) 
filled with films of brown hydrous sesquixide of iron, which 
cause the phenocrystic feldspar to appear 'niacroscopicalli/ like an 
olivine. The plagioclase is partially embraced by the ophitic 
plate, while the smaller laths became entirely enclosed in it. 
The polygonal inters{)ace3, when not occupied by augite, are 
otherwise filled up wâth the fibrous devitrified glass, the latter 
containing globulites, sometimes dendritic, and apatite ; and the 
thick lamellœ of ilmenite traverse the base, but not the plate of 
augite, consequently the crystallisation of the ore must have 
taken place posterior to that of the pyroxene. Sometimes the 
greenish-yellow augite is coarse-granular, and in this case the 
structure approaches to that of inlersertal. Ilaynetile seems to 
be ^vanting. Owing to the coarseness of the structure, the rocks 
are often porous, and the polygonal, angular spaces are often 
filled up with banded, purplish chalcedony. 

I). THE TYPE OF THE LI VINE-LESS BASALTS. 
{PL I, Fig. .3 ; PI. II, Fig.^. 2 and 3.) 

The olivine-less, hypersthene-bearing Basalts are represent- 
ed in my collection by two specimens from ^Yampai", and one 
from each of the following islands, Hoko^', Kin-sho'^', and 
Hatto-sho^'. They are megascopically wet-grey, and fine-granular, 

1) WM 2) Sei-kei ®^ in Hôko 3) ^g| 4) A^^. 



40 KOTO : XOTES ox THE GEOLOGY 

the general microscopic aspect being a crystalline Andésite -like. 
They are all extremely ricli in augite, and the structure is (/rami - 
Utic, Tlie felspars are of two generations {PL II, Fig. 2 and 3), 
and the rock is consequently porphyritic, owing to the presence 
of a few large cr3^stals of plagioclase, though this structure could 
not be easily recognised as such in the ])resent group. 

The p/ienocrystic plagioclase is narrow-tabular with a few twin- 
ning lamellœ (see Iig. 3), and is remarkable in its being traversed 
through by sets of cracks which run approximatel}^ parallel with 
each other. In one instance, only one lamella, out of many twin- 
ned parts after the pericline law, is provided with closely set fissures. 
This anomalous feature can be seen in all the specimens of the 
present type, but not common in others, and the same peculiarity 
recurs also in augite whose granular aspect is due in great measure (o 
the same cause. I cannot offer at present a satisfactory explanation 
to account for this phenomenon ; but, as Judd says, it might in 
part have been cansed by a slow but constant movement of a 
crystallizing magma, and also chilled suddenly, perhaps by the 
access of water at the final stage of consolidation. I may here 
adduce in support of my ground a fact of the special distribution 
of the Hypersthene-basalt which, so far as I am aquainted with, 
occurs only in the outlying islets, excepting the locality Sei-kei, 
on the north coast of Hoko, which also lies not very far from the 
present sea-shore. 

Hyperlhene occurs exclusively, though insignificant in quan- 
tity, in the form of phenocryst {Fl. II, Figs. 2 and 3) and takes 
the place of olivine in the present rock -group. It is sedge-like 
in general shape, and granular in its margin (especially in 
Fig. 2), };eing fringed with grains of common augite, whose 
presence becomes strikingly apparent between crossed niçois. 



OF THE DEPENDENT ISLES OF TAIWAN. 41 

on account of their vivid colours of polarisation in contrast with 
the grey tint of the hypersthene in the interior. Pleocliroism is 
scarcely perceptible. Traces of a few rough cleavages run 
through the hypersthene lengthwise, and as in the case of the 
plagioclase, it is trasversed with many fissures. The hypers- 
thene is of intratelluric origin, and has the general aspect of its 
having been worn out caustically and frittered, and the peri- 
pheral accumulation of augite, already referred to, seems to have 
some genetic relation with the act of degeneration. 

Large, monoclinic augite sometimes makes its appearance in 
company with the hypersthene and plagioclase, forming local 
patches of secretional origin, wiili the hyperitic structure. 

The ground-mass, which constitutes the main bulk of 
the rock, consists of laths of plagioclase and grains of the 
frittered and corroded augite, together with rugged clumps 
of magnetite. The relation of the first two components cannot 
be told in a few words. In one instance, the mutual relation 
is such that we could almost say it is ophitic ; in another, it 
is intersertal in company with a little remnant of brown glass, 
while in the third, no such arrangement could be discovered, 
but a simple aggregate of felspar and grains of augite, there- 
by calling forth the structure which is termed graniditic. The 
augites of both generations are of yelloivish broivn and not violet- 
brown. 

Shingly tridymite fills up polygonal spaces, and the loose 
brushes or tufts of either plagioclase or apatite are thrown 
through the whole mass. A doubtful iddingsite {PL I, Fig. 3) 
was once o1)served, and some rocks are calcareous too. The 
stratigraphie position of this type is not known to me. It may 
be the lava of either the first or the second flow. 



42 KOTO : NOTES ON THE GEOLOGY 

E. THE TYPE OF THE ANALGIME-BASALTS. 
(PL JI, Fig. 5.) 

This to the naked eye is macroscopically deep-grey, and fine- 
granular. Under the microscope it is hypocrystalline and more or 
less ]3orphyritic, either the xenomorphic olivine or the aggregate of 
the automorphic augite being the phenocryst, or sometimes both. 
The texture varies within a wide range, but generally speaking is 
coarse (Fig. o). The porphyritic elements, however, differ general- 
ly not much in size from the crystals of the ground-mass, and 
the mode of arrangement of the several components is graniditic. 
Plagioclase predominates over augite in quantity ; and magnetite 
is not plentiful, and completely soluble in HCl. The paucity of 
iron-ore causes the rock to appear of a grey shade. 

Olivine occurs as a phenocryst in the xenomorphic grains, a 
few of which have been reduced even to mere flecks through gradual 
resorption. Cleavages are not noticeable in contrast to other 
olivines, but in stead of them there are curvilinear cracks, con- 
forming approximately in their direction to the boundary of 
resorption. The substance of the olivine is colourless, and usually 
more or less altered into a greenish or yellowish, fibrous sub- 
stance (not iddingsitic). Brown decomposition is quite foreign 
to the olivine of this type. The present olivine seems to belong 
to a variety rich in magnesia. Phenocrystic pyroxene is scantily 
present in some, but none in others. The augite is of the typical 
Basaltic variety, with a violet-brown type, possessing the hour- 
glass structure, and idiomorphic, flattened on the orthopinacid. 
It occurs singly or in stellar aggregate. There is no felspar- 
phenocryst. 

The ground-mass consists of laths of plagioclase, crystals 



OF THE DEPENDENT ISLES OF TAIWAN. 43 

and crystalloids of violet-brown augite, magnetite, and xeno- 
morpliic olivine, with the interstitial mass of analcime and 
base. The laths are multiple-tsvinned with the parapet-like 
terminations {PL II, Fig. ö) produced by the shifting of lamellse 
to the one end or the other with reference to the adjacent plate. 
The slide treated with HCl shows a considerable corrosion of 
the interior lamellœ of the laths, while the exterior remains 
intact and fresh, as if a frame is enclosing the hollow space. 
The crystals of a violet-brown augite of the short prismatic habit, 
rather flattened towards the ortho-axis, are freely developed, or 
occur in clusters. The augite and plagioclase must have, there- 
fore, crystallized simultaneously, and at their contact the one 
is partially penetrating the other and vice versa. Magnetite is 
idiomorphic, but frequently possesses irregular outlines, owing to 
the penetration of the crystals of plagioclase, augite, and apatite, 
and the larger crystals are anhedrons, as they are moulded upon 
the neighbouring laths of the plagioclase. The magnetite is 
comparatively large and few, excepting its dendritic skeleton 
crystals which are found abundantly in the specimen from Gio-o, 
in company with devitrified glass. In the specimen, which is 
wanting in dendritic magnetite, there are brown, biotite-like 
lamellse usually in association with the hexahedral iron-ore. The 
lamellse are anisotropic, and distinctly pleochroic, and the 
mineral is conjectured to be ilmenite. 

It is of no small interest to note the presence of analcime. 
It occurs sporadically rather in large patches in the cuneiform 
spaces left by other crystals. It is generally fresh and colourless, 
and isotropic, but often sho^YS the optical anomalies so common to 
this mineral. At times, the analcime resolves into a dirty, iib- 
rous natrolite (as in the left, lower margin in Fig. 0). The 



44 KOÏÔ : NOTES ox THE GEOLOGY 

aualcime seems, so far as my experience goes, to be exclusively 
confined to this type, though it is possible that the colourless 
base in minute interspaces of other Basalts of the Hoko Group, 
might turn out to be that mineral, if the means are at hand in 
ascertaining its presence. 

Another accessory to Ije mentioned is apatite in colourless 
prisms, which is especially plentiful in this type. 

The colourless base and analcime are rather unexpected 
guests ill the basic, black roch, such as we have here, and the 
mode of occurrence is that they fill up the polygonal interspaces 
left by the crystals of other components of the rock. If we 
accept the primary origin of the analcime, as Pirsson^^ would do, 
it is all the more very striking to see that the residuum of 
a Basaltic magma should have an exact composition of 
Na20*Al203'4 Si02'2HoO. Yet the analcime seems to all appear- 
ances to be of primary origin, if we take into account the 
perfectly fresh state of the rock in which it is found, and not 
only in the Basalts of the Hoko Group, but in the Teschenite 
of the Nemuro promontory in Hokkaido, I had several occasions 
to observe the same mode of occurrence of the aualcime, so that 
it could not be attributable to a mere accidental circumstance 
to find it in such state, as several foreign writers also noticed 
the same. It excludes the idea of its having replaced the base 
which formerly occupied the place of the now-existing analcime. 

The present mode of occurrence of the analcime may perhaps 
be explained by supposing that, when the Basalt luas consolidating 
0)1 the i^uiface in a quiet state, carrying in it the intratelluric oli- 
vine, the 7iewly created crystals, such as those of plagioclase, augite, 

1) ' The Moncliiquite or Analcite Group of Igneous Eocks.' Journ. Geol., Vol. IV., 1896, 
p. 679. 



OF THE DEPENDENT ISLES OF TAIWAN. 45 

magnetite, apatite, together with the olivine had then sunk down, 
and, formed the heap of cri/stals at the bottom, meanwhile the un- 
consolidated residuum of the magma ivas actually slowly flowing 
through the sieves of crystal-heap, or changed its chemical com- 
position through diffusion, after the manner of liquation as 
ill a metallurgical process. And, then, the solution having the 
composition of the hydrous alumino-sodiura-silicate has finally 
crystallised out in the interspaces of the meshes of crystals. 
Similar process can 'oe frequently observed during the formation 
of crystals on the stage of the microscope. If this be the actual 
condition under which the Analcime- Basalts have consolidated, 
considerable leaching and percolation must have taken place 
during the formation of rocks, and the structure of such a rock 
should better be called the ' leached.' This structure is therefore 
properly seen only either in the granitic or in the granulitic 
rock, consequently it is wanting in the family which has 
a fluxional arrangement of the components. 

The Analcime-Basalts are represented in my collection by 
three specimens from Gio-o, and one from Hoko, The hand- 
specimen from Nai-an^^ in Gio-ô, is, according to Y. Saito, said to 
occur at the water's edge, the main portion of the flow usually lies 
under the level of sea, and constitutes the third sheet of lavas, 
and is the loivest, consequently the oldest of the accessible lava- 
flows of Gio-o. 

Other Analcime-Basalts of the Hoko Group no doubt belong 
to the same horizon. 

1) Nai-an ^^, ^^^%. 



46 KOTO : NOTES ON THE GEOLOGY 



TUB ISLAND Oi' KOTO) (BOTEL-TOBAGOJ. 

Starting from Makiair\ one of the Spice Islands, a long- 
chain of the Moluccan volcanic system runs upwards, and joins 
at the solfataric volcano of Api, in Mindanao, with that of the 
Sangirs, that comes from the north end of Celebes. The united 
system of volcanoes in the Philippines, then, receives the name 
of the Mayon system. It goes right through the whole breadth 
of Mindanao, and enters Caminguin, Leyte, Biliran, and, after- 
wards, the peninsula of Camarines of south Luzon. It is in 
the last-named region that the volcanic activity of the Philippines 
is fully displayed. Albay or Mayon stands foremost in rank 
among the mighty cones. For a while, we lose sight of the chain 
northwards under the Pacific bottom, and it reappears in full force 
at the crater of Cagua near Cape Engano, in north Luzon. 

The northern prolongation of the Mayon system may still 
be traced through the little-known Babuyans,^^ the Batans, and 
the Bîishi islands. AH are said to be of volcanic origin. Among 
the Bashi or Yasshi^', the five larger islands, going from the south 
to the north, are Liayan, IMabudis, Tanem, Maysanga, and Tami, 
the last being the largest of the forlorn isles. An active volcano 
is said to exist in the southern region (?), spreading fire and 
destruction. 

The Balintang Canal at 20° N. lat. separates the Japanese 



1) um- 

2) B, Koto, ' On the Geologic Structure of the Malayun Archipehigo.' This Journal, 
vol. XI, pages 111 aud IIS. Wichmann calls the chain- the 'North Moluccan bow.' ' Der 
Wawani auf Amboina und seine angeblichen Ausbrüche, III.' Tijdschr. v. h. Kon. Nedeii. 
Gen., Jaargang 1899, S. ö2. This bow is now said to start from Batjan, lying to the south 
of Makjan. loc. cil. S. H. 

3) Koto, loc, cit. p. 118. 

4) The Japan Mail, August lUth 1897, ' Forlorn isles.' 



OF THE DEPENDENT ISLES OF TAIWAN. 47 

domain from that of the United States. AVithin the Japanese 
waters lie the Batans, the Bash islands, the rocks of Gadd and 
Forest Belle, the islands of Shô-Kôtô (Little Botel-Tobago) and 
Koto (Botel-Tobago), and, lastly, Kasho (Samasana), as the conti- 
nuation of, I conjecture, the Mayon system of volcanoes {Fig. 1), 

The smaller isle of Koto is, geologically speaking, entirely 
unknown, but the Larger Koto has been several times visited 
by the Japanese, since the first landing of a staff of the gover- 
norship of Taiwan, in April, 1897. Among our University 
men, Mr. Tatla stayed there a week collecting zoological speci- 
mens, and, lately, Mr. Toiii remained longer in this lonely island 
among the aborigines for his anthropological study. I myself 
have not had the opportunity of visiting it, though the island 
has been within my sight for a week long, while travelling the 
pathless beaches of south-eastern Taiwan. 

The island of the Larger Koto {Fig. 1) lies in a south-eastern 
direction about 50 miles off the coast of Pinau, and 35 miles 
north of north-east from the Cape of Galambi in Taiwan. Its 
north-south extent is 3 ri and the breadth 11 ri, with the 
circumference of 9 ri. It is the abode of 1,500 nude aborigines. 
Seen from a distance, this scapula-shaped island appears plateau - 
like in general profile, crowned by a prominence of 120 m., 
somewhat excentrically situated in the north ; and is bounded by 
steep declivity all round the coast, so that it leaves only a nar- 
row patch of lowland on the south-western shore, which serves 
at the same time for the chief anchor-ground of this islet. 

Being situated amidst the stormy and swift Kuro-shiivo cur- 
rent, the narrow beach is highly cobbly, as may be seen from 
Mr. Torii's photographs ; and the steep clifi' undoubtedly owes 
its present form to the abrading action of dashing waves. 



48 KOTO : NOTES ON THE GEOLOGY 

Fringing reefs are said to skirt the shore, some portion attain- 
ing doable the man's height above the water's edge, indicative 
of a recent negative shift of the relative levels. It seems to me 
probable, that they are not the reefs of Neocene time, which 
usually attain a considerable height of more than 200 m., as in the 
Apes Hill of Takao, but those of a comparatively recent date, 
possibly representing a Diluvial formation. The plateau-like 
elevation, which faces the sea in cliffs, seems in parts at least 
in the north-east point to consist of volcanic agglomerate. A 
greater part of the interior seems to be built of volcanic rocks 
with a gabbro-like plntonic mass as the foundation of tlie island 
exposed at the west coast, but their mutual relations and area 
of distribution are quite unknown to me. 

In the following, I will give a succinct account of rocks, 
kindly placed at my disposal by Messrs. Ishii and Torii. 



A. FELSPAR-BASALTS. 
{PL III, Figs. 3 and 4.) 

My slides of Basalts and Andésites are prepared from chips 
of water worn gravels, used as weights attached to a fishing net 
of the aborigines. 

The Basalt is ratlier porous, greyish- brown mass with a few 
phenocrysts of a brown olivine (1-2 mm.) and black common 
augite. Under the microscope, the olivine occurs in two gene- 
rations {Fig. 4). Its forms are acute six-sided, sometimes 
nearly square, truncated at corners, but mostly corroded and 
disfigured, with a few traces of basal cleavage. The crys- 
tals are slightly decomposed in their margin, being either yellow 



OF THE DEPENDENT ISLES OF TAIWAIST. 49 

or brown ; but as a whole the interior is fresh. It is the iron- 
rich variety — hyalosiderite, as is proved by the micro-chemical re- 
actions, which show only a trace of magnesia. The olivine 
encloses a large quantity of regular octahedra or elongated 
crystals of the brown, transparent /»îco^^ïe, mixed with the crystals 
and dust of magnetite. 

Plagioclase, as a phenocryst, is observed only once in ni}' 
three slides ; it is long-rectangular in form, with negative crys- 
tals, filled with a gas. The crystal is multiple-twinned, extin- 
guishing light symmetrically with the maximum angle of 32°; 
consequently it is the calcium-labradorite. The augite is rather 
automorphic, showing, however, a slight corrosion marginally. 
This character is common to all of the specimens. The crystal 
occurs in polysynthetic twins ; the colour yellowish-green and non- 
pleochroic. As usual, it has glass-enclosures with air- pores. 
Sometimes, the augite is internally and nucleaUy resorbed, leaving 
an accumlation of grains of the same in its place. The augite 
is of nearly the same size as the olivine. 

The ground-mass is seen, under the microscope, to make the 
main bulk of the rock. The micro-phenocrysts of olivine and 
augite are tlie same in habit as the macro-phenocrysts. The 
augite is in a few cases fringed with sheleton-crystals. They are 
inbedded in the plexus of the felspar-laths and clum^^s of mag- 
netite, rudely showing a fiow structure. The laths are twinned 
simply or polysynthetically, and in many case hollow, with the 
very thin external rim, partially or entirely filled up with glass. 
So far as I am aware^', such skeleton-crystals of felspar seem to 

1) The same skeleton laths are observed_by E. Elich in the Ampbibole-pyroxene An- 
désite from the Kio Blanco, West Cordillera, Ecuador. Reiss n. Stiibel, ' Eeisen iu^Siid- 
Amerika, Das Hochgebirge der Eepubik Ecuador, I.; Petrographische Untersuchungen, I. 
West Cordillère,' S. 163. 



50 KOTO : NOTES ON THE GEOLOGY 

be of extreme rarity. The laths extinguish light symmetrically 
but in the contrary direction at an angle of 26°-27°, proving 
the felspar to be more acidic than its phenocryst. Interstitial 
space is occupied by a brown glass which contains globulitic 
and rod- shaped bodies. From the foregoing descriptions, it is 
evident that the Basalt of the Island of Koto does not properly 
belong to the category of normal Basalt ivith violet aiigite, present- 
i7ig the intei'sertal structure, which is so common in the 7'ocks of 
the Hoko group, already described. Here exclusively monoclinic 
augite presents the character of diopside. Both the olivine and 
augite, all being equally corroded, present so great a variation 
in size from the macroscopic to tlie microscopic dimensions that 
I could not discriminate the products of the intratelluric from 
the effusive period of consolidation. The ground-mass, as I have 
said, is highly ielspathic, and the structure is Andesitic and 
hypocrystalliue-porphyritic, somewhat resembling a pilotaxitic 
type. Richness in olivine and paucity in iron ores, as well as 
globulitically granulated mesostasis make the rock approach 
to 3, navitic structure [Mg. 4), the only difference being the presence 
of feldpar-laths in the ground-mass. The rock seems to me to 
be a lava-flow, consolidated rapidly, accompanied by a brisk 
liberation of gas from the cooling magma. 

B. HORNBLENDE-ANDESITES. 
[Fl. Ill, Fig. 5.) 

I liave three specimens of rocks in Torii's collection, belong- 
ing to the same category. They differ in colour consequent on 
the various stages of decomposition. A fresh one is greyish and 
porous, speckled with phenocrysts of hornblende (2 mm. in 
length). 



ON THE DEPENDENT ISLES OF TAIWAN. 51 

Plagioclase is long- rectangular along the zonal axis at right- 
angles to 010, and tabular when parallel to that face {Fig. 5). 
It varies in size so that between the phenocrystic felsj)ar and that 
of the ground-mass we could find a series of dimensions. 
Zonary structure is typically developed in almost every indivi- 
dual, especially on the tabular section of 010. It contains, as 
usual, glass arranged in zones ; sometimes encloses crystals of 
augite and hornblende, parallel to base and the positive dome ; 
it extinguishes light in symmetrically opposite directions with the 
maximum angle of 30°-34°. The extinction observed on 010 
amounts to 15° with reference to P/M, the trace of the peric- 
line twins making 1.5° with P/M on the same face. These rough 
observations all point to the labradorite-nature of the felspar. 
Hornblende occurs only as the phenocryst and small in quantity. 
It is a brownish-green variety of optically normal character. 
The crystals are all corroded and enclosed by the opacitic 
margin {Fig. 5) which is composed of confused aggregate of crys- 
talloids and grains of monoclinic pyroxene, and clumps of mag- 
netite. The pyroxene appears in tolerably large size that it could 
be opticall}^ ascertained. Sometimes the substance of the margin 
has been replaced by broiniish, double-refractive fibres. In one 
slide the body of the hornblende is impregnated with countless 
swarms of black dots which lend to the crystal a darker shade. 
With high powers, they resolve into glass-enclosures with bubbles. 

Augite occurs sporadically as a phenocryst. Its coarse dis- 
tinct cleavage, 2:)ale colour, and small angle of extinction (less 
than 32°) prominently characterise this pyroxene, and contrast 
pronouncedly with the brown, Andésite augite. That it is diop- 
side is highly probable, but not proven. In one slide, I observed 
porphyritic aggregate of needles, producing the glomeroporphy- 



52 KOTO : NOTES ON THE GEOLOGY 

ritic structure, and they look more like a druse than like a mass 
of crystals, having a mutual relation, characteristic of plutonic 
rocks. Thus our augiie is remarkable in many respects. 

The ground-mass consists mainly of the idiomorphic plagio- 
clase, long-rectangular or square in shape, and of various sizes, with 
some degree of parallel disposition. The square sections of the 
microliths show occasionally truncation of corners by domal 
face and at other times slightly diverge from recta ngularity on 
the edge 001:010. The traces of cleavage run parallel to the 
same edge, and the sutures of twins run vertically. Symmetrical 
extinction takes place at o0°-32° with reference to the same 
trace, showing that the plagioclase stands just at the middle of 
the series between the sodium and the calcium labradorite^^ 
According to Becker, these square sections, which are prismatic 
sections in vertical positions, are very convenient for the deter- 
mination of the microlithic plagioclase. Intermixed with the 
felspar, we find the less idiomorphic crystals of pale augite, 
together with rounded magnetite and the crystals of apatite. The 
cuneiform space left by minerals being filled with the brown glass, 
densely charged with transparent augite. The structure of the 
rock is therefore that kind which toe call the ' orthophyric.^ In 
the ß variety, minute felspar-needles make the greater part of 
the ground-mass, exhibiting the typical pilotaxitic structure. 

a ÄPOANDESITES. 

{PL III, Figs. 1 and 2.) 

One variety is whitish, bleached and compact, the other is 
green through the presence of a chloritic mineral, having a por- 

1) Becker, Amer. jour. ScL, May, 1898. 



OF THE DEPENDENT ISLES OF TAIWAN. Oo 

phyritic structure with the phenocrysts of plagioclase and horn- 
blende. They are much speckled with glittering iron-pyrites 
(a lirge black spot in Fig. 1), which likely attracted the atten- 
tion of Mr. Xarita, who had brought back the specimens to 
Tai-hoku. 

The phenocrystic plagioclase has a tabular form being nearly 
equidimensional. It has a distinct zonary banding, like the pre- 
ceding; rock. Contrary symmetrical extinction of about 33° on 
both sides of the trace of the albite twins shows the plagioclase to 
be a labradorite of a similar composition as in the rocks, just des- 
cribed. Hornblende is entirely decomposed (in the right halves 
of Figs. 1 and 2) into an aggregate of pistacite, chlorite, and 
calcite-films, which together form the pseudomorph after the 
hornblende of a prismatic habit with the combination of 010, 
as ma}^ be conjectured from the original outlines of the now 
altered mass. The chlorite possesses the normal character, and 
pleochroic, showing a green shade parallel to the axis of fibres, 
which corresponds to ';}(. The epidote occurs in tufts and in rugged 
plates. 

The ground-mass consists of very fine laths, simply twin- 
ned, and they are arranged in more or less parallel disposition 
around the phenocrystic felspar. These minute crystals of fels- 
par swim within the chlorite-lamellœ, mixed with the felspar- 
microlite, magnetite and the pyrites, the last does not contain 
any trace of copper. This Apoandesite is no doubt derived from 
the fi variety of the Hornblende-Andésite, already described, by 
the pneumatolytic process which caused the impregnation of the 
pyrites in the rock-mass. 



54 KOTO : NOTES ox THE GEOLOGY 



I). AMPHIBOLIZED GAB URO. 



A dark-greyisli, coarse rock of gabbroitic aspect, in wliicli 
a cleavable hornblende lies after the manner of plutouics, and 
a plagioclase is moulded upon the amphibole. Patches of 
epidote and iron-pyrites complete the list of megascopic elements. 
Under the microscope, the greenish-brown hornblende is for the 
greater part altered into a nearly isotropic lamellse of chlorite, 
calci te-films, and common epidote. The hornblende has been so 
highly altered that the original substance remains but in few 
stripes. The ^/6r^ioc/«st;-anhedra possess only a few twinned 
lamella?, besides the Carlsbad type of twins. Suitable section 
could not be found for ascertaining the nature of the plagioclase. 
The general deep-greyish appearance of the felspar is due to 
the presence of a penuine-like chlorite in the fissure of it. Com- 
mon epidote occupies the place of the felspar and hornblende 
in rugged plate. Crystalloid of apatite, full of air-pores, was only 
once observed. 

I conjecture this rock to be a metagabhro, though a diallage- 
like augite was never seen in my slides. This gabbroitic mass 
probably makes the foundation of the island, and crops out on the 
ivest coast, together Avitli the Apoandesite and Serpentine. 

E. SERPENTINE. 

Associated with the above rock, there occurs a Serpentine 
which is yellowish-blue in its general appearance. Under the 
microscope, the whole mass presents between crossed niçois a 
beautiful lattice-work, Avhich is a characteristic feature of its having 
being derived from an amphibole. There are found intermixed 
with the Serpentine a little quantity of iron-ore. 



OF THE DEPENDEÎs^T ISLES OF TAIWAN. bO 

THE ISLE OF KASHO (SA3IASANA). 

(Fl III, Fig. G.) 

Kasho is a forest- covered, conical volcanic island {J^'ig. 1), 
only 8 km. in circumference, skirted by fringing reefs. The 
inhabitants are of the mixed blood of the Chinese and the 
Malays. According to Mr. Ishii, who gave me a rock-specimen, 
the island is Andesitic, consisting of Pumice and. lava-flows, and 
carries two craters. My slide show^s the rock to be the Hypei^s- 
thene-hornblende- Andésite. 

To the naked eye the rock resembles very closely those of 
Héradaké, in Shinano, and Hakusan in the Kaga province. It 
is greyish-looking, with the only pheiiocryst of hornblende, 
measuring 5 mm. by 2. The hornblende is the largest of 
phenocrysts (on the right half of Fig. ö), broad-columnar in 
form in combination of 110 and 010, and has always thick 
margin of opacité. The hornblende has dark-brown colour, and 
optically normal. It encloses the grains of felspar after the 
fashion of poikilitic plate, especially on periphery. This fact 
conclusively shows the simultaneous crystallisation of the 
hornblende in its later period with the forerunner of plagioclase. 
The formation of these crystals might have taken place at the 
close of intratelluric period, of the magma. The opacitic margin 
consists, as usual, of the grains of monoclinic pyroxene and 
magnetite. They seem to have been formed by resorption and 
re-combination through the gradual caustic action of the surround- 
ing magma upon the already existing hornblende, at a slightly 
lower pressure and in the upper column of effusive lava than 
the situation in which the original amphibole has crystallized 
out. The majority of crystals seems to have been eaten up by 



56 KOTO : NOTES ox THE GEOLOGY 

the magma, so that there remains nothing but the accumulation 
of magnetite-dust in the place of the hornblende. 

A brown pleochroic hypersthene occurs in few quantity, and 
small in size and less idiomorphic when compared with the 
amphibole. Its base shows no axial poles, but symmetrical 
hyperbolas ; it forms penetrating twins upon a domal face, and 
often moulded upon plagioclase. The plagioclase is of tabular 
or long-rectangular shape ; extinguishes symmetrically in the 
direction at about 30° against the trace of the albite-twins, and 
the trace of pericline lamellœ makes -5° to -10° on 010 with P/M, 
indicating the presence of labradorite. The albite-lamallse are 
clear and definite, but the width varies much from one lamella 
to another, and even in the same the width varies from one 
point to another, — these are also said to characterise labra- 
dorite. Zonary banding is pronounced, the interior abounds in 
glass-enclosures, with the clear shell of different optical orien- 
tations. AVe meet often with the lu'oken crystals, from which it 
may be inferred that the rock is a lava-flow. The ground-mass 
consists of a plexus of augite- needles in a colourless base, inter- 
mixed with a somewhat lai-ger plagioclase of a tabular, or long- 
rectangular form, after the manner of a micro-phenocryst. Twin- 
ned slender sections show symmetrically the opposite extinction 
at an angle of 20°, indicating that the felspar in the ground- 
mass is andesine in lieu of the larger, phenocrystic labradorite. 
3Iagnetite abounds in the glassy base. Tridymite fills free spaces 
in imbricated scales. 



OF THE DEPENDENT ISLES OF TAIWAN. Ö7 



CONTENTS. 



Page- 

I'he llôko Group (IVseadores) 1 

I. Intnxluctoiy 1 

II. Stratigraphical Char;;cteristic.s 5 

Hoko Island 5 

General outlines of the ,i:eology of the island 5 

Local details of ideology 8 

Haku-sha Inland 12 

Kippai Island 13 

Gio-o Island 14 

III. Petrography of the Eifusives 17 

^4. Component-minorals of Basalts ■ 18 

Olivine 18 

Plagioclase 22 

Augite 24 

Hypersthene 26 

Analcime and Natrolile 28 

The Iron Ores 30 

B. Special Descrijjtion of Individual Occurrence uf Basalts 33 

a. The granual type 33 

6. The type of the Iddingsite-bearing Basalts 35 

c. The ophitic type 38 

d. The type of the Olivine-less Basalts 39 

e. The type of the Analcite-bearing Basidts 42 

The Island of Kôtô (Kotel-Tobago) 46 

a. Felspar-Basalts 48 

b. Hornblende- Andésites 50 

c. Apoandesites 52 

d. Amphibolized Gabbro 54 

e. Serpentine 54 

Tlte Isle of IvasUô (Saiuasaiia) 55 



PLATE I. 

(PHOTOGRAMS.) 



PLATE L 

Fig. 1. — A fine coni2'>act basait, with comparatively large 
plienocryst of olivine wliicli is more or less iddingsitized. The 
ground-mass consists of small ci-ystals and grains of angites, 
granular olivine and the laths of plagioclase, with the structure 
typically granulitic. Bo-ryo-san, Haku-sha Island. P. oo. 

Fig. 2. — The same rock-type as the preceding, but rather 
coarse. On the right side in the figure is a augititic patch, 
composed of exclusively the crystals of augite in the base. Hoko 
Island, r. :]:}. 

Fig. ^^ — Olivine-less basalt from Ilatto, ^^outhern Group, 
and it probaldy l)elongs to the same type as Figs, o and 4 in 
Plate TL A doubtful olivine is present in the form of chloritic 
patches, but no visible hypersthene. Cleneral mass consists of a 
plexus of fine grains of augite and fine laths of plagioclase in 
the base. This is quite an anomalous rock. P. 3*). 

Fig. 4. — Iddingsite-l)earing basalt with a large idiomorphic 
olivine, externally changing into iddingsite. Magnified ()-") diame- 
ters. Hoko Island. V. ^.l. 

Fig. "). — Rock belonging to the same type as the preceding. 
It is also from Hoko Island. Olivine on the left side of the 
figure shows various stages of iddingsitization. 

Fig. (). — Also iddingsite-bearing basalt, with olivines chang- 
ing from the interior, as may be seen on the lower side of the 
figure. ]\Iagnified 38 diameters and not <;">, as is stated in the 
Plate. Niçois crossed. Kippai Island, 



Koto, The H oka Group, etc. 



Jour. Sc. Coll. Vol. XIII. PI. I. 




V^fi:^^/^.. 










Fig 1. x^^i 



Fig. 2. -x'^i 



y^_ ,»i.3S3^-_^^ 







%S 






Fig. J. 







(;à 



F«V/. i. x6-5 



4^ 







Fi(7. -5. y.86 




Koto phot. 



Fig. (J. xüö -\- niçois 

imp. Tokyo Printing Co., Ltd. 



PLATE II 

(PHOTOGRAMS.) 



PLATE IL 

Fig. 1. — Iddingsite-bearing andésite, magnified Oö diameters, 
showing the typical intersertal structure. Olivine is here changed 
internally into a red mineral, which the writer believes to he 
iddingsite, as is well seen on the lower right octant in the 
figure (pp. V.) and oö). Kippai Island. 

Fig. 2. — The slide of ophitic basalt (p. 38). Sho-chi-kaku, 
the Island of Hôko. 

Fig. o.— OKvine-less hypersthene-bearing basalt, with two 
large crystals of hyperthene in the centre of the figure. The 
structure is granulitic. Tlie Isle of AVam-pai. V. ol>. 

Fig. 4. — The same rock-type as the pi-eceding, but with 
intersertal structure. Local patches of hypersthene, augitc and 
plagioclase, Avith the hyperitic structure. Sei-kei, the Island of 
Hôko. Pp. o'.> and 41. 

Fig. '"). — Analcime-basalt from ]Xai-an, Gio-o. It has granu- 
litic structure. White patches ai'e filled with analcime, and a 
dirty portion at the middle of the field is the secondary natrolite. 
P. 42. 

Fig. <>. — Foraminiferal rock, consisting of discoidal and spiral, 
water-worn shells of Calcarina Spenglerl, besides fragments of 
corals, bivalves and serpula. In natural size. Kippai Island. 
P. i:5. 



Kofd The Halo Group, etc. 



Jour. Sc. Coll. Vol. XIII. PI. II. 





Fig. 1. XÜ. 



Fig- 2. X -M 





Fig. 3. X 24 



Fig. Â. x88 + niçois 




Fig. 5. X 65 




Koto pilot. 



Fig. (i. Nat. size 

imp. Tokyo Frintituj Co., Ltd. 



PLATE III 

(PHOTOGEAMS.j 



PLATE II L 



The Plate III illustrates the type-rocks from the Isles of 
Koto (Botel-Tobago), and Kasho (Samasaua). 

Fig. 1. — Apo 01" altered andésite, magnified O-j diameters, 
shoAving a pheuoerystie hornblende, with opacité margin (on the 
right side of the figure). The hornblende is entirely decomposed 
into an aggregate of pistacite, chlorite, and calcite-films. l^lagio- 
clase is much decomposed. A dark spot (in the lower left octant) 
is the iron-pyrite (p. '"i^). 

Fig. 2. — The same slide under crossed niçois. 

Fig. o. — A porous, greyish-l)rown basalt, Avith a rather large 
corroded olivine (on the left of the figure). The ground-mass, 
which encloses a corroded diopside-like augite, is highly fels- 
pathic. The structure is hypocrystalline-porphyritic, approaching 
to the pilotaxitic type (p. 48). 

Fig. 4. — Another basalt, with abundant olivine of various 
dimensions. It contains globulitically granulated mesostasis, and 
the structure is navitic (p. -30 ). 

Fig. '). — Hornblende-andesite, with dark hornblende-crystals, 
surrounded by opacitic margin (on the upper and the lower end 
of the figure). The structure is orthophyric (p. '30). 

Figs. l-'3 are all from the rocks of Koto. 

Fig. (). — Hypersthene-hornblende-andesite from the Isle of 
Kasho, with a large ]jlienocryst of hornblende (on the right 
half of the figure). It is enclosed by a thick margin of opacité, 
but enclosing the grains of plagioclase after the fashion of 
poikilitic plate (p. 55). 



Kofo, The lioko Group, etc. 



Jour. Sc. Coll. Vol. XIII. PI. III. 





Fir/. 1. x(i 



Fi(/. 2. X oH + iticol^ 





Fig. o. X o'6' 



Fi(/. 4. X o.S 




Fig. 5 X3Ö 




Koto phut. 



Fig. 6'. xJiS 

imp. Tokyo I'rintiiuj Co., Ltd' 



PLATE IV. 

(MAP.) 



PLATE TV. 



The Plate IV sliows the bathometric condition of tlie neigh- 
bouring seas of the Pescadores or Hoko Gronp. It seems to me 
that the North Group forms itself an independent centre of 
extravasation of magma, in contrast to the South or Rover Grouj), 
from which the Northern is separated by the incurve of the 
forty fathom-line, — the position indicated by the Rover Channel. 
Both groups are, however, located at the north-eastern end of 
the Formosa Bank, which is disconnected on the east from 
Taiwan by a channel of the same name (j). ']). 



THE IIOKO (iKOri*. 

Kotôjhe Hôko Group, etc. Jour.Sc. Coll. Vol. XIII. PI. IV. 



■,0^ 






,^^ 






~~) ( 









f'f^/f/-/fff/ ri/fk 



'!ytJll/lt'/ /YV-/' 'J 






r<fA-f<//f 



Kj 



//U'/i.loiiy. 



-10 1=^^^310-20 f- —?^^ 20-30 

Sr;ilc I . (•>!',(),()()() 



I'-- •■ . ' i30-5( 



Lf-laiJ\ 



' A./ff/'y. 



PLATE V, 

(GEOLOGICAL MAP.) 



PLATE V. 

The Plate V is intended to show tlie geographical dis- 
tribution of the Tertiary basalts with intercalated sedimentaries 
and several Recent formations, one among the latter being the 
coral-reefs which fringe the coast all round. The topographic 
basis for the geological map is compiled by myself from various 
sources, the data being supplied chiefly by Mr. Y. Saito, who 
also ofiered me assistance in colouring the geologic elements 
represented on the map. 



MW OF THF IIOKO (JKOl P. 

Kotô, The Hôko Group, etc. Jour Se Coll. Vol, XIII. PL V 




ZZ] [^DLZZ] 



I 



Change of Volume and of Length in Iron, Steel, 
and Nickel Ovoids by Magnetization. 

By 

H. Nagaoka, Rigahiliakushi, 
Professor of Applied Mathematics. 

AND 

K. Honda, Rigakushi, 
Post-graduate in Physics. 



With Plates VI. & VII. 



1. In our former paper/^ we described some effects of 
magnetization on the dimensions of nickel and iron, as well as 
those of hydrostatic pressure and longitudinal pull on the mag- 
netization. We then showed that there is a reciprocal relation 
between the two, and that the Villari effect in iron is a natural 
consequence of the observed changes of dimensions. Unfortunate- 
ly on that occasion the range of the magnetizing field was 
limited to a few hundred C.G.B. units, so that the investigation 
of the behaviour of these metals in high fields was reserved for 
further experiments. In addition to this, the ferromagnetics were 
not of a shape to be uniformly magnetized with the exception 
of the iron ovoids. It was therefore thought desirable to repeat 

]) Nagaoka and Honda, Journal of the College of Science, 9, 353, 1898; Phil. Mag. 
46, 262, 1898. 



58 H. NAGAOKA AND K. HONDA. 

the experiment on uvoids of ferromagnetic metals, and so to 
extend the investigation into still stronger fields. 

2. In his well-known researches on the changes of dimensions 
of iron and other metals by magnetization, BidwelP' pushed the 
Held strength to loOO ; in the present experiment, the field 
strength is greater than that of Bidwell by 700. In addition to 
ordinary soft iron and steel ovoids, wolfram steel from Bohler 
in Vienna was tested with a result which shoTved a remarkable 
difierence from ordinary steel as regards the change of dimen- 
sions wrought by magnetization. As Avas generally supposed, the 
change of volume is very small in iron and nickel in weak 
fields, but with strong magnetizing force the eftect becomes 
generally pronounced. 

3. The apparatus already described w^as used in measuring 
the change of length and of volume. A small alteration was 
made in the arrangement of the magnetizing coil. Owing 
to the strong magnetizing current, special arrangements w^ere 
made for keeping the interior of the coil at a constant tem- 
perature. A double walled tube of brass was inserted in the 
coil, and a constant stream of cold water was passed in the 
interspace for more than an hour l3efore each experiment. As 
the resistance of the coil was only 0.56i?, the rise of temperature 
was so small, that the ferromagnetics placed in its core w^ere 
scarcely affected. The change of length was measured by an 
optical lever, as before described."* For measuring the change of 
volume, the ovoid was sealed in a glass tube with a capillary neck 
(internal diameter about 0.4 mm.) and so placed in the tube 



1) Bidwell, Phil. Trans. 179, 205, 1889. 

2) Nagaoka. Phil. Mag. [5] 37, 131, 1894; Wied. Ann. 53, 487, 1894.; Nagaoka and 
Honda loc. cit. 



CHANGE OF VOLUME AND OF LENGTH. 



59 



that it rested in the nxial line, and never came in contact with 
the wall of the tube. The magnetizing coil and the tube were 
placed in a horizontal position. The motion of the meniscus 
Avas measured by a microscope provided with a micrometer ocular. 
For more minutely detailed particulars, we must refer the reader 
to the former paper. 

3. The followins: are the dimensions of ovoids used in the 
present experiments : 



Specimen 

No. 


Metal 


a (cm.) 


c (cm.) 


V (c.cm.)l 


f 


N 


1 


Nickel 


0.7.50 


12.50 


31..50 


8.86 


0.125 


2 


)) 


0.500 


10.00 


10.48 


8.86 


0.0848 


3 


Soft iron 


0.750 


12.50 


31.45 


7.84 


J 0.125 


4 


)} 


0.500 


10.00 


10.53 I 


7.83 


0.0848 


5 


Ordinary steel 


0.750 


12.50 


31.60 


7.83 


0.125 


6 


}> 


0.500 


10.00 


10.57 


7.81 


i 0.0848 


7 


Wolfram steel 


0.7.50 


12.50 


31.82 : 


7.90 


' 0.125 


8 


jj 


0.500 


10.00 


10.53 


7.95 


0.0848 



a gives the semi-minor axis, c the semi-major axis, v the volume, 
f> the density, and iY the demagnetizing factor of the ovoids. The 
volume of each specimen was measured by weighing the ovoids 
in water. 

The elastic constants of the metals were measured by flexure 
and torsion experiments on rectangular prisms made from the 
same specimens as the ovoids. The prisms were 14.6 cm. long 
and 0.896 cm. square in cross-section. 



GO 



H. NAGAOKA AND K. HONDA. 



Metal 


E (C.G.S.) 


K (C.G.S.) 


d 


Nickel 


2.07 X 10^- 


0.771 X W 


1.082 


Soft iron 


2.10x10'- 


0.800 X 10'- 


0.844 


Steel 


2.04 xlO^^ 


0.838 X 10"- 


0.384 


Wolfram steel 


2.02 X 10^- 


0.849 X 10'- 


0.306 



K gives Young's modulus, A' ( = n. Thomson and Tait) tlie 

modulus of rigidity, and d a constant defined by the equation 

J^/ ljf2M_,. 
2 I 1 + 3^ )~'^- 

The magnetization of each of these ferromagnetics was deter- 
mined by the magnetometric method, after the ovoids had been 
carefully annealed, with the following results : 



Niclvel (2) 
H I 


Soft iron (4) 
H I 


Steel 
H 


(6) 
I 


Wolfram steel (8) 


H 


I 


0.7 


24.2 


1.0 


62 


1.9 


23 


2.7 


18 


1.4 


49.8 


2.5 


160 


2.4 


44 


6.8 


(^5 


3.0 


138.6 


4.3 


291 


6.9 


183 


12.6 


193 


5.4 


238.0 


9.5 


587 


9.7 


279 


20.2 


498 


10.9 


336.8 


12.7 


750 


13.1 


385 


25.8 


748 


37.8 


395.7 


19.9 


948 


23.3 


651 


44.5 


992 


74.1 


420.0 


37.2 


1111 


32.3 


815 


83.6 


1116 


125.3 


434.5 


99.6 


1255 


50.2 


984 


118.0 


1170 


171.6 


438.7 


155.5 


1309 


116.3 


1196 


191.0 


1224 


240.3 


440.7 


270.3 


1400 


174.4 


1260 


344.6 


1301 


481.4 


443.4 


433.6 


1479 


345.0 


1379 


512.3 


1348 


674.2 


444.5 


584.6 


1520 


520.2 


1440 


ßQQ.C> 


1373 


914.0 


446.8 


792.8 


1546 


873.7 


1489 


940.3 


1400 


1233.0 


447.7 


992.6 


1562 


1149.8 


1512 


1213.3 


1423 


1747.0 


448.7 


1585.8 


1607 


1822.6 


1549 


1674.9 


1452 



CHANGE OF VOLUME AND OF LENGTH. 61 



Change of Length. 

4. Iron (Fig. 1). — The change of length experienced by soft 
iron is too well-known to need any description. The ovoid 
elongates in weak fields till it attains a maximnm, being longer 
by about 3- to 4-millionths of its initial length ; it then decreases 
in length and becomes shorter than in the unmagnetized state. 
The contraction goes on gradually increasing, and, in the present 
experiment, it does not seem to reach an asymptotic value, 
even in fields of 2200 C.G.S. units, wehere the contraction 
amounts to about îoôoôô* '^^^^ present result agrees qualitatively 
w^ith Bidwell's experiment, but the contraction is much greater. 
The discrepancy is perhaps to be chiefly accounted for by the 
difference of shape. 

5. Steel (Fig. 1). — Ordinary steel behaves just like iron, the 
difference being the smallness of elongation and contraction, while 
the field at which the elongation vanishes lies in the stronger. 
The field of maximum elongation in wolfram steel is greater 
than in ordinary steel or iron, that of no-elongation in the unan- 
nealed state being several times greater than in iron or ordinary 
steel. Such a field lies in H=1200. When the wolfram steel 
is annealed, the retraction after reaching the maximum takes 
place very slowly and the characteristic as regards the field of 
no elongation becomes exceedingly pronounced. From the curve 
of length change, it does not appear that it will ever cut the 
line of no-elongation even in intense fields. 

6. The curve of elongation (in dots) plotted against the in- 
tensity of magnetization is given in Fig. 1. The change of length 



62 H. NAGAOKA AND K. HONDA. 

at first takes place very slowly, but on reaching saturation, the 
rate of decrease becomes very rapid. So far as the present ex- 
periment goes, the rate does not diminish except in annealed 
wolfram steel, in which we notice a slight flattening. 

7. Nickel (Fig. 1). — The behaviour of annealed nickel ovoid 
as regards the length change is nearly the same as that already 
observed by one of us. With an increasing field, the contraction 
reaches an asymptotic value, which in the present case is greater 
than that obtained by Bidwell from experiments on a nickel wire. 
The explanation of this discrepancy is to be sought for partly in 
the difference of shape, and partly in the difterence of treatment, 
as will be clearly illustrated by experiments on the change of 
volume. We have also reason to believe that repeated annealing 
alters the elastic behaviour of ferromagnetics as regards the 
strain wrought by magnetization. Plotting the curve of length 
change against the intensity of magnetization, we find a slight 
bend when the magnetization becomes saturated and the con- 
traction approaches its asymptotic value. 

Change of Volume. 

8. Experiments by several physicists prove that magnetiza- 
tion produces change of volume in ferromagnetics, in contradiction 
to the popular belief which is based on Joule's experiment. The 
alteration of volume accompanying the magnetization of ferromag- 
netics is generally very small in weak fields, but as will be seen 
from the present experiment, the phenomenon becomes more marked 
as the field is made stronger. As we have already remarked, the 
change of volume as measured by Cantone^' in an iron ovoid must 

1) Cantone, Mem. delîa K. Accad. dei Lincei 6, 487, 1891. 



CHAXGE OF VOLUME AXD OF LENGTH. 63 

have been exceedingly minute as the magnetizing field was very 
small. Dr. Knott^^ has published several papers on the change 
of internal volume of ferromagnetic tubes, showing that iron, 
nickel, and cobalt are subject to the change by magnetization. 
As our former result regarding the same question was somewhat 
different, especially in the case of nickel, we have thought it 
advisable to settle the discrepancy by fresh experiments. 

9. Iron and Steel (Fig. 2). — Preliminary experiments on soft 
iron and steel ovoid showed that considerable increase in the volume 
change takes place as the ovoids are annealed. The increase 
becomes more significant as the field is made stronger. In steel, 
the effect of annealing is greater than in iron. In strong fields, 
the volume change of the annealed steel ovoid is nearly twice as 
great as in the unannealed state. AYolfram steel is very little 
affected by annealing as regards the volume change, but the 
change itself is much greater than in nickel or iron. The 
motion of the capillary meniscus in the dilatometer can be easily 
followed by the naked eye. The curves in Fig. 2 have been 
plotted from measurements made on annealed ovoids. 

10 Nickel (Fig. 2). — As specimens of nickel almost always 
contain traces of iron, the change of volume will probably depend 
on the chemical nature. In addition to this, the mechanical 
process which the metal had to undergo before it could be brought 
to a form suitable for experiment, nmst have substantially altered 
its elastic behaviour. 

The nickel rod, which we used in the former experiment, 
was hammered from a nickel plate to a prism of square cross- 
section. It contained 1.7Ö <^/o of iron, besides traces of man- 
ganese and carbon. The ovoids used in the present experiment 

1) Knott, Trans. Koy. Soc. Edinb. 38, 527, 1896; 39, 457, 1898. 



64 H. NAGAOKA AND K. HONDA. 

were prepared from a thick plate, and were nearly pure nickel, the 
quantity of iron present as an impurity being inmeasurably 
small. As the material is likely to become homogeneous by 
repeated annealing, the ovoids were carefully annealed for 
about 50 hours. The ovoid was wrapped in asbestus and 
placed in a thick metal tube, the interspace between the ovoid 
and the wall of the tube being filled with fine charcoal powder. 
The tube was then placed in charcoal fire. When the ovoid was 
annealed in this way, there were some traces of surface oxidation. 
The change of volume after each annealing was examined with 
the result that it became evident that the process of annealing 
increases the etfect. It therefore appears that the previous history 
of the specimen exercises an important effect on the magnetization 
and on the dimensions of ferromagnetics as affected by magneti- 
zation. The anomaly in the length change noticed by Bidwell 
in two specimens of nickel wire is probably not the effect of 
temperature, but is perhaps to be ascribed to the cause above 
stated. In contradiction to our former result with a square 
prism, the ovoid showed increase of volume. The amount of in- 
crease was small compared with the decrease noticed in the 
previous experiment. Cantone^^ obtained a tolerably large increase 
of volume in nickel ovoids ; our former result was nearly half as 
large, while in the present experiment, there is a slight increase. 
The discrepancy is probably due to the difference of treatment 
before the specimen can be converted into a proper shape for 
experimenting, and also to its chemical composition. 

11. The volume change of ferromagnetics considered as a 
function of the magnetizing field takes place very slowly in weak 
fields ; it then increases in a more rapid ratio till it reaches the 

1) Cantone, Atti della K. Accad. die Lincei, 6, (1), 257, 1891. 



CHANGE OF VOLUM r: AXD OF LENGTH. 65 

' -Wendepunkt '; after tliat the change becomes slower, but still goes 
on increasing nearly in a straight line. Up to H=2000, the 
rate of change shows no tendency to decrease. Witli a still 
stronger held, the increase of volume will prol)al)ly become 
more considerable. 

12. In our former experiment, the range of the magnetizing 
force was confined to a few hundred C. G. S. units. In the present 
experiment, the increased field strength unveiled the character of 
the change of the volume considered as function of the intensitv 
of magnetization. As will be seen from the curves (Fig. 2.) in 
dotted lines, the increase in nickel and steel takes place quite 
slowly befoi'e the magnetization reaches saturation. As soon as 
the magnetization reaches this state, the increase becomes very 
rapid, so that the branch of the curve ascends nearly parallel to 
the axis of volume increase. There we find that a sliolit increase 
in magnetization is attended with a laige increase of volume. 
As the rate of increase appears to be nearly constant, it would be 
very interesting, if we could push the field strength still farther to 
see whether the volume change ultimately attains an asymptotic 
value. 

The observed changes of volume and of length are exhibited 
in the followins: table : — 



]) Cantone, Atti della R. Accad. dei Lincei, 6, (1), 257, 1891. 



GO 



IT. NAGAOKA AND K. HONDA. 



Nickel (1) 


Soft 


iron (3) 


Steel (.3) 


AVoUram steel (8) 


H 


OV 
V 


H 


V 


H 


Si' 

V 


H 


Si. 

V 


13 


0.09 X lö' 


8 


0.10 xlÖ' 


7 


0.08 X lö' 


19 


0..30 X lÖ' 


30 


0.29 


11 


0.52 


12 


0.47 


42 


1.52 


90 


0.65 


18 


1.5G 


•J.> 


1.95 


93 


3.03 


218 


0.82 


ICI 


3.12 


192 


3.13 


216 


5.01 


282 


0.97 


443 


3.85 


376 


4.69 


442 


8.04 


517 


1.38 


G91 


4.58 


586 


6.22 


692 


11.68 


877 


2.0G 


958 


5.88 


792 


8.01 


1001 


16.68 


1141 


2.44 


1115 


7.18 


1044 


10.16 


1117 


18.96 


1547 


3.24 


1342 


!».47 


1376 


14.07 


1296 


22.75 


1740 


3.53 


1563 


11.45 


1646 


17.20 


1704 


28.82 


2253 


4.12 


2089 


14.C8 


2171 


22.20 


2153 


32.62 



Nickel (2) 


Soft 


iron (4) 


Steel (6) 


Wolfram steel (8) 


H 


cl. 

I 


H 


S/ 

r 


H 


s/ 

r 


H 


0/ 

/ 


4 


- 14.1 xfd 


6 


2.5x10 


13 


3.1 X 10 


18 


4.1 X fd 


6 


- 64.0 


15 


19.0 


19 


7.1 


25 


12.4 


10 


-118.2 


5L 


31.6 


28 


15.1 


39 


21.7 


15 


-163.6 


127 


23.8 


54 


22.2 


62 


28.9 


33 


-217.5 


224 


3.1 


96 


23.1 


106 


32.1 


59 


-264.3 


354 


- 17.7 


160 


17.8 


170 


32.3 


124 


-317.6 


575 


- 52.6 


225 


8.2 


235 


31.7 


302 


-343.6 


698 


- 62.6 


374 


-11.5 


349 


30.2 


561 


-z:)3.s 


883 


- 73.5 


.589 


-36.0 


592 


23.1 


839 


-356.0 


1077 


- 78.9 


844 


-49.2 


781 


18.7 


1145 


-360.0 


1180 


- 82.2 


1061 


— îj^-).^) 


1052 


17.0 


1289 


-360.9 


1324 


- SG.G 


1177 


-59.5 


1188 


15.4 


1483 


-362.2 


1447 


- 89.9 


1361 


- 64.5 


1345 


13.9 


1849 


-362.7 


1538 


- 91.6 


1729 


-69.9 


1697 


12.4 


2322 


-365.3 


2 1 80 


- 102.0 


2172 


-78.1 


2235 


10.9 



CHANGE OF VOLUME AND OF LENGTH. 



67 



Kirchhoff's Constants /' and /.". 
lo. »^üirtiiii'- from the formuUe 



and T^ =]-/.-' + 



00 L/.. , -(A--^-') jy' 



II- 



"4 4 j A'(l + 3^)' 

wliicli give the change of volume and of length of ferroiiuigLietio 
ovoids in terms of Kirchhofes constants // and //', ayc obtain the 
following expressions for these two constants : 



and ]:" = 



2(1 + 3^) 
37 -jj 



2{l + 'Sd)' 

1 4/i:(i+3ö) _^,_7..^o/ 

where p= — -fri ^ + ^~'' + ■^'•^ 



H' 



„„a ,=--Mi+ML;+^?^a+,o+Ä-. 

M' o 

These constants, as calculated from the change of dimensions 
of ovoids, are given in the following table, and gra])hieally drawn 
in Fig. o : 



H 


^sick'l 




Soft 


iron 


St 


eel 


Wolfl 


;iin steel 


/.•' 


/ù" 


¥ 


k" 


/.:' 


//' 


7/ 


k" 


5 


-22U100 


712S00 


21900 


-22610 


1017 


-1865 


348 


-]'252 


10 


-1SS90O 


57SÜ0O 


2;!520 


-28450 


;!840 


— 3322 


986 


-1512 


20 


- 710(10 


216700 


132S0 


-16420 


4248 


-4615 


3.600 


-39s 3 


no 


- ;!G370 


111200 


7302 


- 8050 


4048 


-5080 


4881 


-5440 


60 


- 8163 


34540 


2139 


— 2222 


1738 


-1864 


1946 


-2217 


80 


- CODS 


20060 


1207 


- 1102 


1069 


-1004 


119S 


-13,85 


100 


- 4(;o:î 


14120 


75J 


- 530 


701 


- 546 


794 


- 880 


120 


- 3;;7;> 


10260 


500 


- 255 


477 


- 279 


557 


- 595 


160 


- 12117 


3968 


239 


18 


240 


- 16 


315 


- 3>17 


250 


- 84;; 


2553 


55 


175 


69 


117 


128 


- 109 


300 


- 501 


1700 


25 


175 


38 


124 


88 


- 66 


500 


- 21G 


655 


-9 


130 


-I 


99 


28 


— 8 


800 


- 80 


259 


-9 


70 


-6 


57 


8 


5 


1200 


- oU 


117 


-6 


37 


-4 


3.1 


3 


5 


1600 


•)•_> 


60 


-4 


23 


1 -'^ 


19 


1 


4 


2U0I) 


- 14 


42 


-3 


IC. 


! -2 


13, 


(1 


•■' 



6S H. NAGAOKA AND K. HONDA. 

14. The curves for // and //' present the same general feature 
in iron and steel. ,'/ increases inl[io\v fields; and tliere attaining 
tlie maximum v^alue, it rapidly diminishes till it becomes less than 
zero ; it then reaches a minimum, after Avhich it again gradually 
increases. The exact position of the minimum is very vague ; the 
curve for // ultimately coincides with the axis of IT. ]/' is at first 
negative, and attaining the minimum value, goes on gradually 
increasing till it becomes greater than zero, and then reaches a 
maximum. AVith the farther increase of the field, the value of //' 
decreases very slowly. The position of maximum for k' and that of 
minimum for //' lie nearly in the same field, which is greater for 
wolfram steel than for soft iron, while that for ordinary steel oc- 
cupies an intermediate position. The absolute value of // and //' 
is greater in iron than in steel. In nickel, the values of // and //' 
are far greater than those for iron and steel, and moreover are of 
opposite signs. The maximum of /,", or the minimum of /.', seems 
to lie in a weak field ; the rate of decrease or increase is quite 
rapid and the curves for // and Z " soon approach the axis of//. 
Compared with the results of former experiments, the absolute 
values of /,' and /." are generally small for iron, — far greater for 
nickel. This difterence arises from the fact that for iron, the 
change of length in weak fields is less in this case than in the 
former experiment, and that for nickel the contrary is the case. 
As regards the sign, these two experiments show fair agreement. 

Consequences of the theory. 

15. Effect of longitudinal pull. — The change of magnetization 
produced by the elongation of a wire can be easily calculated 
from the formula 



CHANGE OF VOLUME AXD OF LENGTH. 



69 



.i=zz{,4_3(.'+-^-)|;. 

Putting / = 4.67x lO"", 4.80 x lO-", and 4.8.5 x 10"" for soft iron, ordin- 
ary steel, and wolfram steel respectively, each corresponding to a 
pull of 0.1 Kilog. per sq. mm., we get the following results : 



II 


Soft iron. 


Steel. 


Wolfram .steel. 


ol 


oI 


dl 


10 


0.919 


0.055 


0.044 


20 


1.074 


0.212 


0.170 


30 


0.831 


0.402 


0.3.50 


GO 


0.399 


0.2.54 


0.294 


SO 


0.244 


0.1.53 


0.24!) 


100 


0.127 


0.072 


0. 188 


120 


0.039 


0.005 


0.145 


1(!0 


-0.080 


-0.092 


0.094 


200 


-0.1G4 


-0.14G 


0.0 Gl 


300 


-0.2.58 


-0.210 


0.017 


500 


-0.311 


-0.23G 


-0.002 


800 


-0.275 


-0.205 


-0.038 


1200 


-0.219 


-0.JG3 


-0.037 


IGOO 


^0.183 


-0.134 


-0.033 


200) 


-0.157 


-0.115 


-0.028 



\ 



It will be seen from tlie above table that tliere is an increase 
of magnetization in low fields, till it reaches a maximum, after 
which it gradually decreases. The decrease does not proceed 
continuously, but reaches a maximum, whence the magnetization 
begins to recover. Although the former result here arrived at 
is the well-known Yillari effect, we do not know whether the 
maximum decrease due to longitudinal stress has as yet been 
experimentally ascertained. AVith nickel, we obtain the following 
values for the change of magnetization due to elongation, /= 



70 



H. NAGAOKA AND K. HONDA. 



4.74 X 10"^ wLicli corresponds to a pull of 0.1 Kilog. per sq. mm. 



H 


Ol 


10 


-24. .58 


20 


-18.87 


30 


-14.16 


60 


- 9.09 


80 


- 7.12 


100 


- .3.99 


120 


— .3.22 


160 


- 2.70 


300 


- 2.28 


500 


- 1.39 


800 


- 0.88 


1200 


- O.GO 


1600 


- 0.4.3 


2000 


- 0.3G 



There is nothing remarkable in nickel. Longitudinal pull 
produces decrease of magnetization, which becomes gradually less 
as the field strength is increased. This is such a well established 
experimental fact that we need not enter into further discussion 
of the subject. 

16. Effect of Jiijdrostatic pressure. — We can easily see that 
the change of magnetization ol due to change of volume a by 
hydrostatic pressure is given by 



<?!=-( Z:'+-'|'-)//^ 



If we calculate the chan<>;e of mairnotization due to contrac- 
tions 4.68 x 10"'\ 5,oö x 10", 8,42 x 10 " and t),o3 x 10 " for nickel, 
soft iron, steel, and wolfram steel respectively, each corresponding 
to a pressure of 10 atm., avc obtain the following values : 



CHANGE OF VOLUME AXD OF LENGTH. 



71 



H 


Nickel 


S(ift iron 


Steel 


Wolfram steel 


ÔI 


(57 


dl 


rî/ 





0.000 


0.000 


0.000 


0.000 


10 


0.190 


0.757 


0.230 


0.045 


20 


0.119 


0.840 


0.457 


0.237 


.30 


0.080 


0.713 


0.r)9~) 


0.8.-)8 


GO 


0.037 


0.39S 


0.5G5 


0.G75 


80 


0.030 


0.3G2 


0.495 


0..5.50 


100 


0.024 


0.305 


0.437 


0.4G7 


200 


0.012 


0.178 


0.2G8 


0.2.39 


300 


0.008 


0.1.35 


0.200 


0.185 


500 


0.00.-) 


0.093 


0.134 


0.118 


800 


0.002 


0.062 


0.088 


0.073 


1200 


0.001 


0.042 


O.OGl 


0.048 


] GOO 


0.000 


0.031 


0.043 


0.034 


2000 


0.000 


0.024 


0.034 


0.02G 



It thus appears that in nickel the effect of hydrostatic pressure 
is very small compared to that of longitudinal pull. There 
is increase of magnetization -with the volume contraction of 
the magnet. Such an increase reaches a maximum in low 
fields, whence the effect gradually diminishes. Similar changes 
are also noticed in the case of iron and steel. In our former 
experiment, we found that hydrostatic pressure increases the 
magnetization in nickel, while it decreases it in iron. The 
agreement between theory and experiment is very close in nickel, 
but there is a wide di?icrepancy in iron and steel, as we have 
already noticed. 

17. Effect of torsion on longitudinally or cireularhj magnetiz- 
ed wire. There are other important consequences to be drawn 
from the constant //' with regard to the effect of torsion on 



72 II. NAGAOKA AND K. HOXDA. 

loiigitiulinally mngiietized -wire and on ferromagnetic wire travers- 
ed by an electric current. The strain caused hy twisting a 
circular wire can be resolved in elongation and contraction in 
directions perpendicular to each other and inclined to the axis 
of the wire at 45°. Taking these two principal axes of the strain 
for those of x and v, we have for the strain. 



~d^= ^"^"^ 



du 
dx 

dv - 

-- — =—i(or, 
dij 

where (>^ denotes the amount of torsion and r the distance from 
the axis. Resolving the magnetizing force which is in the direc- 
tion of the axis of the cylinder, along the axis of elongation and 
of contraction, we find that the circular magnetization which will 
be called into play is equal to —\cürl:"H at a distance r from 
the axis, the mean circular magnetization being -(oVHB, where 
R is the radius of the wire. 

The transient current which will be thus induced in the w^ire 
by suddenly twisting it is proportional to —l"H. 

Next suppose that the wire is traversed by an electric current 
of intensitv G. Then the circular maofnetizino; force at a distance 
r from the axis is 

By fipplying similar reasoning, we find that the mean longitudi- 
nal magnetization is equal to - wl" C . We therefore conclude 
that twisting the \Yire carrying the electric current gives rise to 
longitudinal magnetization proj)orlional to —l-"C. Thus the 
circular magnetization produced by tw^isting a longitudinally 



CHANGE OF VOLUME AND OF LENGTH. 73 

magnetized wire has a reciprocal relation to the longitudinal 
magnetization caused by twisting ;i circularly magnetized wire/* 

The view propounded by Prof. Ewing"^ to account for the 
existence of transient current by means of frolotropic suscepti- 
bility is similar to what would follow from KirchhofV's theory, 
but it fails to give the amount of the current or of the magnetiza- 
tion which would be produced by twisting. 

The theoretical inferences which we can draw at a glance 
from the curves of — /." II (Fig. 3) are as follows : 

1. The transient current as well as the longitudinal magneti- 

zation produced by twisting an iron or steel wire is 
opposite to that produced by twisting one of nickel, up 
to moderate fields. 

2. The transient current as well as the lon2;itudinal mao-neti- 

zation produced by twisting an iron, steel, or nickel wire 
reaches a maximum in low fields. 

3. In strong fields the direction of the current as well as 

the longitudinal magnetization is the same in iron, steel, 

and nickel. 
It has been established bv G. Wiedemann'^ that the lona;!- 
tudinal magnetization produced 1)y twisting an iron wire carry- 
ing an electric current is opposite to that produced in a nickel 
one. The opposite character of the transient current in these two 
metals has also been observed by Zehnder*' and independently by 
one of us"'. The existence of a maximum transient current in 

1) Voigt, Kompendium der ihenrethchen rinjsik, 2, 20o, 189(5, Leipzig ; Drude, Wieil. Ann. 
63, J^•, 1897. 

2) Ewing, rroc. Eo:;.'_Soc. 36, 1884. 

3) AViedemann, Elect ricitüt, 3. 

4) Zelnider, Wied. Ann., 38, C8, 1889. 

5j Nagaolva, Phil. Mag. [5] 29, 123,1890; Juunial of tlie College of Science, TGkyG, 3, 
335, 1890. 



74 H. NAGAOKA AND K. HOXDA. 

tlieso two metals lins been clearly establisliecl, altliongli there 
is some difference in tlie field strength between iron and 
nickel. It appears from the experiments of Dr. Knott'* that the 
area of tlie hysteresis curve in the longitudinal magnetiza- 
tion produced by twisting circularly magnetized wire reaches 
a maximum as the field strength is increased ; but on account 
of the feebleness of the current, the existence of the maximum 
in the longitudinal ma2;netization is not well established. To 
judge from the course of the curve given by the same experi- 
menter, it seems highly probable that the maximum would be 
reached if we could push the circularly magnetizing force a little 
farther. The conclusion (3) is still an open question, although 
some experiments of Matteucci'' seem to corroborate the view 
just stated.'"^ 

1(S. Looking at the curves of /.•"//, we cannot but be struck 
with the close resemblance of the curves representing the amount 
of torsion produced by the combined action of the circular and 
the longitudinal magnetizing forces on a ferromagnetic wire. AVe 
can no doubt co-ordinate the effect of torsion on a magnetized 
wire with the Wiedemann effect. The discussion of the last men- 
tioned effect we hope to lay before the public in the near future. 

In spite of the qualitative explanations which Kirchhoff's 
theoi-y afïbrds with regard to the efîect of longitudinal pull, of 
the hydrostatic pressure, and of torsion, there are instances in 
which tlie theory apparently fails in several quantitative details 
that it necessarily calls for modification. We may remark that // 



1) Knott, Trans. Eoy. Soc. Edinb., 36, 485, 1891. 

2) Mattencci, Annales de Chimie et de Physique, 1858. 

•î) "While tliis jiaiier was passing through the press, we fcmnd that the direction of tlie 
transient current prnduceil hy twisting a magnetized iron wire is reversed in strong magneti- 
zing fields. 



CITAXGE OF VOLUME AND OF LENGTH. /Ö 

nnd /." are physically functions of the strain, as is l)Orne out by 
the numerous experiments on the effect of stress on magnetization. 
The present state of the theory of magnetostriction may perhaps 
be compared witli that stage in the history of the theory of 
magnetization when the intensity of magnetization was supposed 
to be simply proportional to the magnetizing force. In fact, the 
theoi'y is still in its infancy, so that there are ample grounds 
for expec!ing further developments on further researches. 



Jour Sc Coll. /Ol. XIII. PI VI. 



























il „. ■ 1 ■ T ■ ■ - — ' 












*J 1 '■ , II / 




- "-3i ^«ir-"-^*?^^"" -++ -1- ^ 










- ^ \ r^-f - - - 






»D [ic,^ T^r^i'i' ^ +-H--- ,-- 














i ^ - 1- .'•..ficrA .^, - - ^^ _ ^ 






















\ -t J 1 i - ------- 






rrWrrrH 








1 




Jour Sc Coll Vol XI 1 1 PI. VII 




Combined Effect of Longitudinal and Circular 

Magnetizations on the Dimensions of 

Iron, Steel and Nickel Tubes. 

By 
K. Honda, Rigakushi, 

Post-£;raduate in Plivsics. 



With Plates VIII. ami IX. 



1. The change of length in the direction of magnetization has 
been made a subject of investigation by several experimentalists, 
but few of them have measured the change in the direction per- 
pendicular to that of magnetization. Joule^^ first observed the 
diminution of length of an iron gas-piping by passing a current 
through an insulated wire inserted into it, and bent over the 
sides, so as to form a circular magnetizing coil of le convolutions. 
His experiment was modified by BidwelP^ who measured the 
change of dimensions in an iron ring. He found that the ring 
becomes thicker in a strong field and thinner in a weak one. 
From the measurement of the internal as well as the external 
change of volume for iron, steel, nickel, and cobalt tubes, Knott^' 

1) Joule, Scientific papers I, 263. 

2) Bidwell, Proc. Eoy. Soc. 56, 94, 1895. 

3) Knott, Trans. Koy. Soc. 39, 457, 1898. 



78 K- HONDA. 

calculated the change of lateral dimension in these tubes. The 
result for iron coincided qualitatively with that of Bidwell. 

The first experiment on the change of length of an iron 
wire by the combined action of longitudinal and circular mag- 
netizations Avas made by Beatson'^ who observed the diminution 
of length at the moment when an electric current was passed 
through a magnetized wire. A similar result was afterward 
obtained by Righi.-^ The same experiment was also repeated by 
Bidwell/' who observed a large increase in the change of length 
by longitudinal magnetization of an iron wire carrying a current. 

2. Through the kindness of Prof. Nagaoka, his apparatus^^ 
for the [measurement of the minute change of length was placed at 
my disposal. The apparatus consists of a small optical lever with 
an arrangement for temperature compensation on the same prin- 
ciple as the gridiron pendulum. The rod, by which the change 
of length is made sensible to the lever, was slightly modified. 

In the annexed figure, T is the tube to be tested, /^^and F' 

are two circular brass rings 
nr LH P protrudins; from the tube 



R' Ji >■ at a distance of 1 cm. from 

Q> the ends, and soldered to a 

brass rod passing through 
the axis of the tube. The magnetizing coil was wound round 
the tube parallel to its length extending from F to F' to envelope 
it completely, and so arranged that the tube could slide in the 
coil with little friction. F in the lower part of the figure shows 



1) Beatson, Archives des. Sc. pliys. et nat. 2, 113, 1S46. 

2) Righi, Mem. di Bologna 4, 1, 1879; Beibl. 4, 802. 

3) Bidwell, Proc. Roy. Soc. .'il, 495, 1892 ; Beibl. 17, 582. 

4) Xagaoka, Phil. Mag. 2T, 131, 1894 ; Wied. Ann. r,s, 487, 1894. 



CHANGE OF DIME^•SIOi\S BY MAGXETIZATIOX. 79 

the front view of these rings. R and R' were two rods in con- 
tact with the ends of the tube. The ends of these rods were ben.t 
upwards and so filed down, that they could easily slide between 
two parallel wires of the coil, which were specially fixed at a dis- 
tance of 1 mm. from each other. The rod ?• served to communicate 
the motion to the prism P. The other parts of the apparatus 
remained unchanged. The apparatus was put into a magnetizing 
coil, 30 cm. long and wound in 12 layers with copper wire of 
2 mm. diameter. The field at the centre of the coil due to a 
current of one ampere was 37.97 C.G.S. units. The current 
through the outer coil produced the change of length by longi- 
tudinal magnetization and that through the inner coil gave rise 
to the change of length by circular magnetization. 

To study the efi'ect of temperature on the change of length, 
the circular magnetizing coil was wound, not by a single wire, 
but by double Avires ; thus connecting the four ends of these 

wires to a reversing key as shown in 
the figure, the circular field can be 
made or annulled by turning the key 
one way or the other. The total 
number of turns of the circular mag- 
netizins; coil was 44 for the nickel 
tube, 40 for the wolfram steel tube and 36 for the soft iron tube. 
The magnetizing currents were measured by Thomson graded 
galvanometers which were compared with a decianipere balance 
before each experiment. 

3. The samples used in the present experiment had the 
following dimensions : 




80 



K. HONDA. 



material 


length, 
(cm.) 


external 
diam. (cm.) 


internal 
diam. (cm.) 


demagnetizing 
factor. 


nickel 


17.02 


1.328 


1.2.52 


0.0261 


wolfram steel 


20.30 


1.124 


1.048 


0.0162 


soft iron 


16.97 


0.966 


0.842 


0.0308 



The tubes of nickel and soft iron are the same as that used 
in the study of the mutual influence between longitudinal and 
circular magnetizations. It was found by analysis that the 
nickel was nearly cliemically pure, the trace of impurity being 
inmeasurably small. 

Results of Experiments. 
1. Nickel Tube. 



4. The tube was carefully annealed, before the circular 
mao-netizins; coil was wound round it. The chanire of length due 
to lono'itudinal field alone was then measured in the usual way. 
The results w'ere compared with that obtained after the circular 
magnetizing coil was w'ound round the tube. The comparison 
showed that there was in general small difference between these 
two, and that the change of length in the former case was 
always greater than that in the latter, the ditïerence amounting 
to nearly 2 or 3 Yo. This is evidently due to the resistance to 
contraction experienced by the tube, although it can easily 
slide along the coil. Whether the apparatus executed its func- 
tion correctly or not was tested before each experiment by 
making a longitudinal field and comparing the deflection so 
obtained with that in the free state ; otlierwise serious mistakes 
would sometimes have arisen. 



f 



CHANGE OF DIMENSIONS BY MAGNETIZATION, 



81 



5. The experiments on the change of length by circular 
magnetization, namely, on the change of dimension in a direction 
perpendicular to the magnetic field, were conducted in the 
following manner. The tube was first demagnetized and a 
circular magnetizing current was made only for a moment and 
the corresponding deflection read. The change of length due 
to magnetization followed almost instantaneously, but the change 
due to the heating of the coil became sensible somewhat later; 
hence these two effects were unmistakably distinguishable so 
long as the magnetizing current was not strong. On this 
account the highc.-t field did not exceed 100 C.G.S. units. 

The eßfect of the longitudinal field on the change of length 
by circular magnetization was also measured. A constant longi- 
tudinal field was first made and the corresponding deflection 
observed ; then currents of different strength were momentarily 
passed through the circular magnetizing coil, and the additional 
deflection was read. These results are given in the following 
table and also in Fig. 1 : 

TABLE 1. 



H 


= 


H = 


= 6.9 


11 = 


-22.1 


11= 


182.9 


h 


4-xlO" 


h 


^xlΠ


h 


^xlO^ 


h 


^xlO^ 


7.6 


4.4 


8.2 


2.7 


8.2 


0.5 


8.5 


1.6 


16.1 


19.0 


17.5 


40.8 


17.5 


18.0 


13.6 


2.2 


26.3 


46.2 


31.1 


87.1 


31.2 


76.2 


21.3 


4.4 


40.6 


75.1 


40.9 


110.4 


40.9 


124.0 


31.5 


9.2 


50.5 


92.5 


49.9 


130.6 


50.2 


157.8 


49.0 


27.2 


64.5 


108.8 


57.5 


141.4 


64.2 


190.4 


64.2 


59.9 


72.1 


117.0 


72.1 


163.2 


70.9 


201.3 


69.5 


81.6 


83.2 


119.7 


81.7 


168.6 


78.8 


217.7 


87.6 


130.6 



82 K. HONDA. 

Here H and li denote effective longitudinal and circular 
fields respective!}', both in C.G.S. units. A represents addition- 
al change of length by circular field. All these changes were 
measured at a constant temperature of about 25° C. 

Fig. 1 shows that the change of length by circular mag- 
netization increases at first showly and tiien rapidly. With the 
further increase of the circular field, the rate of increase becomes 
gradually less. This result agrees in quality with Knott's 
calculation. The circular magnetization combined with a constant 
longitudinal one is always to increase the length which is first 
shortened by the longitudinal magnetization. In weak circular 
fields, the curve of the change of length with a constant 
longitudinal field lies below the cui've with no such field ; 
but in strong fields, the first curve lies above the second. The 
point of intersection of these two curves is displaced into a higher 
field with the increase of the longitudinal. 

6. We shall next pass on to the change of length by longi- 
tudinal magnetization with a constant circular field. The tube 
was first demagnetized by reversals, and then the deflections for 
longitudinal magnetizing currents of different strength were 
measured. During the experiment, the temperature at the centre 
of the magnetizing coil was 18.8° C. The tube was then care- 
fully demagnetized both as regards the longitudinal and cir- 
cular magnetizations. Then a constant current was passed through 
the circularly magnetizing coil so that the field strength became 
null. Owing to the heating of the coil, the tube rapidly ex- 
panded at first, but usually after an hour or two, it reached a 
stationary state ; when that state was reached, the measurement of 
the change of length by longitudinal field alone was commenced, 
which gave the length change at a higher temperature. After 



CHANGE OF DIMENSIONS BY MAGNETIZATION. 83 

the observation was finished, the key was reversed, so that the 
circular magnetizing current was then called into play. During 
this process, no gradual displacement was observed, showing that 
the temperature of the tube remained unclianoed durinsr the re- 
versai, but at the same time an instantaneous deflection was 
noticed, which showed the change of length by circular mag- 
netization. By reading the displaced position of the line in the 
micrometer ocular, tlie deflection corresponding to the longitudinal 
magnetization was noted. The tube was then demagnetized as 
regards the longitudinal magnetization, the circular magnetization 
remaining constant. The same process was repeated for stronger 
fields, till a set of observations w^as completed. 

7. How the rise of temperature afiects the change of length 
by magnetization will be seen from Fig. 2. The change of 
length at ordinary temperature is somewhat less than that w^hich 
Prof. Xagaoka and myself^' have obtained for an ovoid made of 
the same specimen. The diiference may perhaps be explained 
by that of annealing and of the geometrical shape of these 
samples. The temj^erature was measured by inserting a mercury 
thermometer inside the tube. Its effect is thus tolerably large ; 
the rise of temperature is attended with an increase of the 
change of length in weak fields, and is accompanied with a 
decrease in strong fields. From the same figure, w^e obtain the 
relation of temj^erature to the change of length at a constant 
field as shown in Fig. 3. It is well know^n that the magnetization 
of nickel increases with temperature in low fields and decreases 
in strong ones ; but under the temperature of 100° C, the change 
of masjnetization is too small to account for the chan^-e of lenofth. 

1) Nagaoka and Honda, Preceding paper. 



84 



K. HONDA. 



So far as I am aware, Barrett^' is tlie only physicist who has 
investigated tlie effect of temperature on the change of length ; 
his experiment resulted in the decrease of about one-fourth of 
the change of length by a rise of temperature by about 50° C. 
Perhaps his field was too strong to cause an increase. 

8. The results of the change of length by longitudinal 
magnetization with a constant circular field are given in the 
following table and in Fig. 4. The change of length was re- 
duced to the temperature of 18.8° C by using the results above 
obtained. 

TABLE II. 



ll 


=0 


h = 


= 10.7 


h = 


= 16.8 


h = 


=22.9 


H 


-^xlO^ 


H 


i^xlO" 


H 


^- X 10- 


H 


-'f X 10' 


5.3 


- 1.5 


6.9 


- 10.9 


8.1 


- 17.6 


7.2 


- 13.0 


SS) 


- 25.6 


14.6 


- 61.7 


17.8 


- 73.5 


14.2 


- 49.4 


17.0 


- 71.7 


27.8 


-118.1 


27.4 


-118.1 


22.3 


- 93.9 


29.4 


-107.6 


42.7 


-161.6 


46.7 


-174.9 


37.6 


-149.2 


41.2 


-134.7 


62.8 


-198.6 


71.9 


-22.3.4 


65.1 


-198.7 


6J.0 


-163.9 


103.3 


-238.4 






94.0 


-232.7 


101.3 


-202.4 


131.6 


-2.5.5.2 


131.9 


-271.0 


129.7 


-269.3 


184.9 


-241.8 


176.8 


-279.2 


177.7 


-291.5 


175.2 


-29.3.4 


274.9 


-261.3 


254.0 


-301.7 


2.55.4 


-313.7 


255.5 


-317.0 


354.9 


-274.1 


361.0 


-312.9 


363 4 


-329.7 


359.3 


-333.2 


468.6 


-279.2 


.505.0 


- 324.0 


516.1 


-339.4 


516.1 


-347.4 


709.2 


-289.5 


720.2 


-.331.6 


779.0 


-344.5 


725.2 


-3.51.4 



1} Barrett, Phil. Ma,^. [4] 47, 51, 1S74 ; Nature 2«, 515 58G, 1882; Beibl. 7, 201. 



CHANGE OF DIMENSIONS BY MAGNETIZATION. 85 

The comparison of Figs. 1 and 4 shows that the general 
character of the change of length is the same in these two 
cases, except that the sign of the change is opposite. Hence 
simihir remarks as in the former hohl good in the present 
case. 

In the experiment with nickel and cohalt wires traversed by 
an electric cnrrent, Bidwell found that the effect was inmeasur- 
ably small. The discrepancy in nickel perhaps arises from the 
effect of temperature, which he did not take into account ; the 
difference in the method adopted in the present experiment for 
obtaining a circular field and in that of Bidwell does not seem to 
play an important part in accounting for the said discrepancy. 
According to the present experiment, the rise of temperature 
occasioned a comparatively large diminution of the length change 
in strong fields. Hence it can not be denied that in Bidwell's ex- 
periment, the effect of circular magnetization was just as great as 
that of temperature. The same remark will perhaps apply to his 
experiment with cobalt ; but having no cobalt tube at my dis- 
posal, the experimental verification must be postponed till some 
future date. However, a theoretical deduction in favor of the 
view above stated will be given in the last part of the paper. 

It would not be out of place to remark that a klinging 
note of the nickel tube was heard at the make and break of 
circular magnetizing current, a well known phenomenon. Even 
with such a weak current as we obtain from a single Daniell's 
cell was sufficient to produce a distinctly audible sound. 

9. It will sometimes happen that it is convenient to have 
a simple expression for the change of length. For nickel, the 
change of length is very well given by an empirical formula 
of the form 



8G 



K. HONDA. 



I 



air 



l + ßH" ' 

where «, ^9 and n are constants and H is assumed to be positive. 
The determination of these constants from the experimental 
curve gave the following results : 

« = 5.18, ^9=0.0164 and n = 1.017. 

In the calculation, only the fields H = 20, 80, 320 were 
chosen to simplify the calculation. Using these values of the 
constants, the change of length due to fields of different strength 
was calculated and compared with the experimental value as 
shown in the followinir table : 



TABLE III. 



H 


^ (cal.) 


-r (^^^P-) 


10 


- 46x10-' 


- 40x10-' 


20 


- 81 


- 81 


30 


-108 


- 109 


50 


-148 


-148 


80 


-185 


-185 


120 


-215 


-214 


150 


-230 


-229 


200 


-247 


-246 


250 


-258 


-258 


320 


-269 


-269 


400 


-278 


-278 


500 


-285 


-284 


GOO 


— 292 


-289 


700 


-293 


-291 



CHANGt; OF DIMENSIONS BY MAGNETIZATION. 



87 



Thus except in weak fields, the coincidence between these 
two is very close; the difference does not amount to 1 °/o. 
This formula applies, not only for the change of length, hut 
also for every curve which has only one inflexion point and 
becomes asymptotic when one of the co-ordinates increases 
indefinitelv, such as the curve of magnetization. 

2. WoLFKA3i Steel Tube. 

10. The method of procedure with the steel tube was exact- 
ly the same as in the corresponding case of the nickel tube. 
The result of the change of length bv circular magnetization, 
e.g., the dilatation in a direction perpendicular to the field, as 
well as the effect of longitudinal field on the change of length 
by circular magnetization are given in the following table and 
graphically shown in Fig. 5. These observations were taken 
at a constant temperature of about 17^ C. 

TABLE IV. 



H=0 


H = 


= 15.1 


H = 


=31.8 


H = 


=81.3 


h iL X lü" 


h 


1 ^^^' 


h 


'; X 10^ 


h 


4^x10^ 


14.0 - 0.0 


13.2 


- 0.4 


13.6 


- 0.4 


12.9 


- 0.4 


20.S - 8.6 


31.7 


-12.9 


30.9 


- 2.1 


31.2 


- 0.9 


35.2 -20.2 


41.9 


-32.2 


41.9 


-14.6 


41.6 


- 5.2 


51.8 -22.8 


51.7 


-40.4 


51.1 


-25.8 


51.1 


-12.9 


65.1 -26 6 


63.7 


-4.).1 


63.7 


-38.7 


63.7 


-25.8 


78.7 -27.9 


74.7 


- 47.3 


74.7 


-48.1 


75.6 


-30.9 


99.5 -27.9 


88.6 


-49.4 


91.4 


-58.0 


88.6 


-40.0 



88 K. HONDA. 

Here we observe that circular magnetization produces con- 
traction which increases very slowly at first, but afterwards 
quite rapidly, till it reaches a nearly constant value. The 
existence of the field of maxioium contraction is still a question. 
The result is somewhat discordant as compared with that of Bid- 
well with an iron ring, in which case the diminution vanishes 
in a field of about 86 C. G. S. units. Since the behaviour of wol- 
fram steel as regards the change of dimensions by magnetization is 
very different from that of soft iron, the cause of the discrepancy 
is probably to be sought for in the difference of the specimens. 

That the effect of longitudinal field on the change of length 
by circular magnetization is of the same nature as in the case 
of nickel, except that the sign of the change is opposite, is also 
apparent from the same figure. As we have remarked, Beatson 
and Riglii observed the same phenomenon. 

11. The middle curve in Fig. 6 represents the change of 
length by longitudinal magnetization at the temperature of the 
room. The lower curve was obtained at 80.2° C, and the upper 
curve at the same temperature by reversing the key so as to 
produce circular field. From the figure, we see that the be- 
haviour of wolfram steel as re2;ards the clian2;e of leuirth is 
widely different from that of other sorts of iron. It is remark- 
al)le that the length of the tube, after reaching the maximum 
elongation, diminishes very slowly as the field is increased, a 
fact already noticed by the experiment^' referred to. In that 
case, the maximum elongation was somewhat less than in the 
present experiment. The discordance between the two is pro- 
bably due to the difference of annealing and also ^of the shape 
of the specimens. 



1) Nagiioka and lloiiila, loc. cit. 



CHANGE OF DIMENSIONS BY MAGNETIZATION. 



89 



The effect of temperature is to decrease the change of length ; 
the diminution increases with the fiehl, till it reaches a maxi- 
mum, and then decreases very slowly. Barrett'' did not find the 
effect in the case of iron and cobalt. The upper curve shows 
that the influence of circular magnetization on the change of 
length is large for steel. 

12. The effect of circular field on the change of length by 
lonsjitudinal mas-netization is shown in the followino; table and 
in Fig. 7. The results are reduced to the temperature of 
17.2° C. 

TABLE V. 



h. 


= 


h = 


10.8 


li = 


17.7 


li = 


25.8 


H - 


^xlO^ 


H 


i^xlO^ 


H 


J/-xiœ 


H 


i/-xlΠ


12.0 


0.0 


19.3 


3.7 


14.5 


0.2 


11.3 


0.0 


17.7 


8.3 


30.5 


18.3 


26.1 


]1.9 


21.0 


2.1 


29.3 


16.2 


37.6 


24.5 


31.9 


18.0 


31.2 


17.2 


49.4 


27.3 


53.1 


3.5.8 


50.8 


37.2 


57.7 


41.8 


93.0 


38.8 


84.6 


46.3 


75.7 


50.5 


83.7 


56.8 


125.1 


42.9 


13.5.3 


52.4 


105.5 


54.5 


120.3 


67.6 


170.0 


44.9 


182.5 


54.2 


168.4 


60.1 


165.8 


72.5 


348.5 


43.3 


233.3 


55.0 


244.5 


63.7 


246.5 


75.2 


441.0 


42.9 


325.5 


55.6 


350.0 


63.9 


351.5 


73.3 


545.0 


42.5 


505.0 


54.2 


461.5 


63.6 


459.5 


72.3 


728.0 


41.3 


708.8 


52.4 


615.5 


62.8 


671.6 


71.0 



Thus the longitudinal magnetization combined with a constant 
circular one is always to increase the length which is first 



1) loc. cit. 



öo 



K. HONDA. 



shortened by tlie circular magnetization. In weak longitudinal 
fields, tlie curve of the change of length with a constant circular 
field lies slightly below the curve with no circular field ; but 
in strong fields, the first curve lies markedly above the second. 
The point of intersection of these two curves shifts into a high 
field as the circular field is increased. The field of the maxi- 
mum elongation seems to increase with the circular field. 



3. Soft Iron Tube. 

13. The experiments of the change of length by circular 
magnetization and of the effect of longitudinal field on the 
change of length led to the following results, which are graphi- 
cally shown in Fig. 8. The observations were taken at the 
temperature of 18° C. 



TABLE VI. 



H 


= 


H: 


= 5.7 


H = 


=25.8 


H = 


= 67.6 


Ix 


-'l~ X 10^ 


h 




h 


4x10^ 


h 


4^x10^ 


5.3 


- 7.8 


5.3 


- 4.2 


5.3 


- 0.0 


5.3 


- 0.5 


14.0 


-13.0 


13.8 


-11.9 


14.0 


- 5.2 


13.8 


- 1.0 


21.4 


-15.6 


20.7 


-16.6 


21.4 


-10.4 


21.0 


- 4.2 


37.5 


-15.6 


35.7 


-20.8 


37.3 


-20.8 


37.3 


- 9.9 


53.2 


-14.5 


51.8 


-22.3 


53.3 


-26.0 


53.2 


-14.0 


69.4 


-12.5 


66.6 


-22.3 


69.2 


-28.0 


67.8 


-18.2 


81.6 


- 9.3 


81.3 


-20.8 


81.6 


-26.0 


80.5 


-20.8 


98.8 


- 7.8 


97.7 


-19.7 


98.0 


-23.4 


98.1 


-20.8 



By circular magnetization, the length of the tube diminishes 
rapidly at first, till it reaches a minimum, then it gradually 



CHANGE OF DIMENSIONS BY MAGNETIZATION. 91 

recovers. The field at which the tube returns to its former 
length is not yet reached so far as the present experiment 
extends. The result agrees qualitatively with that of Bidwell and 
the calculation of Knott. 

The general form of the curve does not change by the 
application of a constant longitudinal field, but the field of maxi- 
mum contraction shifts into high field as the longitudinal field 
increases. The amount of the maximum contraction increases 
with the longitudinal field, till it reaches a maximum, and then 
it gradually decreases. In weak circular fields, the change of 
length diminishes with the increase of the longitudinal. 

14. As in the case of wolfram steel, three curves in dotted 
lines are given in Fig. 9, two of which correspond to the change 
of length at the temperatures of 18.7°C and 76.1° respectively. 
When the key in the circuit of the circularly magnetizing coil 
was reversed so as to produce a field, the change of length 
corresponding to the third curve was obtained. 

The change of length by longitudinal magnetization at ordi- 
nary temperature is somewhat less than those obtained by previ- 
ous experimenters. The difference is probably to be ascribed to 
the well annealed state^' of the tube ; also, the resistance to the 
elongation experienced by the tube due to the friction of the 
circular magnetizing coil was found to affect the result slightly. 
The oeneral feature of the clianoe of leno;th is so well known 
that farther remarks are unnecessary. It is only to be noticed 
that here the field of the maximum elongation is greater by 20 
C. G. S. units than that of tlie minimum contraction due to 
circular magnetization. 

The rise of temperature is to diminish the change of length 

1) Bidwell. Phil. Mag. 55, 228, 1894. 



92 



K. HONDA. 



in weak fields and to increase it in strong ones. The field at 
which the temperature produces no effect is about 52 C. G. S. units. 
In the case of wolfram steel, this field, if it exists, seems to be 
pushed to an intensely strong field. We also observe that the 
effect of circular field on the change of length by longitudinal 
magnetization is tolerably large, as observed by Bidwell. 

15. The results of the experiments on the change of length 
by longitudinal magnetization with a constant circular field are 
summed up in the following table and graphically shown in 
Fig. 10, these results being reduced to 18.7° C. 



TABLE Vll. 



h 


=0 


h = 


= 5.7 


h = 


.9.2 


h = 


= 26.2 


H 


-^xlO" 


H 


4 X 10^ 


H 


^'L X 10' 


H 


-^L X 10^ 


5.3 


1.1 


6.9 


4.7 


5.3 


1.1 


5.3 


1.7 


10.3 


12.8 


— 


— 


11.2 


17.2 


10.3 


9.6 


21.5 


17.1 


17.9 


22.4 


22.9 


29.6 


20.5 


22 7 


41.3 


19.2 


37.8 


28.8 


39.4 


37.4 


40.6 


35.6 


70.3 


19.2 


61.8 


32.1 


64.0 


42.3 


61.2 


40.0 


97.9 


17.1 


111.1 


31.0 


97.9 


42.2 


90.9 


4].l 


144.3 


11.1 


142.6 


27.8 


145.0 


38.8 


143.5 


38.2 


223.0 


3.G 


218.0 


21.4 


217.6 


31.0 


217.2 


34.5 


318.5 


- 4.9 


320.3 


12.4 


320.0 


20.2 


312.5 


25.3 


481.4 


-20.3 


490.0 


- 4.3 


493.8 


4.2 


475.0 


8.6 


697.0 


-32.5 


704.0 


-13.5 


684.2 


- 5.7 


647.0 


- 2.6 



Thus the nature of the change of length is the same as in 
the reciprocal case already mentioned, except that the sign of 
the change is opposite. As shown in the figure, in strong fields, 
the curve corresponding to the change of length with a constant 
circular field lies alwavs above that witli no circular field. 



CHANGE OF DIMENSIONS BY MAGNETIZATION. 



93 



In weak fields, the curves nearly coincide with each other. The 
field of maximum elongation slightly increases with the circular, 
and the amount of the elongation, after reaching a maximum, begins 
to decrease with further increase of the circular field. Tliough 
Bid well did not observe this point, the present experiment agrees 
quite well with his result. 

16. So far the experiments made on the tubes of nickel, 
steel and iron show that the effect of circular field on the change 
of length by longitudinal magnetization is of the same nature 
as the effect of longitudinal field on the change of length by 
circular masrnetization. 

From the results of the change of length by longitudinal 
and circular magnetizations, the change of volume by magne- 
tization can easily be calculated, provided we assume the material 
to be isotropic, as was already done by Bidwell. If ti and v 
represent these two dilatations respectively, the volume change a 
is given by the formula (t=u +2i\ 

Assuming the isotropy of our specimens, we find the 
calculation leads to the followins; results : 

TABLE VIII. 



H 


Nickel 


Wolfram steel 


Soft iron 


ÔV 


ov 


ov 




V 


V 


V 


10 


- 18.5 X l5' 


0.0x10' 


-J.GxlO 


•20 


-21.0 


- 7.2 


-13.1 


30 


0.0 


-18.8 


-12.8 


40 


18.0 


— 22.2 


-12.2 


60 


4:6.5 


-21.8 


- 7.7 


80 


04.0 


-19.2 


- 1.9 


100 




-16.4 


1.5 



94 K. HONDA. 

We tluis obtain incredibly large values for the change of volume. 
In nickel and soft iron, there is at first decrease of volume, 
and then follows an increase; in wolfram steel, the diminution 
of volume reaches a maximum and then gradually decreases. 
The above result for soft iron agrees feirly well with that of 
BidwelF* for un annealed iron ring. But in the experiment with 
ovoids"^ made of the same specimens, there was always small in- 
crease of volume for nickel, steel and soft iron. The amount of the 
change at the field of 100 C. G. S. units was 0.7 x lO"", 3.1 x 10"^ 
and 2.8 X 10"' for these metals respectively. Hence the question 
now arises whether the change of volume is so influenced by the 
shape of these metals. To settle this point, fresh experiments on 
the change of volume were undertaken with a dilatometer. The 
answer was in the negative, the result being in rough agree- 
ment with that for the ovoids. The initial decrease of volume 
was never observed, but the volume always increased with the 
increase of the magnetizing field. The discrepancy between the 
calculated and the experimental result is perhaps due to the 
œolotropy of the materials. For, if it were not isotropic, the 
lateral dilatation by longitudinal magnetization would not coin- 
cide with the change of length by circular magnetization. It will 
also be explained by the a3olotropy of the specimens that in weak 
fields, Bidwell's calculation resulted in the large diminution of 
volume of iron rings in contradiction to the experimentally 
established fact. 



1) loc. cit. 

2) Nagaoka and Honda, loc. cit. 



CHANGE OF DIMENSIONS BY -MAGNETIZATION. Do 

Concluding Remarks. 

17. From the experiment on the relation between magnetism 
and twist, Knott" concluded that the pure strain effects on a 
ferromagnetic ^vire caused by tension and longitudinal current 
through it are of an opposite character, and also, on the ground 
of Maxwell's explanation for Wiedemann's effect, that in an 
iron or nickel wire carrying an electric current, the change of 
length by magnetization must be greater than when there is no 
longitudinal current. Since the change of length for cobalt is 
scarcely affected by tension, the same must also be the case for 
longitudinal current. The consideration is partially verified by 
the experiment of Bidwell and also by the present one. 

The same phenomenon may also be more concisely explained 
in the following manner. Suppose our samples to be isotropic 
and to have no residual effect. Let / and ^36 two magnetic 
forces acting longitudinally and circularly in two perpendicular 
directions. When these two forces act simultaneously, we have 
a resultant force H\ this force occasions the change of dimen- 
sions in our ferromao-netics. The dilatation in the direction of 
the resultant force, as well as that in the direction perpendicular 
to it, can be expressed by f{H) and F{H) respectively, which 
are even functions of H. To obtain the dihitation in the longi- 
tudinal direction, we have simply to cpnstruct a strain ellipsoid 
at any point of the ferromagnetics and to find the change of 
length of the radius vector in this direction. The simple cal- 
culation gives 

^=/(//)^ + /-(//)Jr. 

1) Knott, Trans. Ftoy. Soc. Edinb. 36, pt. IT., 485. 



96 K. HONDA. 

lu the case of nickel, the change of volume is negligibly small 
compared with that of length ; hence we may put with tolerable 
accuracy f(H) + 2F{H)=^0. With steel and soft iron, the 
change of volume is not very small comj)ared with the change of 
length. But if t does not exceed 50 C. G. S. units, the effect of 
volume change on the change of length by combined action of 
I and t is negligibly small, for in these strong fields at which the 
change of volume is pronounced, the ratio f/H'- in the above 
expression becomes very small. Hence even for these metals, 
we may neglect the change of volume, provided the circular 

> 7- 

field is not very large, and the expression for --^ becomes, 
in all cases, 



L IP 



Since the material is supposed to be isotropic, /{H) is the same 
as the ordinary change of length by longitudinal magnetization. 
Thus the change of length by longitudinal magnetization with 
a constant circular field can be calculated from the change of 
length by longitudinal magnetization aloue. The same expres- 
sion can also be used for the calculation of the change of 
length due to circular magnetization with a constant longitu- 
dinal field. 

In order to compare the above result with that of the 
expeiiment, it is obviously necessary to subtract from -^^ the ex- 
pression F{i) for the change of length by longitudinal magne- 
tization with a constant circular field t, aud/(7) for the reciprocal 
case. 

Assuming for the expression /(//) a suitable empirical for- 
mula for iron, steel or nickel, a simple analytical discussion of 



CHANGE OF DIMENSIONS BY MAGNETIZATION. 97 

tlie expression — j^, or numerical calculation of it for different 
values of / and t from the experimental curve of the ordinary 
change of length leads to the conclusion that for iron, steel 
and nickel, all the points, which we have remarked in connection 
with the curves shown in Figs. 1, 4, 5, 7, 8 and 10, are involved 
without exception in the expression of — ^/— • 

It may be observed that the behaviour of cobalt with regard 
to the change of length is just the reverse of that of iron, and 
therefore every result which we have obtained for iron is also 
applicable to the case of cobalt, provided the sign of the length 
change be properly reversed. Thus in strong fields, the length 
of a cobalt tube should, by the combined eftect of longitudinal 
and circular magnetizations, become shorter than when acted 
upon by the former alone. In weak fields, the result should be 
just the opposite. The field of maximum contraction should 
increase with the circular, and the amount of the contraction, after 
reaching a maximum, gradually decrease. The circular field at 
which the maximum contraction occurs should be far greater 
than that for iron. 

19. The comparison above made is qualitative ; how the 
calculated and the experimental numbers agree with each other 
is seen from the following table : 



98 



K. HONDA. 



TABLE IX. 





Nickel Tube, t=10.7 




I 


L' (cal.) 


U (exp.) 


difiference 


10 


- 20x10' 


- 25xl0' 


5 X f Ö 


2 J 


- 74 


- 85 


11 


30 


-112 


-125 


13 


50 


-162 


-175 


13 


SO 


-204 


-216 


12 


120 


-237 


-250 


13 


200 


-270 


-285 


15 


300 


-291 


-305 


14 


500 


-309 


-323 


14 


700 


-318 1 -331 


13 



Here 1! denotes 



L 



■F{1), and its vakie was calculated from the 



experimental curve for the ordinary change of length. A glance 
in the above table shows a fair agreement between the calculated 
and the experimental values. The difference between these num- 
bers is not of a serious nature, if we remember that one scale 
division of the micrometer ocular corresponds to the change of 
5.12x10"^ for nickel, and that the correction for temperature 
amounts to 11 x 10"' in the most siguilicant case. 

The discrepancy is probably due to the residual effect and 
also to the œolotropy of the tube. If the tube, after it is 
magnetized both longitudinally and circularly, is demagnetized 
by reversals with regard to the longitudinal magnetization, 
the circular field remaining constant, as was actually the case 
in the present experiment, the elongation due to the circular 
field alone is usually increased by one or two scale divisions, 



CHANGE OF DIMENSIONS BY MAGNETIZATION. 



99 



a phenomenon which is perhaps to be attributed to the residual 
effect noticed in ray former paper^'. The constancy of the 
difference in the above table furnishes additional evidence in 
support of this view. The œolotropy of the tube as regards the 
change of length evidently influences the experimental values. 
Moreover tlie change of the intensity of longitudinal mag- 
netization due to the mutual interaction of longitudinal and 
circular fields is not taken into account in the calculation of the 
effective field. These causes, I believe, are sufficient to account 
for the said discrepancy. 

In steel and soft iron, there are comparatively large diffe- 
rences between the calculated and the experimental numbers, as 
will be seen from the following table : 

TABLE X. 



I 


Wolfram steel, t = 17.7 


Soft iron, 


t=26.2 


L' (cal.) 


TJ (exp.) 


// (cal.) 


L' (exp.) 


10 


2x.l0' 


1x10' 


9xl0' 


8xl5' 


30 


22 


18 


23 


30 


50 


31 


37 


29 


38 


80 


39 


51 


31 


41 


120 


46 


57 


29 


41 


200 


48 


Q2 


21 


36 


300 


48 


64 


12 


26 


.500 


47 


63 


— 5 


7 


700 


45 


62 


-17 


- 5 



For iron and steel, the sensibility of the apparatus was about 
2xlO~' and the correction for temperature amounted to 5xl0~^ 



1) K. Hoiidu, Jour. Sc. Coll. XF., 311, 1899. 



100 K. HONDA. 

in the most significant case. I believe that the principal causes 
of discrepancy above ennmerated are sufficient to account for 
the difference between the calculated and the experimental 
numbers. 

19. Thus qualitatively the above result and the experiment 
are in complete agreement with each other, although there are some 
discrepancies in quantitative details ; there are, however, probable 
causes to account for the discrepancies. According to Knott, the 
change of length in cobalt by longitudinal magnetization is very 
little affected by the presence of a circular field, but the above 
consideration leads to a result which contradicts his anticipation. 
Hence a single experiment on this point for cobalt will deci- 
dedly establish the correctness of the one explanation against 
that of the other. 

In conclusion, I wish to express my best thanks to Prof. 
H. Nagaoka, and also to Prof. A. Tanakadate for useful advice 
and kind guidance. 



Jour Sc Coll. Vol XIII PI. VIII 



^Si;^;^il;pH|p:!;;!piPilMN!!N^;i^Hüi^;Eli;H;IH:^^ 


^^S^pj£}jfeff^-HKM 


1 r N M^'g •'■t-H 1 ^-H ni-^'an-ii: ' q: 


'"'""' li 1 A'i;rA.'«;/ lube .. | ^ | - .''' ^ _-) -j- 1- T T Z _ 

, 1 \— jJ- _ — I--X .-^ ==>?--? -'- 




-^ — 1— •+ -^ -- 7^ - -;:- = s:2' "" " ^" i^ "" ~ -1 -«««Î iuij, .-. _ :. ~ z ^ ---?-»;; ------- 


»=dTtiii±^-^i:i^:-i^3+$++;2-:::: + ::::::::-+:jî°::¥H 










■ • : 1 ' ! ' J^ y^ .-' : _ _ . _ :^ __ _ _.. _ . _ ,?. <!__ _ _ _ _ 


1 ' 1 Ti Ml -n L-rl ^L M J'^ U 












' i\^^ *\^^ ---1- - -t-' -r ■'3^^. ~"T"^— — — ^— ^ l [.-- t.. ^ L 


■ TSUK H fHJJL, ■ (- - -1- - ^i- ' 1 1 1 1 i 1 1 1 1 II 1 1 1 1 1 1 II 1 1 IR^^^^^=s5ès=^=:fc^H^ 1 ■ - - - - --- . 1 1 1 1 1 1 ][■■ Hü 


;l Pl-fl^ pfl f\|] woijhmisteu .. |. |.]|.|...|| J ' ■ :.rpr~ ^^•T^~:^;===d::^^^___; , , ^rt-b^f:;;!"^ 


, -'- "'-^" >■ '^ "^•»^ "^ ~ ï'o-o'" Il ■■■■'' ' ■ ■ ' ' '1 ■ ff-rai«^ 












" — " "'~5 vi "- _!:!:,::": " ■ -_^_.___-^it ___ . _ . _ :;:;_: 




-- -- -- - + ^5 ^^8^^- ''^j: " " i- zL: - ± - : - .- _ .:._,._ :-.:-: : :: ". :::": 




„,;." _ 2^ ^ X-- - - - -0- - iS- - 2S-- as -i £12 -_is -ä < îî g 
















___ _- — - _^-_ _^^- -^ - ^_^ -J- ^ JJ ^ - - J,,^^ ^ . - ^ ^ _ - ^ - 










j^^^ FiffS 1, ,1^--- 


_. ^ »^ Mllll 1 1 - - -- -H — 


_'V ^ -iv -- J'II'' -- ^ 4 J- - _ -_ _ - 


--lS-- ït-- s^ -± Soft tron ._ ::. ^ \ i ------- - ------- 


1^*. ^ OW - --- X-H _„-- -_ - -_ --!_-- 


^t ^vî. ,-r ^^ \ 










i-^s ï> - .:s _.^,-- _ - - ._ 's^ii ^--^ - - - - - . - 


-i^ \ - ~^__,;: - _;::;:::: ;_ . -_:: !s^k^s '~ 


-?-, = ^j . = .-'55; + ::-s^o?;:::e;;:: :::_::::::::::::::::::::::::::::::::: :::::::::;:::::::-: 


:::::::::- :^v:-:::::::-ï;::::::::^::::::::^:::::::::::::::::::::^^ 


,^----±---::: = ;;:5-:::::+:::::::: = s;:::::::::::::::::::::::::::;::-::::j::^.;-;;::^ 






-j- _ J L_.J I^:-s===c,^__g^,^_,_^^r=rtr^_-_., _ -__^_^i^j,B-, 


-- -_..-- --,.: 1 X 5; X--^-M- 1--J r^_p = :. = ,.: — -^__^^^^;„: ,, . ;-,^__/i-ka 








_ -- _ __ _~~ ___-_- ^ - _____ ____ 1 ,.,- -j 





1( 

iE 

Ol 

tl 

n 

ai 
di 

C£ 
Cl 
li 
C( 

H 
d. 
tl 



B 
ai 



Jour. Sc. Coll. Vol. XIII PI. IX. 




Studien über die Anpassungsfähigkeit einiger 
Infusorien an concentrirte Lösungen^). 



Von 



Atsushi Yasuda, Rigakushi, 

Professor der Naturirescliiclite an der zweiten Hoclischule zu Sendai. 



Hierzu Tafel X-XII. 



Einleitung. 



In der Natur finden wir Tliiere und Pflanzen stets den 
obwaltenden Bedingungen angepasst. Diese Anpassung an die 
Umgebung ist in der That erst im Verlaufe langer Generationen 
hervorgebracht worden. Wenn dann aber die Lebensbedin- 
gungen sich ändern, wie werden diese Organismen dadurch beein- 
flusst? Unsere bisherigen Erfahrungen lehren uns, dass den 
Organismen im Allgemeinen die Fähigkeit innewohnt, sich diesen 
veränderten Verhältnissen genau anzupassen und so dauernd 
leben zu können, nur unter der Bedingung, dass die Veränderung 
nicht plötzlich stattfindet, oder, wenn sie rasch vor sich geht, 
sie doch verhältnissmässig unbedeutend ist. 



1) Die vorläufige Mittheilung dieser Arbeit erschien in The Botanical Magazine. 
Tokyo 1897. Vol. XI, iSTo. 121. pp. 19-24. und Aunotationes Zoologicte Japonenses. 1897. 
Vol. I, Part I et 11. pp. 23-29. 



102 A. YASUDA : ANPASSUNGSFAEHIGKEIT 

Bekanntlich giebt es in der Pflanzenwelt die "Wasserformen 
der amphibischen Gewächse, wie Polygonum amphibium und 
Ranunculus aquatilis, die sich in ihrer morphologischen und 
anatomischen Beschaffenheit ganz anders als ihre Landformen 
verhalten. Auch sind die Hydrophyten in Bezug auf Gestalt 
und Struktur einer grossen Metamorphose unterworfen, die sie 
zum Leben im Wasser befähigt. Aehnliche Thatsachen finden 
wir auch in der Thierwelt. Hierher gehören z. B. unter den 
Amphibien die Anuren, deren aus den Eiern ausschlüpfende 
Larven durch Kiemen athmen, aber im erwachsenen Zustande 
durch Lungen, während die Tliiere, welche zu den Perenni- 
branchiaten gehören, fortdauernd Kiemen besitzen, weil sie 
lebenslang im Wasser wohnen und niemals ein oberirdisches 
Leben führen, so dass sie sich jenem Medium völlig accommodirt 
haben. 

Es dürfte daher von Interesse sein, wenn wir die Beschaf- 
fenheit der in der Natur gefundenen Medien verändern, künst- 
liche Nährmedien anfertigen und die Anpassungsfähigkeit 
gewisser für diesen Zweck ausgewählter Organismen an diese 
künstlichen Medien studiren. Es liegen bereits Untersuchungen 
mancher Forscher über derartige Kulturversuche bei niederen 
Organismen vor. Im nächsten Abschnitt fasse ich die Resultate 
der wichtio'sten einschlä2;i2;en Arbeiten zusammen. 

Vorausschicken will ich noch, dass ich die Anpassung der In- 
fusorien an solche künstlichen Medien studirt habe, die aus hetero- 
genen Lösungen von höheren Concentrationen bestanden. Fol- 
gendes sind die Fragen, die ich zu beantworten versucht habe : — 
1) Welche Grade der Conc-entrationen der Aussenmedien können 
die Infusorien ertragen ? 2) Welche relative Widerstandsfähigkeit 
haben sie im Vergleich mit Algen und Pilzen ? 3) Welche 



EINIGER INFUSORIEN. 103 

Veränderungen ihrer äusseren und inneren Gestalt werden 
dadurch hervorgebracht, und wie wird ihr Bewegungs- und 
Vermehrungsvermögen durch diese heterogenen Medien beein- 
flusst ? In Folgendem werde ich der Reihe nach die diesbezüg- 
liche Litteratur, die Methodik meiner Versuche, die Beschreibung 
der Experimente und endlich die Zusammenfassung der Resultate 
mittheilen. 

Die vorliegende Arbeit wurde grossentheils im botanischen 
Institut der kaiserlichen Universität zu Tokyo unter freundlicher 
Anregung des Herrn Prof. Dr. M. Miyoshi ausgeführt, wel- 
chem ich für seine liebenswürdige Rathschläge meinen verbind- 
lichsten Dank ausspreche. Herrn Prof. Dr. J. Matsumura 
sage ich auch an dieser Stelle für das Interesse, welches er meiner 
Arbeit entgegengebracht hat, meinen besten Dank. 



Litteratur. 



Ueber die Accommodation niederer Pflanzen giebt es ziem- 
lich viele Versuche; so beobachtete StahP), dass Aethalium 
septicum sich allmählich an Traubenzuckerlösungen anpasste und 
der Einwirkung einer 29^igen Lösung widerstand. Richter") 
experimentirte mit Cyanophyceen und fand, dass Eivularia 3^o 
ige, Gloeocapsa 6^ ige, Anabaena 69^ ige und Oscillaria 10 ^o ige 
Kochsalzlösung ertragen konnten. Auch gelang es ihm, Dia- 
tomaceen in einer 7 9^ igen Kochsalzlösung ein Jahr und in einer 
10 9^ igen einen Monat lang leben zu lassen. Er zog ausserdem 

1) E. Stahl. Zur Biologie der Myxomyceten. Bot. Ztg. 1884. Nr. 11. p. 166. 
*) A. Richter. Ueber die Anpassung der Süsswasseralgen an Kochsalzlösungen. Flora. 
1892. pp. 18-56. 



104 A. YASUDA : ANPASSUNGSFAEHIGKEIT 

verschiedene andere Algen in den Bereich seiner Untersuchungen, 
darunter Zygnema, Mougeotia, Spirogyra, Cosmarium, Chlorella^ 
Tetraspora, Chaetophora, Vaucheria, Oedogonium,, Ohara u. s. w.; 
eine gewisse Anzahl von ihnen vermochte sogar in IS'^iger 
Lösung zu existiren. Ferner theilte Klebs^) mit, dass einige in 
den Lösungen organischer Verbindungen kultivirte Süsswasser- 
algen anfänglich eine Plasmolyse zeigten, die aber nach einigen 
Stunden vollständig ausgeglichen war, worauf sie ohne Schaden in 
den neuen Medien fortlebten. Nach demselben Autor gedieh Zyg- 
nema in einer 10 bis 209^igen Glycerinlösung eine Woche lang. 
Auch 10-50'?^oige Rohrzuckerlösungen vermochten dieselbe Alge 
im Leben halten, aber mit verschiedenen Wirkungen je nach 
der Concentration : lO^^ige Lösung veranlasste lebhafte Kern- 
theilung, 20-259^ ige Längenwachsthum, 309^ ige Zellhautbildung 
und 40^ ige Assimilation und Stärkebildung, während in 60% 
iger Lösung die Alge nur wenige Tage lebte. 

Unter den Meeresalgen nahm Jan se") eine ähnliche Er- 
scheinung bei Chaetomorpha wahr, und zwar hervorgebracht durch 
Kalisalpeter- und Kochsalzlösungen. Er fand nämlich, dass, 
wenn man diese Alge in eine solche Lösung legt, in Folge ihrer 
Anpassung an dieselbe nach kurzer Zeit ihre Widerstandsfähig- 
keit bedeutend gesteigert wird. Oltmanns") machte Experi- 
mente über den Einfluss der Concentrationsänderung des 
Meerwassers auf Fucus, der bei niedriger Concentration sich dem 
neuen Medium gänzlich accommodirte. Eschenhagen^) kulti- 

G. Klebs. Beiträge zur Physiologie der Pflanzenzelle. Berichte der deutsch, bot. 
Gesellsch. 1887. Bd. V, Heft 5. p. 181. 

') J. M. Janse. Plasmolytische Versuche an Algen. Bot. Centralbl. 1887. Bd. XXXIT. 
p. 21. 

^) F. Oltmanns. Ueber die Bedeutung der Concentrationsänderung des Meerwassers für 
das T^ben der Algen. Sitzb. d. Königl. preuss. Akad. d. Wissensch. zn Berlin. 1891. p. 193. 

*) F. Eschenhagen. Ueber den Einfluss von Losungen verschiedener Concentration 
auf das Wachsthnm von Schimmelpilzen. Stolp. 1889. 



EINIGER INFUSORIEN. 105 

virte Aspergillus niger, Pénicillium glaucum und Botrytis cinerea 
in verschiedentlich concentrirten Lösungen von Traubenzucker, 
Glycerin, Natronsalpeter, Kalisalpeter, Chlornatrium und Chlor- 
kalium, und wies sowohl die Grenzpunkte ihrer Accommoda- 
tion als auch ihr Wachsthumsverhältniss zu diesen Substraten 
nach. Bachmann^) bewies durch zahlreiche Experimente, dass 
Thamnidium elegans durch veränderte äussere Bedingungen 
gezwungen werden kann, diese oder jene Art von Sporangiolen 
zu bilden oder die Bildung derselben gänzlich zu unterdrücken. 
Kay^) säte die Sporen von Sterigmatocystis alba in Medien, 
welche aus Zucker, Stärke, Möhren, Kartoffeln, Gelatine und 
mineralischen Salzen bestanden, und erhielt verschiedene aus 
diesen Sporen entwickelte Pilzformen. 

Auch für das Thierreich fehlt es an diesbezüglichen Unter- 
suchungen nicht. Als Beispiele seien folgende angeführt : — 
Schmankewitsch") beobachtete, dass Branchipus stagnalis, der 
immer in Süsswasser gefunden wird, sich, wenn man ihn in 
versüsstem Meerwasser züchtet, der Form von Ärtemia Milhau- 
senii, einer das Brackwasser bewohnenden Art, nähert, und, 
wenn man das Brackwasser so lange concentrirt, bis dasselbe 
den Salzgehalt des Meerwassers erreicht hat, sich in Ärtemia 
salina, eine Meerwasser-Art, verwandelt. Herbst^) züchtete die 
Larven einiger Seeigel in verschiedenen Lösungen von Lithium-, 



1) J. Bac h mann. Einfluss der äusseren Bedingungen auf die Sporenbildung von 
Thamnidium elecjans Link. Bot. Ztg. 1895. Abt. I. p. 128. 

2) M. J. Ray. Variations des Champignons inférieurs sous l'influence du milieu. 
Revue générale de Botanique. 1897. T. IX. pp. 193-259 et pp. 283-304. 

3) W. Schmanke witsch. Zur Kenntniss des Einflusses der äusseren Lebensbeding- 
ungen auf die Organisation der Thiere. Zeitscli. f. wiss. Zool. 1887. Bd. XXIX. p. 429. 

*) C. Herbst. Experimentelle Untersuchungen über den Einfluss der veränderten 
chemischen Zusammenseizung des umgebenden Mediums auf die Entwicklung der Thiere. 
I. Theil. Zeitsch. f. wiss. Zool. 1892. Bd. XV. p. 446. 



106 A. YASUDA : ANPASSUNGSFAEHIGKEIT 

Natrium- und Kaliumsalzen, und fand, dass die Wirkungsstärke 
auf die Entwicklungsstufen derselben in einer Reihe der Salze 
von demselben Kadical ihrem Molekulargewicht umgekehrt pro- 
portional ist, d. h. ihre Wirkungsstärke nimmt um so stärker 
ab, je mehr ihre Molekulargewichte zunehmen. Nach ihm 
wurde in einem Falle die Gastrulation der Seeigel bedeutend 
verzögert, in einem anderen Falle wurde die Pluteusorganisation 
entweder mit der runden und gedrungenen Gestalt ohne Fortsätze 
oder sogar ohne eine Spur des Kalkgerüstes gewonnen. Cohn^) 
bemerkte, dass eine plötzliche Concentrationsänderung des 
Aussenmediums der Infusorien eine schädliche oder tödtliche 
Einwirkung ausübt. Fabre-Domergue") beobachtete auch das 
Verhältniss der Ernährung in den Körpern einiger Infusorien, 
und gelangte zu folgendem Schluss : ,,Dans des conditions par- 
faites de nutrition prise dans l'acception la plus large du mot 
il se produit des aliments de réserve qui disparaissent quand 
des conditions deviennent défavorables à la vie.*' Weiter 
studirte Bokorny^) die Veränderungen der Bewegung, der Ge- 
stalt und der Grösse der Vacuolen von Paramaeciiim unter dem 
Einfluss gewisser Basen, wie Coffein, Ammoniak und Kali, deren 
1 promill. oder noch dünnere Lösung im Allgemeinen die Bewe- 
gung verlangsamte, die Gestalt abrundete und sowohl Vergrösse- 
rung der Vacuolen als auch das Auftreten von neuen verursachte. 

*) F. Cohn. Entwickelungsgescliichte der microscopischen Algen und Pilze. Nova 
Acta Akad. Caes. Leopold. 1851. Bd. XXIV, Tli. 1. p. 132. 

-) M. Fabre-Domergue. Eecherches anatomiques et physiologiques sur les infu- 
soires ciliés. Ann. d. Sc. nat. Zool. 1888. Ser. VII, T. 5. p. 135. 

^) Th. Bokorny. Einige vergleichende Versuche über das Verhalten von Pflanzen 
iid niederen Thieren gegen basische Stoffe. Pflüger's Archiv. 1895. pp. 557-562. 

Bokorny gab auch in einer anderen Schrift (Vergleichende Studien über die Giftwir- 
kung verschiedener chemischer Substanzen bei Algen und Infusorien. Pflüger's Archiv. 
1896. pp. 262-306.) eine genaue Untersuchung über die Giftwirkung von Basen und Säuren 
unorganischer Natur, Salzen, Oxydations-Giften, Pliosphor, organischen Säuren, Alkoholen, 
Alkaloiden u. a. ni. auf das Leben der Infusorien und anderer Organismen. 



EINIGER INFUSORIEN. 107 

Endlich müssen noch die KesuUate der Untersuchungen von 
Davenport und NeaP) Erwähnung finden ; sie züchteten Stentor 
2 Tage lang in einer 0,00005% Sublimat enthaltenden Kultur- 
lösung ; die Thiere Hessen sich sehr wohl acclimatisiren und 
erwiesen sich gegen eine 0,0019^ ige sonst tödtliche Sublimat- 
lösung ca. vier Mal länger widerstandsfähig als diejenigen, die 
im Wasser kultivirt worden waren. 

Ueberblickt man die Ergebnisse dieser Untersuchungen, so 
ersieht man, dass sowohl den niederen ïhieren als auch den 
niederen Pflanzen die Fähigkeit innewohnt, sich geänderten 
Aussenmedien leicht anzupassen. Da aber diese Fähigkeit bei 
verschiedenartigen Organismen verschieden stark ausgeprägt ist 
und unter Umständen mannigfaltig auftritt, so muss jeder specielle 
Fall genau erforscht werden. Meine vorliegenden Studien sollen 
in dieser Hinsicht einen kleinen Beitrag bringen. 



Methodisches. 



Als Versuchsmaterial wählte ich solche Infusorien aus, die 
in Gräben und Teichen stets gefunden werden können. Da aber 
die in der freien Natur vorkommenden Infusorien nie in reiner 
Kolonie vorhanden sind, so Hess ich sie in einem Gefässe 
sich massenhaft entwickeln und unter Vorsichtsmassregeln eine 
längere Zeit fortleben. 

Genau nach den Angaben von Miyoshi") kultivirte ich 
die Infusorien in einem mit Spirogyra gefüllten Gefässe. Sobald 



1) C. B. Davenport and H. V. Neal. On the Acclimatization of Organisms to 
Poisonous Chemical Substances. Archiv für Entwicklungsmechanik der Organismen. 1895. 
Bd. II, Heft 4. p. 581. 

2) M. Miyoshi. Physiologische Studien über Cilia ten. The Botanical Magazine. 
Tokyo 1896. Vol. X, No. 112. p. 43. 



108 A. YASUDA : ANPASSUNGSFAEHIGKEIT 

die grüne Masse der Alge sich allmählich zu verfärben begann 
und die ursprünglich klare Flüssigkeit immer mehr getrübt 
wurde, bemerkte ich die Eutwickelung verschiedener Arten von 
Infusorien, die sich oft mit einer erstaunlichen Schnelligkeit 
vermehrten und sich auffallenderweise in bald fadenförmigen, 
bald netzartigen Kolonien gruppirten. Die microscopische 
Untersuchung ergab, dass diese Kolonien nur aus wenigen Arten 
bestanden, die im Kampfe ums Dasein den Kest überwunden 
hatten. Aus dieser Mischkultur isolirte ich einzelne Arten, indem 
ich mittelst einer Pipette eine kleine Menge der Kolonie zusam- 
men mit Wasser herausholte und in ein ebenfalls mit Brunnen- 
wasser und der Alge gefülltes Gefäss versetzte. Bei gewöhnlicher 
Zimmertemperatur zeigten diese Kulturen ca. in einer Woche üppige 
Eutwickelung ; nach vier oder fünf Wochen aber nahm die 
Vermehrung ab, und endlich nach sechs Wochen konnte nur 
noch eine ausserordentlich kleine Anzahl in der Flüssigkeit 
gefunden werden. Um eine und dieselbe Art immer in üppiger 
Kultur zu halten, legte ich deshalb alle drei Wochen neue 
Kulturen an und trug dafür Sorge, dass sie nicht etwa durch 
Bactérien inficirt wurden. 

Nachdem ich die gewünschten Arten auf solche Weise in 
Kultur hatte, wurden die Experimente auf zweierlei Weise 
ausgeführt : einerseits prüfte ich die Anpassungsfähigkeit der 
Infusorien in dem Zustande, wie sie in der Natur vorkommen, 
d. h. in ihrem Zusammenleben mit Bactérien ; anderseits wandte 
ich zu demselben Zwecke die Beinkultur jedes Infusors an, also 
frei von Bactérien. 

Der grösste Theil meiner Experimente wurde mit unreinen 
Kulturen ausgeführt ; in einigen Fällen wiederholte ich die 
Experimente an Beinkulturen, um zu wissen, ob die Gegenwart 



EIlSriGER INFUSORIEN. 109 

der Bactérien etwa das Ergebniss der Experimente modificivt 
hätte. Die Resultate stimmten aber bei beiden Kulturen vollkom- 
men überein, wie wir nachher sehen werden. 

Als äussere Medien verwendete ich Lösungen von Rohrzucker, 
Traubenzucker, Milchzucker, Glycerin, Kalisalpeter, Natron- 
salpeter, Chlorkalium, Chlornatrium und Chlorammonium in 
verschiedenen Concentrationen. Diese Stoffe waren chemisch 
rein und wurden vor dem Gebrauch vollständig getrocknet. 
Folgende Infusorien wurden bei meinen Studien ausschliesslich 
verwendet : Euglena viridis, Chilomonas ijaramaecium, Mallomonas 
Plosslii, Oolpidium colpoda und Paramaeciwn caudatimi. Alle 
Kulturen, sowohl unreine als reine, wurden bei Zimmertem- 
peraturen von 25°-30° C gehalten und in den Wintermonaten 
in einen Thermostat von etwa 30° C gestellt. 

Für unreine Kulturen benutzte ich eine grosse Anzahl der 
5 cm hohen und 3 cm weiten cylindrischen Glasgefässe, deren jedes 
30 ccm der Versuchslösung und etwa 1 Gramm Spirogyra-FïiàQw 
enthielt. Da die Infusorien im Brunnenwasser weit besser gedeihen 
als in destillirtem Wasser, so wendete ich bei der Zubereitung 
der flüssigen Versuchsmedien das letztere als auflösendes Mittel für 
verschiedene Substanzen au, wobei die Menge der darin gelöst 
vorhandenen Stoffe mit Ausnahme von Kochsalz^) so unbedeutend 
war, dass ich sie ohne grosse Ungenauigkeit ausser Acht lassen 
konnte. Bei vielen Kulturen, die gleichzeitig gemacht wurden, 
nahm ich keinen Anstand, eine Kontrollkultur, in welcher nur 
Brunnenwasser und Spirogyra-F'éiàQn angewendet wurden, anzu- 
fertigen und zum Vergleiche dienen zu lassen. 

In Bezug auf die Reinkultur der Protozoen im Allgemeinen 



') Die chemische Analyse des Brunnenwassers zeigte, class es 0,0%% Kochsalz 
enthielt. 



110 A. YASUDA : ANPASSUNGSFAEHLGKEIT 

haben viele Forscher^) in neuerer Zeit versucht, sie entweder auf 
festen Substraten oder in flüssigen Medien zu züchten. Bei der 
Isolirung der Infusorien befolgte ich genau die Methode von 
Ogata") mit positivem Kesultat. Ich liess mir nach seiner Vor- 
schrift feine Glascapillarröhren anfertigen, deren Durchmesser je 
nach der Grösse des Versuchsobjects variirten. So war z. B. bei 
C%ilomonas 'paramaeciumf dessen Körper 25-30 n Länge und 
10-12 li- Breite hat, das Capillarrohr etwa 0.1 mm in innerem 
Durchmesser und 10 cm in Länge, während es bei Calpidium 
colpoda, dessen Körper 60-70 /j- lang und 25-30 /^ breit ist, einen 
inneren Durchmesser von 0,15 mm und eine Länge von 10 cm 
hatte. 

Sobald das Capillarrohr nach dem Eintauchen in eine 
sterilisirte Nährlösung mit der letzteren grossen theils gefüllt 
war, brachte ich das nämliche Ende desselben in die Mischkul- 
turflüssigkeit von Bactérien und Infusorien und liess das Rohr 
sich mit der Flüssigkeit völlig füllen. Untersucht man ein 
solches Capillarrohr unter einem Microscope, so findet man an 
der Capillarrohrmündung eine grosse Anzahl von Infusorien in 
Bewegung. Einige streben sich ins Innere zurückzuziehen, bald 
aber kommen sie nach dei* Mündung zurück. Wegen der starken 
Aërotaxis und schwachen Chemotaxis der Organismen^) gelingt 

*) Während Beijerinck (Kulturversuche mit Amöben auf festen Substraten. Centralbl. 
f. Bak. u. Parasit. 1896. Bd. XIX, No. 8), Celli (Die Kultur der Amöben auf festen Sub- 
straten. Centralbl. f. Bak. u. Parasit. 189G.Bd. XIX, No. 14/15.), Schardinger (Reinkulturen 
von Protozoen auf festen Nährboden. Centralbl. f. Bak. u. Parasit. 1896. Bd. XIX, No. 14/15.), 
Gor in i (Die Kultur der Amöben auf festen Substraten. Centralbl. f. Bak. u. Parasit. 1896. 
Bd. XIX, No. 20), Tisch utkin (Ueber Agar-Agarkulturen einiger Algen und Amöben. 
Centralbl. f. Bak., Parasit, u. Infekt. 1897. Bd. III, No. 7/8.) und andere Forscher Amöben 
auf festen Substraten künstlich züchten konnten, ist es Ogata (Ueber die Reinkultur gewis- 
ser Protozoen-Infusorien. Centralbl. f. Bak. u. Parasit. 1893. Bd. XIV, No. 6.) auch gelun- 
gen, Polytonia uvella in flüssigen Medien rein zu kultiviren. 

=^) M. Ogata, loc. cit. p. 168. 

^) M. Mi y os hi. loc. cit. p. 48. 



EINIGER INFÜSORIEX. Ill 

es nicht immer, die Infusorien auf diese Weise hervorzuloeken, und 
ihrer habhaft zu werden. Xur wenn sie zufällig in die sterilisirte 
Flüssigkeit tief eindringen, kann man unter dem Microscope das 
Capillarrohr an der betreffenden Stelle abbrechen und dann das 
Ende des Rohrs zuschmelzen. Sodann impft man den infu- 
soriumhaltigen Capillarrohrinhalt, indem man das Rohr mit 
einem sterilisirten Pincet abbricht und den Inhalt in ein mit 
sterisirter Nährlösung gefülltes Reageusglas hineinbläst. 

Bei meinen Versuchen impfte ich wenigstens zwei Individuen 
in ein und dasselbe Reagensglas, um erstens den Effect der Inocu- 
lation zu sichern, und zweitens mit der Hoffnung, dass sie, wenn 
alle beide in dem neuen Medium unversehrt fortlebten, durch 
Copulation sich vermehren könnten. Bei der Zimmertemperatur 
von 20° C blieb die geimpfte Nährflüssigkeit nach zwei oder 
drei Tagen vollkommen klar, und erst nach ungefähr zehn Tagen 
erschienen Hunderte von Individuen, die nahe der Oberfläche 
der Flüssigkeit als sehr kleine weisse Pünktchen hin und her 
schwammen. Die Zahl dieser weissen Pünktchen nahm hernach 
allmählich zu, und dieselben waren nicht allein am oberen Theil 
des Reagensglases, sondern auch am mittleren und unteren Theil 
desselben zerstreut sichtbar. Eine solche Erscheinung bedeutet, 
dass die Reinkultur gut ausgefallen ist, und dass man in jener 
Nährlösung nichts anders als die isolirte Art der Infusorien 
findet. Wenn aber die Nährlösung während der Impfung von 
Bactérien inficirt wird, so tritt immer eine starke Trübung zu 
Tage, und man bekommt in diesem Falle selbstverständlich 
keine Reinkultur der Infusorien. 

Die auf diese AYeise hergestellte Reinkultur gedieh 4-5 
Wochen lang, vorausgesetzt dass die Nährstoffe in der Kultur- 
flüssigkeit nicht völlig erschöpft waren. Durch erneuerte Wieder- 



112 A. YASUDA : ANPASSÜNGÖFAEHIGKEIT 

impfung konnte ich die Organismen in reinem, gutem Kultur- 
zustande eine lange Zeit erhalten. 

Die Nährlösung, die ich für die Reinkultur gebrauchte, 
stellte ich nach der Vorschrift von Ogata an, und zwar war 
ihre Zusammensetzung folgende : 

Flfcischextract 1 g 

Rohrzucker 20 „ 

Concentrirt gekochte Lösung von Porphyra vulgay^is. 250 ccm 
Destillirtes Wasser 729 ,, 



Beschreibung der Versuche. 

Wie gesagt, stellte ich die Experimente hauptsädilich mit 
unreinen Kulturen an, in der Absicht, den Einfluss, welchen die 
äusseren Bedingungen auf die Infusorien in ihrem natürlichen 
Vorkommen ausüben, festzustellen. Dabei versäumte ich aber 
nicht, Kontrollversuche mit reinen Kulturen zu machen und die 
beiden Resultate zu vergleichen. 

Bei allen Versuchen mit unreinen und reinen Kulturen Hess 
ich die Beschaffenheit des Mediums sich plötzlich ändern und 
prüfte die Anpassungsfähigkeit unserer Organismen an das neue 
Medium. Hatte ich unreine Kulturen, so verglich ich gewöhn- 
lich im Verlauf von 1-7 Tagen, zuweilen aber erst nach einem 
Monat, die Wirkungen der verschiedenen Mediumsconcentrationen 
auf das Reproductionsvermögen und die Gestaltänderungen der 
Versuchsorganismen. Speciell bei den Versuchen mit Rohrzucker 
war es nöthig, eine Reinkultur anzuwenden, weil bei unreiner 
Kultur der Rohrzucker durch vorhandene Bactérien oder Pilze 
nach und nach invertirt und schliesslich gespalten wurde. Um 



EINIGER INFUSORIEN. 113 

ZU erkennen, nach wie vielen Tagen der Kobrzucker zum Trau- 
benzucker invertirt wird, prüfte ich mit der Fehlin g 'scheu 
Lösung und fand, dass in meinen Versuchen nach etwa 4 Tagen 
eine kleine Inversion stattgefunden hatte. Meine unreinen 
Eohrzuckerkulturen waren daher binnen der ersten drei Tage 
doch noch brauchbar. 

Dass diese Inversion ausser durch Bactérien und Pilze auch 
unter Mitwirkung der Infusorien stattfände, ist schon von vorn 
herein unwahrscheinlich. Um dies aber experimentell zu con- 
statiren, stellte ich einige Versuche mit Kohrzuckerreinkulturen^) 
an, und gelangte beim Prüfen der fraglichen Flüssigkeit wie 
erwartet zu negativem Resultat. Die Koutrollkultur mit Asper- 
gillus glaiicus zeigte eine starke Inversion. 

Bei allen Kulturen mit verschiedenen Stoffen machte ich 
immer Kontrollkulturen, und bei den kritischen Versuchen, wie 
z. B. der Bestimmung der Concentrationsgrenze einer Flüssigkeit, 
in welcher die Organismen sich mehr oder minder anpassend 
leben können, wurden dieselbe Kulturen einige Male wiederholt. 

Ich gehe nun zur Beschreibung der einzelnen Versuche bei 
jeder Art meiner Versuchsorganismen über. 

(a) Euglena viridis Ehrbg.^) 
Dieser Organismus hatte in der Kontrollkultur folgende 



') Da der Rohrzucker bekanntlich bei langem Kochen zum Theil in Trauben- und 
Fruchtzucker verwandelt wird, so kann bei den Rohrzuckerreinkulturen die gebräuchliche 
Sterilisirung durch Hitze nicht ohne Vorsichtsmassregeln angewendet werden. Ich sterilisirte 
deshalb den Rohrzucker mit absolutem Alkohol und brachte ihn dann in die vorher 
sterilisierte Nährhisung ein. 

^) Figuren in Friedrich Ritter v. Stein, Der Organismus der Infusionsthiere. 
Leipzig 1878. Abt. III, Heft I. Taf. XX., W. Sa ville Kent, A Manual of the Infusoria. 
1880-81. Vol.1, n. XX. und O. Biitschli, H. G. Bronn's Klassen und Ordnungen des 
Thierreiches. 1ÖS3-87. Bd. I. Protozoa, Abt. 11. Taf. XLVII. 



114 A. YASUDA : ANPASHUNGSFAEHIGKEIT 

Merkmale : Gestalt gewöhnlich spindelförmig, Hinterende schärfer 
zugespitzt, aber wegen der Metabolie sich mannigfaltig verän- 
dernd. Aus dem Schlünde entspringt eine lange Geissei. Chro- 
matophoren zahlreich vorhanden, klein, scheibenförmig und rein 
grün gefärbt. Eine contractile Vacuole nahe dem Vorderende 
gelegen. Dicht bei demselben Ende befindet sich auch ein rother 
Augenfleck. 

Versuch 1. Rohrzucker, C1.2H22O11. — Von einer l'?^igen 
Lösung anfangend Hess ich in anderen Kulturen die Concentra- 
tion um je l^^o steigen. Obgleich die Accommodation schwer 
wurde, als die Concentration zunahm, so lebte das Infusor doch 
bis zur 15 9^ igen Lösung, welche die Maximalconcentration für 
den Organismus war. l^^ige, 29^ ige und 3^oige Kulturen 
zeigten keine wesentliche Veränderung am Körper des Organis- 
mus. Bei einer 4'^ igen Lösung aber begannen die Chromato- 
plioren an Grösse zuzunehmen. Von l°/o iger Lösung bis zu 
79^iger war die spirale Bewegung des Organismus lebhaft, 
dagegen über S^^o wurde sie allmählich langsamer, während die 
Chromatophoren selbst sich merklich ausdehnten ; als die Con- 
centration des Mediums zunahm, wurde auch die Vermehrung 
verhindert. Bei 12 9^ iger Lösung konnte das Thier nicht mehr 
normal gedeihen, bei 139ö überlebte eine kleine Anzahl, die 
jedoch nach einer Woche alle zu Grunde gingen ; bei 14^^ lebten 
noch einige Individuen, aber nicht länger als 4 Tage, während 
sie bei 15"?^ kaum einen Tag lebendig blieben. Da der Organis- 
mus metabolisch ist, so konnte keine deutliche Veränderung an 
seiner äusseren Gestalt beobachtet werden. 

Versuch 2. Traubenzuclcer, Q^li-O^^. — Unser Organismus 
konnte 1-11 ?o ige Concentrationen ertragen. Bei l"/o- und 1%- 
Kultur war noch keine merkliche Veränderung wahrzunehmen, 



EINIGER INFUSORIEN. 115 

aber schon bei o^^iger Lösung dehnten sich die Chroraatophoren 
ein wenig aus, und bei der Concentration über 3% wurden sie 
noch etwas grösser. Die Bewegung des Thierkörpers schien bei 
1-69^ -Kulturen normal zu verlaufen ; erst bei 79^ wurde sie 
langsamer mit gleichzeitiger Verminderung der Vermehrungs- 
fähigkeit. Bei einer 9% igen Lösung vermehrten sich die Tliiere 
überhaupt nicht mehr, und nach einer Woche war nur noch 
eine kleine Anzahl am Leben. Alle Individuendes Infusoriums 
gingen bei einer 109^- Kultur nach einer Woche, und bei einer 
von 11^ schon nach einigen Tagen zu Grunde. 

Versuch 3. Milchzucher, Ci2H220n + .H2O. — Unter den oben 
erwähnten Zuckerarten schien unser Infusor sich an Milchzucker 
am besten anzupassen. Die Maximalconcentration, welcher es 
widerstehen konnte, war eine 179^ige. Von 4?^ an aufwärts 
schienen die Chromatophoren sich zu vergrössern. Bei 1-11^^- 
Kulturen nahm die Multiplication rasch zu, aber über 12% 
wurde sie etwas vermindert und auch die Bewegung wurde 
einigermassen träge. Eine VJ^oigQ Lösung erwies sich als die 
Grenzconcentration für das Versuchsthier. 

Versuch 4. Glycerin, CsHgOg. — Die Versuche lehrten uns, 
dass sich unser Infusor an Glycerin weit schlechter anpasste als 
an eine der oben erwähnten drei Zuckerarten, denn das Thier 
konnte nur 1-6^ ige Lösungen ertragen. Bei einer 29^ igen 
Lösung erweiterten sich die Chromatophoren ; bei 3 9^- Kultur 
lebte eine kleine Anzahl noch am fiuiften Tage, und bei 69^ 
blieben nur wenige Individuen noch einige Tage am Leben. 
Die Bewegung wurde bei einer 49^ igen Concentration schon 
vielfach retardirt, und bei 69^ hörte sie fast gänzlich auf. 
Ferner war in der letzteren Lösung eine pathologische Erschei- 
nung wahrzunehmen, indem die Cuticula des Körpers um die 



116 A. YASUDA : ANPASSUNGSFAEHIGKEIT 

Chromatoplioren etwas einschrumpfte, sodass ihr Umrîss im 
optischen Schnitte gesehen zickzackförmig aussah, und die 
Chromatophoren selbst verschmolzen mehr oder weniger mit 
einander. 

Versuch 5. Schivefehaures Magnesium, MgS04. — Unter den 
unorganischen Substanzen erwies sich das schwefelsaure Magnesium 
als dem Leben des Organismus am besten zusagend. Der 
Organismus konnte die Concentration von 1-QYo vertragen. Die 
Chromatophoren nahmen in ihrer Grösse fortwährend zu, als die 
Concentration von 1,5?^ bis auf ihr Maximum stieg. In einer 
3,49^ igen Lösung zeigte das Thier eine sehr träge Bewegung, 
und schon bei 4-6 9^ igen Lösungen ging es beinahe zum 
Stillstand über. Betrug die Concentration nur 1-2,5 %, so gedieh 
unser Thier einen Monat lang vollkommen normal, aber von 
einer 2, 6^^ igen Concentration an aufwärts büsste es seine Ver- 
mehrungsfähigkeit ein, und endlich bei b-Q>% blieben nur 
vereinzelte Individuen am Leben. 

Versuch 6. Salpetersaures Kalium, KNOo. — Der Organismus 
widerstand einer 2,49^ igen Concentration. Von 0,W an fingen 
die Chromatophoren an sich zu erweitern ; über 2?^ wurde die 
Bewegung sehr langsam. Im Allgemeinen schien das vorliegende 
Salz auf die Mulptiplication des Infusors hemmend zu wirken, 
da selbst bei verdünnten Lösungen das Thier eine unbedeutende 
Vermehrung zeigte. 

Versuch 7. Salpeter saures Natrium, NaN03. — Dieses Salz 
verhielt sich fast wie das vorige. Eine 2'^ige Lösung war das 
Maximum, welches unser Thier ertragen konnte. Die Chromato- 
phoren dehnten sich schon von 0,8^ an aus. Die Bewegung 
war bei 2^^iger Lösung nach zwei Tagen sehr träge. 

Versuch 8. Chlorhalium, KCl. — Nächst dem Magnesium- 



EINIGER INFUSORIEN. 117 

Sulfat erwies sich unter den unorganischen Stoffen das Clilor- 
kalium für die Vermehrung des Organismus am günstigsten. 
Eine 0,7 "^o ige Lösung verursachte sowohl Zahl Vermehrung als auch 
Volumenerweiterung der Chromatophoren. In 0,2-l9oigen 
Coucentrationen gedieh der Organismus noch nach 40 Tagen. 
Stieg die Concentration auf 2,S^o, welches die maximale Grenze 
für den Organismus war, so hörte die Bewegung fast gänzlich 
auf, während die Chromatophoren sich theilweise zu grösseren 
Körnern verschmolzen. 

Versuch 9. Chlor natrium, NaCl. — 0,2-1 ,8^0 waren die Cou- 
centrationen, bei welchen das Thier am Leben blieb. Die 
Chromatophoren schienen bei einer 0,89^igen Lösung au Grösse 
zuzunehmen, und bei einer l,6?^igen Concentration zeigte der 
Organismus noch eine langsame Bewegung. 

Versuch 10. Chlorammonium, NH4CI. — Dieses Salz wirkte 
unter allen oben genannten Stoffen am ungünstigsten auf das 
Leben des Organismus ein, sodass die Anpassungsgrenze hier am 
niedrigsten war. 0,2-0,69^ Kulturen gediehen noch am Ende der 
dritten Woche, aber über 1?^ vermehrte sich das Thier nicht mehr, 
und bei 1,4?^ lebten kaum noch einige Individuen. Bei 0,69^ 
iger Lösung nahm die Grösse der Chromatophoren zu, und bei 
lYo verschmolzen sie sich zu wenigen grösseren Körnern. Eine 
1,49^ ige Lösung verursachte immer die Verschmelzung der 
Chromatophoren und hob gleichzeitig die Bewegung des Orga- 
nismus auf. 

Ich wiederholte dieselbe Versuche zehnmal mit Reinkulturen 
und verglich die Resultate mit denjenigen bei den unreinen Kul- 
turen. Die Ergebnisse stimmten in beiden Fällen völlig überein. 
Weiter beobachtete ich, dass bei den Versuchen mit Reinkulturen 
die Schnelligkeit der Multiplication für die Lösungen verschiede- 



118 A. YASUDA : ANPASSUNGSFAEHIGKEIT 

ner Coiicentrationen eines und desselben Stoffes niclit allzu 
gleich war, obgleich sie gleichzeitig geimpft worden waren 
Im Allgemeinen nahm die Vermehrungsenergie in dem Masse 
ab, wie die Concentration des Mediums stieg ; so war zum 
Beispiel bei einer Traubenzuckerkultur, im Verlaufe von 2 
Woclien nach der Impfung, die Vermehrung eine starke bei 
2^/0, eine massige bei 4*?^, eine sehr unbedeutende bei 69^, eine 
noch spärlichere bei 8Y0 und keine bei 10^. Ferner war die 
Vermehrung bei derselben Kultur nach 4 Wochen bei 2% eine 
sehr starke, bei i^o eine starke, bei 69^ eine massige, bei 89^ 
eine unbedeutende und bei 109^ eine höchst spärliche Vermehrung 
zu beobachten, während sich am Ende der sechsten Woche bei 
2-4^/o eine sehr starke, bei 6^/0 eine starke, bei 8% eine massige 
und bei lO^ö eine spärliche Vermehrung zeigte. Auch beim 
Milchzucker wurden ähnliche Thatsachen constatirt. So gediehen 
nach 2 Wochen 2-49^ -Kulturen ausgezeichnet ; 69^ -Kultur zeigte 
eine starke, 89^ eine massige, 109o eine spärliche, 129^ eine noch 
schwächere und 149^ gar keine Multiplication mehr. Nach 4 
Wochen aber vermehrten sich die Organismen bei 2-69^ sehr 
stark, bei 89^ stark, bei 109^ massig, bei 129^ spärlich und bei 
149^ sehr spärlich. Endlich nach 6 Wochen gedieh die Multi- 
plication stark bei 109^, massig bei 129^ und spärlich bei 149^. 
Auch für schwefelsaures Magnesium, salpetersaures Kalium, 
Chlornatrium u. s. w. habe ich ähnliche Erscheinungen wahr- 
genommen. 

(b) Chilomonas paramaecium Elirbg.^) 
Der Organismus in Kontrollkultur hatte folgende Charak- 
teristika : Körper nicht metabolisch, sondern plastisch. Gestalt 

') Figuren in Friedrich Ritter v. Stein, loc. cit. Taf. XIX, W. SaviUe Kçut. 
loc. dt. n. XXIV und O. Bütschli. loc. cit. Taf. XLV. 



EINIGER INFUSORIEN. 119 

länglich oval, seitlich comprimirt. Aborde rende breiter und schief 
abgestutzt, Hinterende dagegen rundlich zugespitzt. Zwei 
Geissein am Vorderende, eine derselben rollte sich, wenn der 
Organismus im Ruhezustande war. Zahlreiche sphäroidische 
Amylumkörner dicht unter der Körperoberfläche. Eine contractile 
Vacuole am Vordertheil des Körpers. 

Versuch 1. Rohrzucker. — Mit einer l^igen Lösung anfangend 
liess ich die Concentration um je l^/o steigen, wie es bei Euglena 
viridis der Fall war. Von 2% an aufwärts dehnten sich die 
Körnchen fortwährend aus. Ueber 6^^ nahm der Organismus 
an Dicke und Breite zu, nicht aber au Länge, und sah so 
einfach oval aus. Die Vermehrung wurde schon bei 4.^/o verzögert; 
bei 7?^ lebten einige Individuen noch eine Woche lang. Die 
Individuen aus der 7 feigen Kultur waren so träge, dass sie an 
einem bestimmten Platze still lagen und nur eine zitternde Be- 
wegung zeigten ; vor ihrem Tode hüpften sie einige Male 
rückwärts. Die letztere Erscheinung wurde auch bei den con- 
centrirten Lösungen anderer Stoffe beobachtet. 

Versuch 2. Traubenzucker. — Der Traubenzucker wirkte stärker 
als der Rohrzucker. Die höchste Concentration, die der Orga- 
nismus vertragen konnte, war 69^. Bei 4^^ -Kultur ging die 
Vermehrung nicht mehr gut vor sich, und bei 59» blieb nur 
eine kleine Anzahl der Individuen am Leben. 2^/oige Con- 
centration bewirkte, dass die Körnchen sich vergrösserten, und 
bei 5^/0 wurde die Bewegung sehr langsam. Bei 6^ kam der 
Organismus fast zum Stillstande, und wurde eine Unebenheit des 
Körperumrisses hervorgerufen. 

Versuch 3. Milchzucker. — Der Organismus widerstand l-S^/o- 
igen Concentrationen. Ueber 3% vergrösserten die Körnchen ihr 
Volumen, bei 6?« wurde die Multiplication verhindert und end- 



120 A. YASUDA : ANPASSUNGSFAEHIGKEIT 

lich bei S^/o lebten nur noch einige Individuen eine Woche lang 
weiter, mit dem erwähnten UnebeD werden der Körperumrisse. 
Merkwürdig war, dass der Körper an Dicke und Breite zunahm, 
als die Concentration stieg. 

Versuch 4. Glycerin. — Eine 49^ige Lösung war die Maxi- 
malconcentration für die Accommodation des Thieres. Bei 29^ 
erweiterten sich die Körnchen, bei Sfo hörte die Vermehrung auf, 
und bei 4/o lebten nur noch einige Individuen, deren Körper 
unregelmässige Umrisse zeigten. 

Versuch 5. Schwefelsaures Magnesium. — In 1-3 'folgen Lö- 
sungen lebte der Organismus fort. Von 0,8 "^o an aufwärts ver- 
grösserten sich die Körnchen, und in 3'?oiger Lösung wurden 
dieselben auffallend gross. Eine 1,4?« ige Lösung verhinderte die 
Multiplication. Bei höheren Concentrationen trat bei einigen In- 
dividuen ein unregelmässiges Aussehen zu Tage. Diese Gestalt- 
änderung wurde bei einer 2,59^ -Kultur besonders gut beobachtet, 
indem alle Individuen noch 2 Wochen mit einer ungewöhnlichen 
Unebenheit ihrer Körpergestalt fortlebten. 

Versuch 6. Salpetersaures Kalium. — Eine 2'?öige Concentra- 
tion schien die obere Grenze der Anpassung zu sein. Eine 
0,89^ ige Lösung veranlasste eine Vergrösserung der Amylum- 
körner, die bei einer 1 böigen Lösung nach einer Woche einen 
sehr grossen Durchmesser zeigten. Die Vermehrung ging nur 
bei niederen Concentrationen gut vor sich. 

Versuch 7. Salpeter saur es Natrium. — Der Organismus konnte 
in 0,2-1,2^0 igen Lösungen leben. Die Volumenzunahme der 
Körnchen fand von 0,69^ an statt. Bei höheren Concentrationen 
gedieh unser Organismus nicht. Im Uebrigen fast dieselben 
Erscheinungen wie beim vorhergehenden Versuche. 

Versuch 8. Chlorkalium. — Eine 0,8?« -Kultur am Ende des 



EINIGER INFUSORIEN. 



121 



dritten Tages nach der Impfung untersucht zeigte sowohl Yer- 
grösserung der Körnchen als auch Abrundung des Körpers. Nach 
Verlauf einer Woche besassen einige Individuen in einer l^/o 
igen Lösung eine fast scheibenförmige Gestalt, lieber 29^ konnten 
sie nicht mehr leben. 

Versuch 9. Chlornairium. — Eine 0,4^« ige Lösung Hess die 
Körnchen sich erweitern. Bei 0,8-1^^ -Kulturen dehnte sich 
ihr Volumen bedeutend aus, indessen der Körper des Organismus 
sich verkürzte und rundlich wurde. Bei l9^iger Concentration 
war die äusserste Grenze der Accommodation erreicht. 

Versuch 10. Cldorammoniiün.—Der Organismus konnte sich 
an O,l-O,69oige Concentrationen anpassen. Bei einer 0,2^o 
igen Lösung trat schon Körnchenvergrösserung ein, und bei 
0,4^0 zeigten einige Individuen unebene Umrisse, mit gleichzei- 
tiger Abschwächung ihrer Bewegung. Wie bei EiigJena viridis 
so auch bei Ckilomonas paramaeciiim übte der vorliegende Stoff 
von allen auofewandten Chemikalien die stärkste Einwirkung aus. 



•ft^ 



(c) Jlallomonas Plcsslii Perty.^) 

Der Organismus in der normalen Kultur zeigte folgende 
Merkmale : Gestalt oval, am Vorderende etwas schmaler. Die 
ganze Cuticularoberfläche mit langen, biegsamen, borstigen 
Wimpern bekleidet ; am Hinterende mit einer langen Geissei 
versehen. Anstatt der Amylumkörner war eine Anzahl von 
Vacuolen vorhanden. Eine contractile Vacuole befand sich nahe 
dem hinteren Ende. Das Tliier schwamm mit lebhafter Bewegung, 
wobei es oft plötzlich stillstand. 

Versuch 1. Rohrzucker, — Der Organismus vertrug Anj^as- 



Figuren in W. S. Kent. loc. cit. PI. XXIV. 



122 A. YASUDA : ANPASSÜNGSFAEHIGKEIT 

sungscoiicentrationen von 1-7 ^/o. Auch bei höheren Concentra- 
tionen nahm die Grösse des Körpers mehr oder minder zu, 
während die Multiplication und die Lebhaftigkeit der Bewegung 
allmählich sanken. Ferner wuchs die Zahl und Grösse 
der Vacuolen mit der Concentrationserhöhung bedeutend an ; 
diese Erscheinung trat schon bei einer 2"^ -Kultur ein. Erst 
von 4% an wurde die Multiplication schwächer und bei 7fô 
erlitt die Bewegung eine Betardirung, welche zu der sehr 
schnellen, normalen Bewegung in grossem Contrast stand. 

Versuch 2. Traubenzucher. — Eine 6'^^ -Kultur war das Maxi- 
mum der Anpassung des Organismus. Die Vergrösserung des 
Körpers und die Zunahme der Vacuolen bei stärkeren Concen- 
trationen waren wie beim Bohrzucker. Eine 2?oige Lösung 
erweiterte die Vacuolen einigermassen, und von o9^ an nahm 
die Vermehrung des Organismus ab. Die Bewegungshemmung 
war schon bei A^o zu beobachten, noch stärker bei b°/o. 

Versuch 3. MilchzucTcer. — Das Infusor ertrug 1-9 9^ ige 
Lösungen. Das allgemeine Besultat stimmte mit dem bei den 
anderen Zuckerarten überein ; der einzige Unterschied war der, 
dass der Milchzucker eine schwächere Einwirkung auf den 
Organismus ausübte als die anderen Zuckerarten. Die Vacuolen 
fingen erst bei einer 39^ igen Lösung an sich zu vermehren, und 
die Multiplication wurde erst von 7?o an etwas verlangsamt. 

Versuch 4. Glycerin. — Die Grenze der Accommodation war 
eine 4'?oige Lösung. Bei einer 2^,o- und noch auffallender bei 
einer 39o -Kultur fand Zunahme der Zahl und Grösse der Vacuolen 
und Anschwellen des Organismus statt. 

Versuch 5. Sc! iwe feisaures Magnesium. — Der Organismus 
vermochte sich 1-3,4?^ igen Lösungen anzupassen. Eine 0,8 ?ö 
ige Concentration verursachte sowohl Vermehrung als auch 



EmiGER INFUSORIEN. 123 

Vergrösserung der Vacuolen, und bei höher concentrirten Lö- 
sungen war Verhinderung der Multiplication und Retardation der 
Bewegung wahrzunehmen. 

Versuch 6. Salpeter saur es Kalium. — Eine 0,7 9o ige Lösung 
vergrösserte die Vacuolen etwas. Die Bewegung begann bei 
1-1,5?^ igen Lösungen sehr langsam zu werden. Bei den Kul- 
turen höherer Concentrationen pflegte der Organismus sich nicht 
zu vermehren, und im Verlaufe einiger Tage ging der grössere 
Tlieil der Individuen zu Grunde. Eine 1,59^ ige Lösung bildete 
die Grenze der Anpassung. 

Versuch 7. Salpetersaures Natrium. — Für diesen Stoff besass 
der Organismus eine besonders grosse Resisteuzkraft. Er ver- 
mochte sich sogar einer 2,6^ igen Concentration anzupassen, 
wenn auch mit grosser Schwierigkeit. Die Bew^egung war bei 
1,09^ noch lebhaft. 

Versuch 8. Chlorhalium. — Bei einer 0,8?^ -Kultur vergrösser- 
ten sich die Vacuolen. Die Grenze der Anpassung des Organis- 
mus war bei l,4'?^iger Lösung zu beobachten. Mit der Con- 
centrationssteigerung trat Körperabrundung ein. 

Versuch 9. Chlornatrium. — Die Maximalconcentration war 
l,5fo. Bei 0,89ö -Lösung schien der Körper nach 5 Tagen sich 
abzurunden. In einer 1 böigen Lösung konnte das Thier 3 
Wochen lang gedeihen, aber in l,o?o starb es schon am Ende 
des vierten Tages. 

Versuch 10. C%lor ammonium. — Für diesen Stoff besass der 
Organismus die kleinste Anpassungsfähigkeit, ganz wie es bei 
den anderen Infusorien der Fall war. Eine 0,8^ ige Lösung 
war das Maximum. Die Cuticularoberfläche des lebenden 
Organismus zeigte in dieser Lösung nach einem Tage einige longi- 
tudinale Falten, 



124 A. YASUDA : ANPASSUNGSFAEHIGKEIT 

(d) CoIpicUum colpoda Ehrbg.^) 

Merkmale des Organismus in der normalen Kultur : — 
Körper mittelgross, nierenförmig. Eückenseite massig gewölbt; 
Bauchseite in der Nähe des Mundes etwas eingebuchtet. Vor- 
derende viel schmaler als das abgerundete Hinterende. Cuti- 
cularwimpern auf der ganzen Oberfläche des Körpers reichlich 
vorhanden und an Grösse alle gleich. Mund in massiger Ent- 
fernung vom Vorderende, in einer die Bauchseite querenden 
Einbuchtung. Eine contractile Vacuole und einige Nahrungs- 
vacuolen vorhanden. Bewegung lebhaft. 

Versuch 1. Rohrzucker. — Die Concentrationsdififerenz der 
Versuchsserie war l^/o. SYo wurde als das Maximum erkannt. 
Schon bei einer 3?^ igen Lösung begannen die Vacuoleu sich 
etwas zu vermehren und zu vergrössern. Diese Erscheinung 
wurde mit der Concentrationserhöhung immer mehr merklich, 
lieber 4o.{, sah der Körperumriss rundlich aus und die Grösse 
nahm merkwürdig zu. Multiplicationshemmung schon bei 6Yo. 

Versuch 2. Traubenzucker. — Der Organismus lebte in 1-7?^ 
igen Lösungen. Vermehrung und Vergrösserung der Vacuolen 
schon bei 2^/o und Abrundung des Körpers bei o^/o. Bei einer 
4,59^ -Kultur wurde die Multiplication sehr verzögert, bei ߥ> 
lebte am Ende des fünften Tages noch eine kleine Anzahl der 
Thiere ; bei 1¥o waren nur noch vereinzelte Individuen am 
Leben, welche schliesslich nach 5 Tagen abstarben. 

Versuch 3. llilchzucker. — 1-109^ ige Lösungen wurden ver- 
tragen. Vacuolenvergrösserung von o^o an aufwärts und Kör- 
perabrundung über 4^/o. Bei einer 7^^ igen Lösung wurde die 
Vermehrung verzögert, und bei lOfo konnten nur einige Indi- 



») Figuren in O. Bütschli. hc. cit. 1887-89. Abt. III. Taf. LXII. 



EINIGER INFUSORIEN. 125 

viduen 10 Tage lang leben. Auch hier fand mit der Concentra- 
tionssteigernng Grössenziinahme des Körpers statt. 

Versuch 4. Glycerin. — Maximalconcentration b^/o. Vermeh- 
rung und Vergrösserung der Vacuolen bei 2°/o ; Körperabrundung 
bei 39^. Bei 5'?^ lebten nur noch vereinzelte Individuen wenige 
Tage lang. 

Versuch o. Schivefelsaiires 3Iagnesiuni. — Anpassungscoucen- 
tration : 1-0%. Vacuolen vergrösserung und Körperabrundung 
begannen bei 2"/o. Bei schwächeren Concentrationen gedieh der 
Organismus gut, aber über ?y% schlecht. 

Versuch 6. Salpetersaures Kalium. — Maximalconcentration 
2%. Das Thier gedieh bei 0,89^ iger Lösung nicht mehr. 
Zahlzunahme und Vergrösserung der Vacuolen waren wie ge- 
wöhnlich. 

Versuch 7. Salpetersaures Natrium. — Anpassungsconcentra- 
tion : 0,1-2%. Ueber O,89o nahm die Vermehrung stufenweise 
ab. Gestaltänderung u. s. w. waren ähnlich wie in den vorher- 
gehenden Fällen. 

Versuch 8. Chlorhalium. — Anpassungsconcentration : 0,2- 
1,69^. Die höheren Concentrationen über 0,89« verursachten 
Multiplicationshemmung. Körperabrundung von 0,69^ an. 
Vacuolen vergrösserung fand auch bei stärkeren Lösungen statt. 

Versuch 9. Ghlornatrium. — Maximalconcentration 1,590. 
Volumenvergrösserung der Vacuolen wie gewöhnlich. 

Versuch 10. Chlorammonium. — Der Organismus konnte sich 
nur äusserst verdünnten Lösungen anpassen. Schon bei 0,29^ 
trat Vacuolenvergrösserung ein, bei 0,8^0 Bewegungshemmung 
und Un regelmässig werden der Körperumrisse. Bei 1% , der 
höchsten Concentration, welcher das Thier widerstand, waren 
einige Individuen nach 2 Tagen noch lebendig. 



126 A. YASUDA : ANPASSUNGSFAEHIGKEIT 

(e) Paramaecium caudatum Ebrbg.^) 

Merkmale des Organismus in der normalen Kultur : Körper 
verlängert, spindelförmig, biegsam. Wimpern überall an der 
Oberfläche des Körpers, dicht und gleichmässig. Trichocysten 
senkrecht zur Oberfläche in der unter der Cuticula unmittelbar 
befindlichen Rindenschicht gelegen. Mund nahe der Mitte der 
Bauchseite. Schlund ziemlich lang. Zwei contractile Vacuolen 
am vorderen und hinteren Ende, mit strahligen zuführenden 
Kanälen. Nahrungsvocuolen vorhanden. Bewegung lebhaft. 
Vermehrung langsam. 

Versuch. 1. Rohrzucker. — Anpassungsconcentration 1-79^. 
Von 3% an aufwärts bis 7% Zahl- und Durchmesserzunahme 
der Vacuolen. Ueber 4% Dick werden des Körpers ; bei 7Yo 
blieb das Thier noch viele Tage lang lebendig. 

Versuch 2. Traubenzucker. — Anpassungsconcentration : 1-5^. 
Vacuolenvergrösserung bei ca. 2fo, Körperabrundung bei 3Yo. 
Sonst wie beim vorhergehenden Versuch. 

Versuch 3. llilchzucker. — Maximalconcentration 8^. Ver- 
mehrung und Vergrösserung der Vacuolen bei 3% u. s. w. In 
höheren Concentrationen erreichten die Durchmesser der Vacuolen 
bedeutend grössere Dimensionen, und der Körper erhielt ein 
fleischiges Aussehen. 

Versuch 4. Glycerin. — Anpassungsconcentration 1-39^. In 
diesem Medium konnte das Versuchsinfusor nicht lange am Leben 
bleiben, Vacuolenvergrösserung und Körperabrundung wie bei 
den vorigen Versuchen. 

Versuch 5. Schwefelsaures Blagnesium. — Maximalconcentration 
2,49^. Obgleich der Körper bei 0,2^^ verlängert war, so wurde 

Figuren in O. ßütschli. loc. cit. 1887-89. Abt. III. Taf. LXIII. 



EINIGER INFUSOßlEN. 127 

er doch bei 2,49^ viel fleischiger, wobei sich auch die Vacuoleu 
vergrösserten. 

Versuch 6. Salpetersaures Kalium. — Anpassungsconcentration 
0,2-1'^. Die durcli dieses Medium hervorgebrachten Gestalt- 
änderungen waren fast dieselben wie die von Colpidium colpoda 
bei demselben Medium. 

Versuch 7. Salpetersaures Natrium. — Maximalconcentration 
1,2*^. Dickenzunahme des Körpers von ca. 0,1 % an. In 1,2'^ 
iger Lösung lebten nur noch vereinzelte Individuen mit schwacher 
Bewegung. 

Versuch 8. Chlorkalium. — Anpassungsconcentration 0,2-19^. 
In einer 1 9^ -Kultur starb das Thier nach 3 Tagen gänzlich ab. 

Versuch 9. Ghlornatrium. — Anpassungsconcentration 0,2-1'%. 
Hier fand mit Concentrationssteigerung auch Abrundung des 
Körpers statt. 

Versuch 10. Chlorammonium. — Maximalconcentration 0,5'%. 
Das Infusor accommodirte sich an dieses Medium am schwersten ; 
in keiner Kultur blieb es lange am Leben. 



Allgemeines und Schlussbemerkungen. 

Aus den oben angeführten Versuchen ergiebt sich, dass mit 
der Steigerung der Concentration unabhängig von der chemischen 
Beschafienheit die Cuticularoberfläche der Infusorienkörper ein- 
schrumpft, wenn die Organismen plötzlich in das Medium 
gebracht werden, weil durch couceutrirtere Medien das Wasser 
aus dem Thierkörper herausgezogen wird. Zugleich wird ihre 



128 A. YASüDA : ANPASSUXGSFAEHIGKEIT 

Bewegung, die bisher lebhaft gewesen war, immer langsamer, und 
nach einem kurzdauernden Zittern an einem Platze kommen die 
Thiere endlich zum Stillstande. Wenn aber die Concentration 
des Mediums nicht zu stark ist, so können es die Infusorien 
ohne grossen Schaden ertragen, und die einmal gebildeten longi- 
tudinalen Falten der Cuticularoberfläche verschwinden nach 
einiger Zeit wieder. Sogar bei concentrirteren Lösungen findet 
man nicht selten einige Individuen, welche mit der contrahirten, 
unebenen Körperoberfläche noch einige Tage lang fortleben 
können. 

Je höher die Concentrationen der Medien sind, desto 
schwerer wird selbstverständlich die Anpassung, und wenn sich 
schliesslich die Maximumgrenze nähert, so stirbt der grössere Theil 
der Individuen ab. Im Falle gelungener Anpassung an ein 
gewisses Medium sieht man stets Volumen- sowie Zahlzunahme 
der Chromatophoren, Amylumkörner und Nahrungsvacuolen. 
Gleichzeitig nimmt der Körper selbst an Dicke und Breite zu, 
dagegen an Länge etwas ab, so dass er ein einigermassen abge- 
rundetes Aussehen erhält. Zu2;leich ist ausserdem Grössenzunahme 
des ganzen Körpers wahrnehmbar, wie ich dies ausschliesslich 
mit Zuckerarten bei Golpidium colpoda, Mallomonas Plosslii und 
Chilomoîias paramaecium nachgewiesen habe. 

Als eine allgemeine Regel gilt auch, dass die Vermehrungs- 
fähigkeit bei höherer Concentration stark beeinträchtigt wird. 
Unsere Versuche mit den Reinkulturen von Euglena viridis^ 
Chilomonas paramaecium und Colpidium colpoda bieten hierfür 
unzweideutige Beweise dar. Zum Vergleich führe ich einige der 
bei Schimmelpilzkulturen gewonnenen Erfahrungen an. Es che n- 
hagen^) constatirte, dass das Wachsthum einiger Schimmelarten 

^) F. Eschen h agen. /oc. cit. p. 55. 



EINIGER INFUSOEIEN. 129 

sich clurcli stärkere Concentrationen des Substrates stark ver- 
zögerte ; Klebs^) beobachtete, class das Auftreten der Konidien- 
träger und die Perithecienbiklung von Eurotium rei^ens, die 
Keimung und die Sporangienbildung von Mucor racemosus durch 
die Steigerung der Concentration des Mediums retardirt wurden. 
Gelegentlich fand ich'-) auch, dass Aspergillus niger, der in Mag- 
nesiumsulfat-Nährlösungen von verschiedener Concentration 
gezüchtet wurde, nach 4 Tagen verschiedene Grade der Ent- 
wickelung zeigte. Der Pilz wuchs in einer 59^ -Kultur am besten, 
minder gut bei 10%, während bei 20^0 und oO% nur noch eine sehr 
schwache Entwickelung zu beobachten war. Die weisse Anlage 
der Konidienfrüchte trat bei ö°/o und 10% nach 4 Tagen, bei 
20% nach 5 Tagen und bei 30% erst nach 6 Tagen ein. 

Unter den zehn von mir angewendeten Stoffen — vier organi- 
schen und sechs unorganischen Verbindungen — passten unsere 
Infusorien sich den Zuckerarten am besten an, und wieder unter 
den Zuckerarten erwies sich der Milchzucker als das beste An- 
passungsmedium. Ihm folgt der Kohrzucker in seiner Concen- 
trationshöhe, während der Traubenzucker schon in weit verdünn- 
teren Lösungen auf die Organismen schädlich einwirkt. Glycerin 
steht als Anpassungsmedium den Zuckerarten sehr nach. Unter 
den unorganischen Verbindungen, deren Einwirkung stets viel 
stärker ist als die der organischen Substanzen, ist schwefelsaures 
Magnesium zur Vermehrung der Infusorien am geeigneststen, 
während Chlorammonium für ihr Gedeihen das unpassendste 



1) G. Klebs. Die Bedingungen der Fortpflanzung bei einigen Algen und Pilzen. 
Jena 1896. pp. 446-535. 

2) A. Yasuda. Ueber den Einfluss verschiedener unorganischer Salze auf die Fort- 
pflanzungsorgane von Aspergillvs niger. The Botanical Magazine. Tokyo 1898. Vol. XU, No. 
141. p. 370. 



130 



A. YASUDA : ANPASSUNGSFAEHIGKEIÏ 



Medium ist, und unter den übrigen Chlorkalium eine mittlere 
Stellung einnimmt. 

Die verschiedenen Infusorien zeigten in Bezug auf ihre 
Anpassungsfähigkeit grosse Unterschiede, und zwar wohnte unter 
unseren Infusorien Euglena viridis die grösste Widerstandsfähig- 
keit inne, während Paramaecium caudatum die kleinste Resi- 
stenz besass. Folgende Tabelle zeigt die Grenze der Concentra- 
tionen, bis zu welcher die Infusorien am Leben blieben : 



Stoffe 


1 




Ö 

CD 




il 

0) bß 

'S S 


Salpetersaures 
Kalium 


1 


S 

.5 

ü 


Chlor natrium 


Chlorammo- 
nium 


Formeln 





0" 


ü 


d 
ü 






6 

ce 


ü 

)4 




ce 


ü 


Concentrationen der 
mit 0,1 Aeq. KNO3 
isotonisclien Lösungen') 


% 

5,40 


% 
5,13 


% 

2,70 


1,38 


1,80 


1,01 


% 
0,85 


% 
0,75 


0,59 


% 
0,54 


Maximalconcentrationen, welchen 
sich die Infnso)-ieu anpassen können 


Euglena 
viridis 


17 


15 


11 


6 


6 


2,4 


2 


2,8 


1,8 


1,4 


C'hilomonas 
paramae- 
cium 


8 


7 


6 


4 


3 


2 


1,2 


2 


1 


0,6 


Mallomonas 
riossUi 


9 


7 


6 


4 


3,4 


1,5 


2,6 


1,4 


1,5 


0,8 


Colpidium 
colpoda 


10 


8 


7 


5 


5 


2 


2 


1,6 


1,5 


1 


Paramae- 
cium 
caudatum 


8 


7 


5 


3 


2,4 


1 


1,2 


1 


1 


0,5 



1) Hugo de Vries. Eine Methode zur Analyse der Turgorkraft. Jahrb. f. wiss. Bot. 
1884. Bd. XIV. pp. 536—537. 



EINIGER INFUSOEIEN. 131 

Wie aus der vorstehenden Tabelle ersichtlich ist, muss die 
Wirkung der angewandten Substanzen auf die Infusorien nicht 
allein dem Grade ihrer Concentration zugeschrieben werden, 
dao;eo;en zeigt sich ein annäherndes Verhältniss zu den isotoni- 
sehen Concentrationen jedes Stoffes. Ich sage ausdrücklich ,, an- 
nähernd," weil die wahre Beziehung zwischen beiden durch unsere 
Versuche noch nicht sicher gestellt ist. Folgende Tabelle 
dient zum Vergleiche der isotonischen Concentrationen mit den 
entsprechenden gefundenen Werthen der maximalen Anpas- 
sungsconcentrationen^) : 



') Die bisherigen Untersuchungen ergaben in Bezug auf den Zusammenhang zwischen 
isotonischen Concentrationen und Reaktionsgrösse durchaus negative Resultate. Man ver- 
gleiche hierüber B. Stange, loc cit. Nr. 22. p. 364, und C. B. Davenport and H. V. 
Neal. loc. cit. p. 579. 



132 



A. YASUDA : ANPASSUNGSFAEHIGKEIT 



Stoffe. 


s-, 

1 

'o 


1 

S 


■1» 
ce 


a 
8 


te 

il 

1^ 




d 

SI 


1 

là 

o 


'S 

ce 

Ö 

$-( 

o 

Ü 


'S 

o 

Ü 


1 




Coucentrationeu der 

Stoffe, die mit 15% 

Eobrzucker isoto- 

uiscli sind 


15 


15,8 


7,9 


4 


5,3 


3 


2,5 


22 


1,7 


1,6 


Gefundene Werthe 
der maximalen 

Aupassuiigsconcen- 
tration 


15 


17 


11 


6 


6 


2,4 


2 


2,8 


1,8 


1,4 


II 


§ 1 
S -S 
o o 

J i 

«1 


Conceutvationen der 

Stoffe, die mit 7% 

Rohrzucker isoto- 

niscli sind 


7 


7,4 


3,7 


1,9 


2,5 


1,4 


1,2 


1 


0,8 


0,7 


Gefundene Wertlie 

der maximalen 

Ani)assuugsconcen- 

tration 


7 


8 


6 


4 


3 


2 


1,2 


2 


1 


0,6 


III 


1 i 


Concentration der 
Stoffe, die mit 7% 
Rohrzucker isoto- 
nisch sind 


7 


7,4 


3,7 


1,9 


2,5 


1,4 


1,2 


1 


0,8 


0,7 


Gefundene Werthe 

der maximalen 

Anpassungscoucen- 

tration 


7 


9 


6 


4 


3,4 


1,5 


2,6 


1,4 


1,5 


0,8 


IV 


II 


Concentvationeu der 
Stoffe, die mit 8% 
Rohrzucker isoto- 
nisch sind 


8 


8,4 


4,2 


2,2 


2,8 


1,6 


1,4 


1,2 


0,9 


0,8 


Gefundene Werthe 

der maximalen 

Anpassungsconcen- 

tration 


8 


10 


7 


5 


5 


2 


2 


1,6 


1,5 


1 


V 


i 1 

i1 
1 ^ 


Coucentrationen der 

Stoffe, die mit 7% 

Rohrzucker isoto- 

uisch sind 


7 


7,4 


3,7 


1,9 


2,5 


1,4 


1,2 


1 


0,8 


0,7 


Gefundene Werthe 

der maximalen 

Anxjassuugscoucen- 

tration 


7 


8 


5 


3 


2,4 


1 


1,2 


1 


1 


0,5 



Dieselbe Tabelle zeigt auch zugleich, class die Anj^assungs- 
grenzen unserer Infusorien an verschiedene Concentrationen im 
Allgemeinen weit niedriger sind als diejenige der niederen Algen 
und Schimmelpilze, So kann Zygnema nach Klebs^) 509^ Rohr- 

*) G. Ivlebs. Beitrüge zur Physiologie der Pflanzenzelle. Eericlile der deutsch bot. 
Gesellsch. 1887. Bd. V. p. 187. 



EINIGEE INFUSORIEX. 133 

zucker und 20¥o Glycerin vertragen. Dass dieselbe Alge sich 
auch einer 69oigen Chlornatriumlösung anpassen kann, ist von 
Richter^) erwiesen worden. Was die Schimmelpilze anbelangt, 
so zeigen sie ebenfalls eine weitaus grössere Widerstandsßihigkeit 
gegen starke Concentrationen. Ich gebe hier zum Vergleiche die 
von Eschenhagen'") erhaltenen Ergebnisse wieder. 



Traubenzucker. (Tlvcerin. ^'>^atHuar^^ Chlornatrium. 

Aspergillus nicjer 53o/o 439^ 2lo/o \lo/o 

Pénicillium glaucum 55 ,, 43 ,, 21 „ 18 „ 

Botrytis cinerea 51 ,, 37 „ 16 „ 12 „ 

Unsere Infusorien zeigen gegen dieselben Stoffe folgendes Ver- 
halten : 



Traubenzucker. 
Euglena virivlis 11 9^ 

Colpidium colpoda 7 ,, 

Malloraonas Plosslii 6 ,, 

Chiloraonas j)aramaeciu7n 6 ,, 
Paramaecium caudatura 5 „ 

Daraus geht hervor, dass die Resistenzkraft unserer Infusorien 
gegen die augewendeten Stoffe hinter derjenigen der Schimmel- 
pilze weit zurücksteht. 



Glycerin. 




leitisiiures 

atrium. 


Clilornatriun 


6 


¥o 




2 0/, 


1,89^. 


5 


}j 






1,5 „ 


4 


/", 




2,6 „ 


1,5 „ 


4 


,) 




1,2 „ 


1 „ 


3 


,> 




1,2 „ 


1 „ 



1) A. Richter, loc. cit. p. 24. 

-) F. Eschenliagen. loc. cit. p. 55- 



134 A. YASUDA : ANPASSUNGSFAEHIGKEIT 



Zusammenfassung. 

(1) Isotonische Lösungen der in Rede stehenden Substanzen 
üben auf die von mir geprüften Infusorien nur eine ,, annä- 
hernd" gleichartige Wirkung aus. 

(2) Die Grenzen der Concentration, welcher sich diese 
Infusorien unter gewöhnlichen Verhältnissen anpassen können, 
liegen im Allgemeinen weit niedriger als die der niederen Algen 
und Schimmelpilze ; selbst das widerstandsfähigste darunter, 
Ewjlena viridis, vermag nur verhältnissmässig schwache Concen- 
trationen zu ertragen. 

(3) AVenn die Organismen plötzlich in Lösungen höherer 
Coucentrationen gebracht werden, so treten erst an der Cuticular- 
oberfläche ihrer Körper longitudinale Falten auf, aber während 
ihre Anpassung an das neue Medium stattfindet, dehnen sich die 
Falten allmählich aus, bis sie zuletzt gänzlich verschwinden. 

(4) Die höhere Concentration des Mediums verlangsamt die 
Vermehrung der Infusorien. 

(5) Durch Steigerung der Concentration des Mediums wird 
die Bewegung der Organismen vielfach retardirt. 

(6) Bei Zuckerlösungen stärkerer Concentration vergrössern 
sich die Körper der Infusorien bis zu einem gewissen Grade. 

(7) Die Vacuolen, Chromatophoren oder Amylumkörner 
nehmen in dem Masse an Grösse zu, al« die Mediumsconcentra- 
tion steigt. 

(8) Je mehr die Concentration des Mediums zunimmt, desto 
mehr runden sich die Körper der Organismen ab, und die Kör- 
perumrisse werden uneben. 

(9) Wenn das Maximum für die Accommodation ein nie- 



EINIGER INFUSORIEN. 



135 



driges ist, so finden die Veränderungen der Körper der Infusorien 
schon bei niederen Concentrationen des Mediums statt. 

(10) Wenn sich die Concentration des Mediums dem Maxi- 
mumpunkt nähert, so verschmelzen die in den Körpern der 
Organismen befindlichen Chromatophoren oder Amylumkörner 
mehr oder weniger mit einander. 

Tokyo, 30. November 1898. 



Inhaltsübersicht. 



Einleitung Seite 101. 

ms. 

107. 
112. 
127. 
134. 



Litteratur 

Methodisches 

Beschreibung der Versuche 

Allgemeines und Schlussbemerkungen 
Zusammenfassung 



136 A. YASüDA : ANPASSUNGSFAÈHIGKEIT 

Erklärung der Figuren. 

Sämmtliche Figuren wurden nach den lebendigen Tliieren in 
den unreinen Kulturen unmittelbar skizzirt, weil ihre Gestalten 
bei den getödteten Individuen sich mehr oder weniger veränderten. 

TAFEL X. 

F\'^. 1-46, Cltilomonas paramaecium Ehrbg. Vergr, 420. 
Fig. 1. Individuen aus einer Nährlösung mit 1 ^ Milchzucker. 
•2 

}> •^- )} V }) }) 

A 

)) ^* )} >> }) }} 

5) •^- ;> }J )) ?) 

jj ^- j) 5) '> >; 

>) ' • ;? J> }•, »5 

q 

j' ■ • jj )! J7 j; 

12 
14 

}> ■^'^- )J }} ,v J! 

>J -'■'-*• >î >) ,•> )! 

)} •*■ ' • j) >> >) ;> 

1 S 
10 
20 
21 
22 

') ^-'' 77 77 77 77 

77 "^"^ 77 77 77 77 

24 



^ 


77 




77 


3 










77 




77 


4 


7 7 




77 


5 


77 




7 7 


6 


77 




77 


7 






77 


8 


77 




77 


1 


% 


Rohrzucker. 


•7 








■" 


77 




7 7 


3 


7' 




7; 


4 


77 




77 


5 


77 




77 


6 










77 




77 


_ 








/ 


7 7 




77 


1 


9^^ 


T 


raubenzucker. 


2 










77 




77 


3 










77 




77 


4 


77 




77 


.5 


77 




77 


6 


77 




77 


1 


¥o 


G] 


ycerin. 


2 


77 




7) 



EINIGER INFUSORIEN. 137 

Fig. 25. Individuen aus einer Nährlösung mit 4 o/g Glycerin. 
„ 26. „ „ ,, „ „ 0,5^ Schwefelsaures 

Magnesium. 
97 

29 

31 

•^9 
" '^^^ j» ,, }} >) 

!i '^^' )} J, !) )) 

,J ^ • • ,, ;? jj ,? 

00 

40 
41 
42 
43 

44 

4^) 
46 

TAFEL XI. 

Fig. 1-11. Euglena viridis Ehrbg. Vergr. 420. 
Fig. 1. Individuum aus einer Nährlösung mit 1 9^ Traubenzucker. 



J, 


1—1 




,' 


1,5 „ 




J, 


9 




V 


•2,5 „ 




V 


3 „ 




j) 


0,59^ 


Salpetersaures 

Kalium 


}) 


1 „ 


,, 


;? 


1,5. 


,, 


V 


9 


,, 


') 


0.69^ 


Salpetersaures 

Natrium, 


>> 


1,2 „ 


,; 


)) 


0,DO/a 


Chlorkalium. 


;, 


1 „ 


,, 




1,5 „ 
9 


,, 


;? 


0,5«/^ 


Chlornatrium. 


,, 


1 „ 


„ 


,, 


0,2«/. 


Chlorammonium . 


J> 


0,4 „ 


,, 


>5 


0,6 „ 


„ 



2 2 

3 3 

4 4 



138 A. YASUDA : ANPAÖSUNGSFAEHIGKEIT 

Fig. 5. Individuum aus einer Nährlösung mit 5 «^ Traubenzucker. 

,, D- j) V Ti r V ^ JJ ?' 

,, 7. ,, ,, V >) ?) ' }j j> 

„ 8. ,, „ ,, ., ;> ^ J5 JÎ 
Q 9 

,, •-'• M JJ )5 '^ ?' " " 

„ 10. „ „ ,, J> ,V i'-' 5) V 

,, 11. 5> )> 5J >> " ^^ " " 

Fig. 12-41. Oolpidmm colpoda Ehr hg. Vergr. 420. 

Fio- 12 Individuum aus einer Nährlösung mit 1 o/g Milchzucker. 

o* 

)j 14. j, jj j7 jj 

>; !'• ?) )) 3' '' 

)) J-^' >3 M JJ '• 

jj J-'J» )J )) 5! >> 

„ -0. ,, jj jj }; 

94 

JJ ■^^' }> 3) 33 3' 

53 wO. ,, ,, 33 33 

), ^t). ,, ,, 5J )} 

)3 -'•• '3 '3 33 33 

9 8 

)3 -"^' 33 33 33 33 

33 ~''^' 33 3' 33 33 

33 'JO. ,, ,, ,> 1} 

33 "^l- 33 33 33 33 

•^9 

33 '^-" 35 J3 33 33 

33 '''^' 33 33 33 33 

i> 9.D. ,, ,5 )) )5 



2 




33 


3 




33 


4 




33 


5 




J3 


6 




33 


7 




33 


8 




33 


9 




33 


10 




33 


1 


9^. 


Rohrzucker. 


2 




33 


3 




'3 


4 




33 


5 




33 


6 




3» 


7 








33 


8 




33 


1 


y. 


Traubenzucker, 


2 




33 


3 




33 


4 




33 


.5 




33 


6 




33 



EINIGER INFUSORIEN. 139 

Fig. .36. Individuiira ans einer Nährlösung mit 7 o/^ Traubenzucker. 

„ 37. „ ,, „ „ „ 1 9^ Glycerin. 

QO 9 

}i »JU. ,, ,, ,, ,, „ ~l ,, ,, 

.39 3 

40 4 



TAFEL XII. 

Fig. 1-21. 3'kd/omonas Plosslii Perty. Vergr. 420. 
Fig. 1. Individuen aus einer Nährlösung mit 1 o/^ Milchzucker. 
9 
3 
4 
5 

V -'• >} )) }> )7 

7 
9 

;j ''• )} 55 55 55 

55 ^^' 55 55 55 55 

55 ^^' 55 55 55 J5 

55 -■■ — • 55 55 55 55 

13 

14 

55 -••^* 55 55 55 55 

55 -^"* 55 5> 55 55 

55 ^^' 55 55 55 55 

17 

55 -^ ' • 55 55 55 55 

55 -'-"• 55 55 55 55 

19 

55 •^^' 55 55 55 55 

90 

55 '^^- 55 55 55 55 

21 



2 


55 


55 


3 


55 


55 


4 


55 


55 


5 


55 


55 


6 


55 


55 


7 


55 


55 


8 


55 


5J 


9 


55 


55 


1 


fo 


Rohrzucker. 


2 


55 


55 


3 


55 


55 


4 


55 


55 


/) 


55 


55 


6 


55 


55 


7 


55 


55 


1 


fo 


Traubenzucker, 


2 


55 


55 


3 


55 


55 


4 


55 


55 


5 


5» 


JJ 



140 



Fig. 22-28. Faramaecîum caudatum Ehrbg. Verg. 240. 
Fig. 22. Individuum aus einer Nährlösung mit 1 0/^ Rohrzucker. 
23. 
24. 
2.5. 
26. 
27. 
28. 



2 




3 




4 




5 




G 




7 





Jour. S C.Co II. Vol. XIII. PI.X, 



13 



14^ 



15 



10 



lô 



f i 

5 



11 



11 



,ÇN 



J2 



la 



19 



20 



21 



22 



23 



24^ 



2& 



26 



27 



23 



29 



30 



31 



32 



33 



34 



35 



36 



3 7 



38 



39 



40 



41 



42 



43 



45 



46 



Jour. Sc. Coll. Vol. XIII. PI. XI. 








J2 



14^ 




13 



16 






to&-^£<; 








J«S 



JÔ 



21 



23 



es 



.:^\ 



"^^^^^ 




24 



2 6 



28 



2P 




30 








32 




32 



/^K 




34^ 



^^B^ 




35 



TAI 



36 






37 




3Ô 



^: 






Jour. Sc. Coll. Vol. XIII. PI. XII 



W 



^ 



/^ 



im 



12 



16 



^ 



14 



^ fb,. 



17 



& 






IS 




V 



20 



21 



22 



24 



1 



2 7 



28 



Ueber die Wachsthumsbeschleunigung einiger 
Algen und Pilze durch chemische Reize. 



VON 



N. Öiio, Rigahishi. 



Hierzu Tafel XIII. 



I. Einleitung und Litteratur. 

In seiner Arbeit : Etudes chimiques sur la végétation^), hat 
Kaulin schon im Jahre 1869 darauf aufmerksam gemacht, dass 
Zink- und Siliciurasalze in geeigneter Dosis das Wachsthum von 
Aspergillus niger befördern. Auf dieser an sich richtigen Beo- 
bachtung fussend war der genannte Autor mit Unrecht der 
Ansicht geneigt, dass diese Substanzen zur normalen Entwickelung 
unseres Pilzes nothwendig seien, indem er diese unter ,, les 
éléments chimiques essentiels" rechnet und eine Nährlösung von 
recht complicirter Zusammensetzung für Pilze vorschreibt. 

Ueber die Mineralstoffbedürfnisse der Pilze wurden seither 
von einigen Forschern Untersuchungen gemacht, von denen die 
Arbeit Naegeli's^) in der ersten Linie zu nennen ist. In 



1) Ann. d. Sc. nat. Bot., Ser. V, T. XL, 1869, S. 91. 

2) V. Naegeli, Der Ernähi'ungschemismus der niederen Pilze. Sitzungsberichte d. Kgl. 
Bayr. Akad. d. Wiss. Math.-phys. Cl. 1880. 



142 N. ONO : WACHSTHUMSBESCHLEUNIGUNG 

neuerer Zeit wurde dies Thema von Molisch^), sowie auch von 
Benecke") wiederaufgenommen. Der Mühe dieser Autoren 
verdanken wir unsere heutigen Kenntnisse in dieser Kichtung. 
Nach den übereinstimmenden Angaben der genannten Autoren 
stellt weder Zink noch Silicium einen eigentlichen Nährstoff dar 
und kann wohl von Kulturflüssigkeit ausgeschlossen werden. 

Dass die wachsthumsbeschleunigende Wirkung gewisser 
Metallradikale auf einer chemischen Keizung beruht, wurde von 
Pfeffer^) in 'seiner im Jahre 1895 erschienenen Arbeit zum 
ersten Male klar gestellt und später in allgemeinen Zügen in der 
S**^" Auflage seiner Pflanzenphysiologie erörtert. Er bemerkt in 
dem letztgenannten Werke, dass anscheinend geringfügige Um- 
stände in der That nicht selten einen erheblichen Einfluss auf 
Gedeihen und Wachsen haben, und sagt : ,,Vermuthlich handelt 
es sich in dieser beschleunigenden Reizwirkung um eine der 
mannigfachen Reaktionen, die darauf abzielen, durch intensitivere 
Thätigkeit einen benachtheiligten Einfluss thunlichst entgegen- 
zuarbeiten oder Schädigungen auszugleichen/'^) Im vorhergegan- 
genen Jahre wurde eine Reihe Versuche von Richards^"*) angestellt, 
deren Resultate die Ansicht Pfeffer's bestätigen. Er zog zu seinen 
Untersuchungen verschiedene Schwerenmetallsalze wie Zink-, 
Kobalt-, Nickel-, Eisen-, und Mangansalze und einige andere 
giftige Substanzen heran und stellte die Thatsache fest, dass fast 



1) H. Molisch, Die mineralische Nahrung der niederen Pilze. Sitzungsberichte d. 
Wiener Akad., Oct. 1894. 

2) W. Benecke, Die zur Ernährung der Schimmelpilze nothwendigen Metalle. Pringsh. 
Jahrb. f. wiss. Bot. Bd. XXYIII, 1895, S. 487. 

3) Pfefl'er, Election organischer Nährstoffe. Pringsh. Jahrb. f. wiss. Bot. Bd. XXVIII, 
1895. 

4) Pflanzenphysiologie. 2" Aufl. Bd. I. S. 374. 

5) H. M. Kichards, Die Beeinflussung des Wachsthums einiger Pilze durch chemische 
Reize. Pringsh. Jahrb. f. wiss. Bot. Bd. XXX, 1897, S. 665. 



EINIGER ALGEN UND PILZE. 143 

alle geprüften Substanzen mehr oder weniger die Pilzernte zu 
vermehren vermochten. 

Auch anderweitige Beispiele für Erhöhung der Lebens- 
thiitigkeiten durch geringe Zusätze giftiger Stoffe findet man in 
der Litteratur. So versuchte Schulz^) zu zeigen, dass die durch 
Saccharomyces verursachte alkoholische Gährung bei Gegenwart 
von geringeren Quantitäten gewöhnlich als Hefegifte sich verhal- 
tender Substanzen wie Sublimat, Jod, Salicylsäure, Brom, Arse- 
nige Säure u. a., auf längere oder kürzere Zeit wesentlich gehoben 
werden konnte. Diese Tliatsache aber war bereits gewissen 
Gewerben nicht unbekannt gewesen, dass Zuführung von sonst 
gährungshemmenden Stoffe, wie z. B, Kupfervitriol oder Sali- 
cylsäure, unter Umständen die Hefe zu energischer Thätigkeit 
veranlasse. Derselbe Autor hat früher eine ähnliche Erhebung 
der Lebensäusserungen bei dem thierischen Organismus beobachtet 
und folgenden Satz ausgesprochen, dass ,, Jeder Reiz auf eine 
einzelne Zelle sowohl wie auch auf aus Zellengruppen bestehenden 
Organen entweder eine Vermehrung oder eine Verminderung 
ihrer physiologischen Leistungen bedinge, entsprechend der 
geringeren oder grösseren Intensität des Beizes""). 

Derartigen Verallgemeinerungen begegnen wir auch bei 
Hueppe''). Hauptsächlich auf die auf bakteriologischem Gebiete 
constatirten Thatsache sich stützend verkündigt er die Erschein- 
ung als den Ausdruck des allgemein giltigen Gesetzes für die 
Wirkung von Chemikalien auf Protoplasma. Er bezeichnet dieses 
als das ,, Biologische Grundgesetz," welches er in folgendem 



1) H. Schulz, Ueber Hefegifte. Pflüger's Archiv f. Physiologie. Bd. 42., 1888. S. 517. 

2) H. Schulz, Zur Lehre von der Arzneiwirkung. Virchow's Ai'chiv. Bd. 108, 1877. 
S. 427. 

3) F. Hueppe, Naturwissenschaftliche Einführung in die Bakteriologie. 1896, Wies- 
baden. S. 55. 



144 N. ONO : WACHSTHUMSBESCHLEUNIGUNG 

Satze formulirt : ,, Jeder Körper, der in bestimmter Concentration 
Protoplasma tötet und vernichtet, in geringeren Mengen die Ent- 
wickelungsfähigkeit aufhebt, aber in noch geringeren Mengen, 
jenseits eines Indifferenzpunktes, umgekehrt als Reiz wirkt und 
die Lebenseigenschaften erhöht." 

Zu erwähnen ist noch, dass auch bei höheren Pflanzen die- 
selbe oder eine wenigstens sehr nahe verwandte Erscheinung 
vorkommt. Bekanntlich pflegt man seit einigen Jahren die 
Weinreben mit Kupferpräparaten, der sogenannten Bordeaux- 
brühe, zur Bekämpfung der Pilzkrankheit zu bespritzen. Eine 
derartige Behandlung ruft ausser der indirekten Einwirkung, die 
Schädigung durch parasitische Pilze herabzusetzen, auch eine Schar 
auffallender Erscheinungen seitens der bespritzten Pflanze hervor, 
die mit Kumm^) vielmehr als direkte Wirkung der angewandten 
Chemikalien auf den Pflanzenorganismus selbst zu bezeichnen 
sind. Solche sind die Steigerung der Chlorophyllbildung und 
daraus resultirende vermehrte Stärkeproduktion, reichlicherer 
Traubenansatz, Beschleunigung der Peifung u. a. Pumm ist der 
Ansicht, dass die Steigerung der Chlorophyllbildung einem 
chemischen Peiz zuzuschreiben sei. Im darauf folgenden Jahre 
führten Frank und Krüger^) einige Bespritzungsversuche an 
Kartoffeln aus, wobei sie auch eine ähnliche Thatsache consta- 
tirten. Ueber das eigentliche Wesen der Wirksamkeit jener 
Stoffe ist vorläufig nichts weiteres zu sagen. 

Wie aus der oben angeführten Skizze ersichtlich ist, bezogen 



1) Rumm, Ueber die Wirkung der Kupferpräparate bei Bekämpfung der sogenannten 
Blattkrankheit der Weinrebe. Ber. d. deutsch. Bot. Ges. Bd. XI. 1893. S. 709. 

Rumm, Zur Frage nach der Wirkung der Kupfer-Kalksalze bei Bekämpfung der 
Peronospora villcola. ebenda S. 445. 

2) Frank u. Krüger, Ueber den Reiz, welchen die Behandlung mit Kupfer auf die 
Kartofiël hervorruft. Ber. d. deutsch. Bot. Ges. Bd. XII, 1894. 



EINIGER ALGEN UND PILZE. 145 

sich, wenn wir von der zuletzt besprochenen Thatsache absehen, alle 
bisherigen diesbezüglichen Versuche mit pflanzlichen Organism- 
en ausschliesslich auf chlorophyllose niedere Organismen der 
Pilze. Wenn die erwähnte Erscheinung, wie vielerseits behauptet 
wird, allgemeine Geltung haben würde, so sollte es a priori zu 
erwarten sein, dass auch chlorophyllhaltige niedere Organismen 
in gleichem Sinne reagiren. Dies aber benötigt einer experimen- 
tellen Bestätigung. 

Herbeigezogen wurden zu meiner Untersuchung verschiedene 
Schwerenmetallsalze, wie Zink-, Nickel-, Kobaltsulfat u. a., welche 
auf unsere Versuchsaigen gewisse Reizung auszuüben vermochten. 
Ferner wurden eine Reihe Parallelversuche mit Pilzen angestellt, 
deren Ergebnisse die oben genannten Richards' sehe Unter- 
suchung bestätigt und erweitert. 

Die vorliegende Arbeit wurde auf Veranlassung und unter 
Leitung von Herrn Prof. Dr. Miyoshi im Botanischen Institut 
der Kaiserlichen Universität zu Tokyo während des Zeitraums 
von August 1898 bis Juni 1899 ausgeführt. 

Es ist mir eine augenehme Pflicht, meinem hochverehrten 
Lehrer für die vielseitige Anregung meinen verbindlichsten Dank 
auszusprechen. 

Herrn Prof. Dr. Matsumura sage ich an dieser Stelle auch 
meinen besten Dank für die Belehrung und das Interesse, welches 
er meiner Arbeit entgegengebracht hat. 



II. Methodisches. 

Bei unseren Versuchen kommen stets die Reinkulturen in 
Betracht. Als Kulturgefösse wurden Erlen meyer'sche Kolben 



146 N. ONO : WACHSTHUMSBESCHLEUNIGUNG 

von ca 200 ce Inhalt angewendet, und zwar von gleicher Gestalt 
und Qualität in einer Reihe von Parallelversuchen benutzt. Nach- 
dem die Gefässe zuerst mit Salzsäure gründlich gewaschen worden 
waren, wurden sie mit Leitungswasser, dann mit destillirtem 
Wasser wiederholt ausgewaschen, getrocknet und gebraucht. 

Wasser. — Zur Zubereitung der Nährlösungen und zum Auf- 
lösen der Reizmittel benutzte ich doppeltdestillirtes Wasser. 
Das auf übliche Weise gewonnene destillirte Wasser war von 
Glas zu Glas nochmals destillirt worden. 

Chemische Präparate. — Die als Nährstoffe sowohl als 
auch als Reizstoffe dienenden Chemikalien stammten grössten- 
theils aus Merck's ,,garantirt reinen" Reagentien. 

Nährlösungen. — Für Algen bediente ich mich der bekann- 
ten Knop'schen Lösung^), die ich nach folgender Vorschrift 
bereitete : 

(A) MgSO^ + TH.O 10.25g (B) Ca(NO;3)o 20.00g 

KNO3 5.00,, 

KH.PO, 5.00,, 
AVasser 175.00cc Wasser 175.00 cc 

Beim Gebrauch wurden je 10 cc von A und B mit 880 cc 
Wasser verdünnt. Diese bezeichne ich als Original-Nährlösung 
für Algen. 

Die Original-Lösungen für Pilzkulturen bestanden aus fol- 
genden drei Serien : 



1) Aus keinem besonderen Grunde benutzte ich liier Ca-lialtige Külirlösung. Die Ent- 
behrlichkeit der genannten Metalle bei Pilzen und niederen Algen ist bekanntlich in 
neuerer Zeit von Moliscli uiad Beuecke erwiesen worden. 



EINIGER ALGEN UND PILZE. 147 

(A) 



kh;po4 


0.50 g 


MgSO, 


0.25 „ 


NH4NO3 


1.00 „ 


Eisen 


Spuren 


Rohrzucker 


5.00 g 



Wasser 90.00 CG. 

(B) 
Wie bei A. Asparagia 0.5g statt NH^NOa 1.0 g 

(C) 

Wie bei A. Dextrose anstatt Rohrzucker. 

Bei fast allen Kulturreihen wurde A angewendet, während 
B und C nur ausnahmsweise benutzt wurden. 

Kulturaustellung. — Ichgossin 5Kolbenje 135 cc Original- 
Lösung (Knop'sche bezw. Rohrzuckernährlösung). Sodann setzte 
ich zum ersten Kolben 15 cc destillirtes Wasser hinzu und liess 
dies als Nährlösung für Controlkulturen dienen, zum 2^'^'', 3*™... 
5^'"'' je 15 cc betreffend verdünnte Lösung von Reizstoffen, deren 
Wirkungen versucht werden sollten. Bei den Algenkulturen 
geschah die Verdünnung in absteigender geometrischer Reihe im 
Verhältniss 1:5, bei den Pilzkultureu aber mit 1:2. Alle Kolben 
wurden darauf gut geschüttelt, um die Lösungen aufs innigste 
zu mischen, und dann die in jedem Kolben enthaltene Lösung in 
drei Kolben s-leicbmässio; vertheilt, so dass wir 3 Serien von 
je 5 Kolben, deren jede 50 cc Nährflüssigkeit enthielt, vor 
uns haben. 



148 N. ÖNO : WACHSTHUMSBESCHLEUNIGUNG 

Die auf diese Weise zubereitete Kulturflüssigkeit enthielt 
bei der Knop'sche Lösung ca 2.5X wasserfreie Salze und bei der 
Pilznährlösung etwa 5^ Rohrzucker^). Dann folgte bei Pilzkul- 
turen die Sterilisation in einem Koch'schen Dampftopf, welche 
Y2 — 1 Stunde dauerte. 

Bei Pilzen fand die Impfung in üblicher Weise statt, wäh- 
rend ich sie bei Algen in der Weise ausführte, dass ich mittelst 
Platindraht oder Pipette eine möglichst kleine Algenmenge aus 
den zuvor in Nährlösung von derselben Concentration oder auf 
Agar bereiteten Reinkulturen herausnahm und in Versuchsgefässe 
brachte. 

Die Kulturen wurden in Zimmertemperatur (ca 15° C im 
Mittel) ausgeführt, und in kälteren Jahreszeiten ins Treibhaus 
(16-21° C) gebracht. 

Die Kulturdauer variirte unter Umständen zumeist zwischen 
8 und 25 Tagen bei Pilzen und etwa einem Monatlang bei Algen. 

Bestimmung des Trockengewichtes. — Für die Beurthei- 
lung des Gedeihens hat fast stets die Ermittelung des Trockenge- 
wichtes der gebildeten Algen- bezw. Pilzmassen Aufschluss 
gegeben. Die Bestimmung wurde folgendermassen ausgeführt. 
Nach Beendigung der Versuche wurde die Kulturflüssigkeit mit der 
Erntemasse insgesammt durch vorher einzeln gewogene Filter 
filtrirt. Dabei befreitete ich den an der Glaswand haftenden 
Theil mittelst eines mit einem Kautschuk-Hut versehenen Glas- 
stäbchens. Dann spülte ich die Erntemasse mit kaltem destillirtem 
Wasser, um dadurch etwa noch vorhandene Nährflüssigkeit 
möglichst zu entfernen, und wenn sie ziemlich lufttrocken ge- 
worden war, trocknete ich sie im Paraflinofen bei 100° C und 
wog sie nach dem Erkalten. Das auf diese Weise ermittelte 



1) Diese Lösung wurde von PfefTcr und Bichards vielfach benutzt. 



EINIGER ALGEN UND PILZE. 149 

Trockengewicht der Ernte ist in den angeführten Tabellen als 
Ernteertrag notirt. 

Bei den Algenkulturen geschah es vielfach, zumal bei 
denjenigen, welche während der kälteren Jahreszeit angestellt 
worden waren, dass die Vermehrung nur sehr langsam vor sich 
ging, und dass nach monatelangem Stehenbleiben eine nur 
schwache Entwickelang sich zeigte. In solchen Fällen musste 
ich mich damit begnügen, durch das Aussehen der Kulturen die 
Stärke der Entwickelung zu beurtheilen. 

Was nun die specielle Ausführungsmethode anbelangt, so 
wird sie an geeigneten Stellen berücksichtigt. 



III. Vorbemerkungen über Versuchsobjekte. 

Für Pilzkulturen bediente ich mich der gewöhnlichen 
Schimmelpilze Aspergillus niger und Pénicillium glaucum. 

Benutzt wurden bei meinen Algenversuchen die folgenden 
Formen : 

Protococcus sp. 
Chroococcum sp. 
Hormidium nitens. 
Stigeoclonium sp. 

Da wir zur Zeit über die Lebensbedingungen der Algen 
überhaupt nur wenig wissen, so war es mir nicht immer ge- 
lungen, die im Freien rasch wachsenden Algenarten im Labora- 
torium unbeschädigt gedeihen zu lassen. Besonders schwierig war 
die Aufgabe, die grösseren Formen in reinem Zustande längere 
Zeit in einer bestimmten Nährflüssigkeit zu kultiviren. 

Nach einigen darauf bezüglichen Vorversuchen kam ich 



150 N. ONO : WACHSTHUMSBESCHLEUNIGUNG 

schliesslich auf die oben genannteu niederen Formen zurück, die 
ziemlich leicht rein zu erhalten waren, ausserdem sicheres Gedeihen 
zeigten und ferner leichtere Kontrolle der Impfmasse gestatteten. 

Unsere kleineren Algen waren zumeist in den für grössere 
Algen bezweckten Kulturgefässen spontan aufgetreten, und so 
wurden diese durch wiederholte Uebertragung rein gezüchtet. Die 
ersteren Algen Hessen sich auch auf festem Nährboden, welcher 
aus 72 Proc. Agar und 2.5 Promille Nährsalz enthaltender 
Gallerte bestand, gut vegetiren und solcher Plattenkultur bediente 
ich mich bei einigen Beimpfungen. 

Hormidium nitens, welches sich auf der Oberfläche einer 
Vaucheria-K\i\t\xv als eine charakteristische seidenglänzende 
Decke bildete, zerfiel nach dem Uebertragen in unsere Nähr- 
lösung in einzelne Zellen und vegetirte als solche weiter. 

Es bleibt noch zu bemerken übrig, dass die Vermehrung 
bei niederen Algen {Protococcus wurde zunächst untersucht) 
während der kälteren Jahreszeit so gut wie vollständig herab- 
gesetzt worden ist, wenn auch die Kulturen im Treibhaus bei 
16-20° C sich befanden. Die Ursache wolle man nicht in man- 
gelndem Licht suchen, da dieselben Kulturen mit noch massigerem 
Lichtgenuss vor einem gegen Norden gerichteten Fenster eine 
gute Entwicklung bis Anfang November zeigten. Ob es sich 
hier etwa um eine Vegetationsperiodicität handelt, beabsichtige ich 
im kommenden Jahre näher zu untersuchen. 

IV. Die Veränderungen in der Wachsthumsweise 

und die Correlation zwischen Fortpflanzung 

und Wachsthum. 

Wie wird die Wachsthumsweise einer Pflanze beeinflusst 
werden, wenn ihr Wachsthum bei Gegenwart von Eeizstoffen 



EINIGER ALGEN UND PILZE. 151 

Über die Norm hinaus gesteigert wird ? Um diese Frage zu 
beantworten, wurden bei meinen Untersuchungen einige Beobach- 
tungen gemacht, um dabei auftretende Wachsthumsmodificationen 
zu kennzeichnen. 

Bei von mir untersuchten Algen konnte ich keine bemer- 
kenswerthe Veränderung der Wachsthumsweise beobachten. Die 
Zellengrösse blieb unverändert ; so lag z. B. bei Prolococcus sp. 
die Zellengrösse jedenfalls zwischen 7-10 ,«. Sie zeigte ferner 
keinen Unterschied in Amylumeinschlüssen. 

Bei Pilzkulturen liess sich aber die Veränderung in der 
Wachsthumsweise mehr oder weniger schon makroskopisch er- 
kennen. So war bei den meisten versetzten Kulturen die 
Beschaffenheit des Mycels ungewöhnlich. Während bei Kontroll- 
kulturen die Hyphe in Nährlösungen durchsichtig und zart 
waren, bildeten diejenigen der versetzten Kulturen ein dickes, 
weisses, hautartiges Geflecht, -welches bei längerem Stehen sich zu 
einer aufgerollten Masse umgestaltete. 

Auch das Turgorverhältnis in den versetzten und den nicht 
versetzten Kulturen wurde vielfach studirt, um zu ersehen, ob 
hier etwa ein nennenswerther Unterschied zwischen beiden vor- 
handen ist. Meine Versuche ergaben in diesem Punkte kein 
positives Besultat. 

Viel auffallender sind die Beziehungen zwischen Fortpflan- 
zung und AVachsthum. 

Bekanntlich stellen das Wachsthum und die Fortpflanzung 
zwei miteinander in engster Wechselbeziehung stehende Lebens- 
thätigkeiten dar. So kommt es nicht selten vor, dass bei einer 
Pflanze, die in üppigen^ Wachsthum begriffen ist, ihre Fortpflan- 
zungsfähigkeit zeitweilig suspendirt wird ; und wenn hingegen die 



152 N. ONO : WACHSTHUMSBESCHLEUNIGUNG 

Fortpflanzung herabgesetzt worden ist, so schreitet das Wachsthum 
kräftig fort. 

Es war mir daher nicht ohne Interesse, diese Verhältnisse 
in unserem Falle kenneu zu lernen. Ich konnte jedoch das 
Studium nur bei Pilzen ausführen, nicht aber bei Algen, da 
meine Versuchsalgen hauptsächlich einzellige Formen waren, die 
nur durch Theilung sich vermehrten. 

Bei Pilzen hingegen war die Wechselbeziehung zwischen der 
Mycelentwickelung und der Sporenbildung deutlich zu erkennen. 

In fast allen Fällen übten die Versuchsstoffe auf die Pilze 
Sporen- bezw. Conidienbildung verzögernden Einfluss aus. Be- 
sonders ausgeprägt trat dies bei Zusatz von ZnS04 und NaFl 
ein. Ich konnte vielfach constatiren, dass bei normalen, in 
Zimmertemperatur (ca 15° C im Mittel) gezüchteten Äspergilliis- 
Kulturen das Mycelium schon nach 1-2 Tagen sich ausbreitete 
und mit angelegten Sporangienträgern versehen war, deren Köpfe 
nach weiteren 2 Tagen durch Reifung der Sporen ganz ge- 
schwärzt worden waren. Bei den Kulturen mit Zusatz von 0.005 
Proc. ZnSO, daaceo-en habe ich selbst nach Verlauf einer Woche 
vergeblich nach reifen Sporangien gesucht. In den letzteren 
Fällen fand ich nur auf dem ziemlich stark angewachsenen 
häutigen Mycel etwas Sporangienträgeranlage mit etwas aus- 
geschwollenen Köpfchen, nebst einer Anzahl von wohl als Luft- 
mycel anzusehenden Gebilden. Erst nach weiterem einwöchigen 
Stehen wurden sie mit bräunlich-schwarzen Sporen besetzt. 

Die Grösse der Sporen, welche gegen verschiedener Einflüsse 
sehr empfindlich ist, blieb in unseren Fällen unverändert. Die 
Beschaflenheit des Myceliums aber war nicht unbedeutend 
beeinflusst. 

Wie erwähnt, fand ich in allen von mir untersuchten Stoßen 



EINIGER ALGEN UND PILZE. 153 

mehr oder minder die Tendenz, die Sporenbildunji; zu verzögern 
oder wenigstens zu verspäten. Dies gilt ohne weiteres auch für 
Pénicillium sowohl, wie für Aspergillus. 

Am ausgeprägtesten und am schönsten aber konnte ich 
dieses Verhältnis bei Aspergillus-KwMwY mit Zusatz von NaFl 
kennzeichnen. 

Ich nehme hier aus meinem Protokolle folgendes Beispiel : — 

I. Normal Ganze Oberfläche mit schwarzen Sporen bedeckt, 

II. 0.00209^ NaFl Etwa Vs der Decke mit Sporen bedeckt, die übrigen 
"/a steril. 

III. 0.005 ,, „ Nur sehr spärliche Sporenbildung, weiss. 

IV. 0.0 10 „ „ Steril, hautbildend, weiss. 
V. 0.021 ,, „ Ganz steril,, weiss. 

(Beobachtet nach 10 Tagen seit der Aussaatzeit der Sporen,) 

Alle NaFl enthaltenden Kulturen zeigten eine üppige vege- 
tative Entwickelung des Myceliums und bildeten hautartige 
Decken. 

Nach weiteren zehn Tagen waren die Kulturen wesentlich 
unverändert im Aussehen, Bei III. sporadisches Auftreten der 
Sporangienträger mit unreifen Sporen, bei lY. Sterilbleiben 
der Decke mit einigen haarigen Luftmycelen, bei V immer steril. 
Die Trockengewichtbestimmung nach der Beendigung der 
Kultur zeigte folgendes Kesultat : 

I. IL III. IV. V, 

0.314g 0.336g 0.385g 0.3l6g 0.274g 

(Für näheres vgl. Tabelle Pilze H) 

Eine Photographie am Ende dieser Arbeit veranschaulicht 
das eben gesagte Verhältnis. 

Die Unterdrückung der Sporenbildung in diesem Falle kann 
allem Anschein nach eher dem direkt hemmenden Einfluss des 



154 N. ONO : WACHSTHUMSBESCHLEUNIGUNG 

betreffenden Stoffes zugeschrieben werden, als der durch stärkere 
Mycelentwickelung hevorgerufenen Correlationserscheinung, welche 
häufig bei günstigstem Nährboden zu Stande kommt. Um diesen 
Punkt aufzuklären, stellte ich die Versuche derart an, dass ich 
ein Stück Mycelium, welches für 19 Tage in 0.015% NaFl Lösung 
gezüchtet worden war und keine eigentliche Fruktifikation, 
abgesehen von einigen schon besprochenen haarigen Gebilden, 
aufnimmt, nach Bespülung mit Wasser in Normallösung brachte. 
Schon nach 2 Tagen kamen die angelegten Träger zur Reife 
und schwärzten, während ich auf der noch in 0.0159^ verblei- 
benden Decke vergeblich nach den reifen Sporangien suchte. 
Dieses E-esultat spricht ohne weiteres für die gesagte Ansicht. 

Es fragt sich nun, ob die stärkere Entwickelung des vege- 
tativen Organs der Pilze bei der Zugabe einer kleinen Dosis 
giftiger Stoffe nicht eher als ein specieller Fall der Correlations- 
erscheinungen zwischen Fortpflanzung und Wachsthum zu be- 
trachten ist, indem der direkt hemmende Einfluss der Substanzen 
auf die Sporenbildung durch Correlation die vegetative Funk- 
tion befördert. Es ist schon bestätigt wurden, dass Sporenbil- 
dung der Pilze durch einige Gifte^) viel empfindlicher beeinflusst 
wird als die vegetative Mycelentwickelung. Fasst man dieses 
Verhältnis ins Auge, so ist es wohl begreiflich, dass bei einer 
genügenden Verdünnung der angewandten Stoffe jene Concentration 
erreicht ist, welche an sich für die Mycelentwickelung unschäd- 
lich, aber für die Sporenbildung hemmend wirkt, und dass so 
infolge der Correlation das Wachsthum des vegetativen Theils 
ungewöhnlich gesteigert wird. Dieser indirekten Wirkung der 
betreffenden Stoffe schreibe ich, ausser dem direkten Peizeffekt, 
die Wachsthumssteigerung zu. 



1) O. Loew, Ein natürliches System der Giftwirkungen 1893. 



EINIGER ALGEN UND PILZE. 155 

V. Einfluss der Reizstofife auf die Betriebsstoffwechsel. 

Die Pflanze nimmt clurcli ihre Lebensthiitigkeiten die Nähr- 
stoffe auf und verwendet einen Tlieil derselben zum Aufbauen ihres 
Körpers, hingegen den anderen Tlieil znm Oxydationsmaterial, um 
dadurch die nothwendige Betriebsenergie sich zu verschaffen. 
Von diesem Standpunkte aus betrachtet, lassen die im Pflanzen- 
organismus sich abspielenden Stoflumsätze sich, wie bekannt, in 
zwei Kategorien : Bau- und Betriebsstoftweclisel trennen. Um 
die Grösse jedes von diesen Stoffwechseln zu ermitteln, hat man 
einigermassen einen Maasstab. So ist bei der Betriebsstofiwechsel- 
thätigkeit das Trockengewicht maassgebend, und für den Betriebs- 
stoffwechsel giebt die Ermittelung der Kohlensäureproduktion, die 
Ausscheidung gewisser Stoffw^echselprodukte, einige Aufschlüsse. 

AVie aus der in der Einleitung angeführten Skizze zu ersehen, 
beziehen sich die bisherigen Untersuchungen über die Erhöhung 
der Lebensthätigkeiten durch chemische Reize zumeist nur auf 
Baustoffwechsel. Bichards untersuchte in seiner Arbeit nur den 
Ernteertrag, berücksichtigte aber nicht den Betriebsstofiwechsel. 
Schulz beobachtete stärkere Entwickelung der Kohlensäure bei 
Hefen. Beim ersten Anblick scheint dies auf nur erhöhtem 
BetriebsstofPvvechsel zu beruhen, doch blieb hier die Frage immer 
offen, ob man es in diesem Falle mit der Erhöhung der Gähr- 
thätigkeit einzelner Individuen zu thun hat, oder ob diese 
Erscheinung von der durch die Beizwirkung verursachten Ver- 
mehrung der Gährungserreger bedingt sei. 

Es fehlen bisher meines Wissens einschlägige Versuche über 
die Beeinflussung des Betriebsstoffwechsels in Gegenwart von 
giftigen Stoffen. Irgend ein Beitrag in dieser Richtung durfte 
wohl nicht ohne Interesse sein. 



156 N. ONO : WACHSTHUMSBESCHLEUNIGUNG 

Für solchen Zweck bieten die Algen keine geeigneten Objekte 
dar, wohl aber die Pilze, welche sich dafür bequem anwenden lassen. 

Einige Pilze, insbesondere Aspergillus niger, produciren eine 
nicht unbeträchtliche Menge Oxalsäure als Stoffwechselprodukt, 
wie aus der bekannten Arbeit We h m er 's^) hervorgeht. Dieses 
Stoffwechselprodukt bot bei meinen Versuchen einen Angriffspunkt, 
und so werden eine Peihe Bestimmungen über die Säurenmenge 
angeführt. Ich muss hier bemerken, dass ich die Säure nur auf 
titrimetrischen Wege bestimmte. Die Methode ist untauglich, wenn 
Oxalsäure nicht nur als solche, sondern auch als Salz vorkommt. 
We hm er zeigt aber in seiner oben besprochenen Arbeit, dass in 
NH4N03-haltiger Zuckernährlösung Oxalsäure stets als freie Säure 
bei Äspergillus-K.\x\iviY&[i auftritt und da meine Kulturen haupt- 
sächlich derartige waren, so war die Titration zuverlässig. Dies 
geschah mit Liquor Alkali Decinormalis und Phenolphthalein 
als Indikator. 

Nachdem ich von der durch Titration ermittelten gesammten 
Säure die ursprüngliche Acidität der Nährlösung subtrahirt hatte, 
rechnete ich diese als Oxalsäure um, welche in der angeführten 
Tabelle gegeben ist^). 



1) Welimer, Entstehung und physiologische Bedeutung der Oxalsäure im Stoffwechsel 
einiger Pilze, Bot. Ztg. 1890. 

2) Hier werde ich einige Versuche, um die titrimetrisch ermittelten Zahlen mit denjenigen, 
die gravimetrisch bestimmt waren, zu vergleichen, ausgeführt angeben : — 

Silurenmenge in g in 10 cc Niihrflüssigkeit. 



Zusatz von NiSO^ 


Titration 


Gravimetrisch 





0.042 


0.048 


0.003°^ 


0.052 


0.060 


0.007 „ 


0.050 


0.058 


0.014 „ 


0.066 


0.075 


0.028 „ 


0.064 


0.060 





0.042 


0.040 


003 „ 


0.047 


0.048 


0.007 „ 


0.052 


0.058 


0.014,, 


0.060 


0.065 


0.028 „ 


0.051 


0.054 



EINIGER ALGEN UND PILZE. 157 

Vergleicht man nun die Säurenraenge bei Gegenwart von 
Reizmitteln mit derjenigen der Kontrollversuche, so findet man in 
unseren Versuchen nur mit der einzigen Ausnahme von NiSOj 
stets das Minus im ersteren Falle. Dieses Verhältnis ist ersicht- 
lich aus der Colonne ,, Säure pro lg Pilzsubstanz" in den Tabel- 
len. Steigt der Zusatz von Reizstoffen, so ^Yird die Menge der 
Säure um so kleiner. 

Man kann jedoch nicht annehmen, dass die aufgefundene 
Säurenmenge die sämtliche Menge der ausgeschiedenen Säure 
darstellte, da bekanntlich die Ausscheidung der Säure mit ihrer 
Zersetzung Hand in Hand gehen sollte. Daraus geht hervor, dass 
die Erklärung der besprochenen Verhältnisse nicht einzig in ihrer 
Art sein kann. Es könnten einige Möglichkeiten, welche für 
diese Thatsache sprechen, angegeben werden. 

Erstens, wenngleich die Oxalsäure als normales Stoff- 
wechselprodukt unseres Pilzes auftritt, ist sie doch als ein Pro- 
dukt unvollkommener Oxydation anzusehen, und wenn die Stoff- 
wechselthätigkeit auf einmal gesteigert wird, so wird als das 
Produkt vollkommener Oxydation mehr Kohlensäure entstehen, 
dagegen weniger Oxalsäure. 

Zweitens könnte dieselbe Erscheinung auftreten, falls die 
einmal entstandene Säure durch Wiederverarbeitung seitens der 
Pilze verschwindet. Dabei könnte sie entweder als Baumaterial 
wiederaufgenommen werden oder, ohne wieder den Pilzen nutzbar 
zu werden, zersetzt werden. Doch, wie schon Wehmer^) betont 
hatte, stellt die Oxalsäure einen nur sehr armen Nährstoff für 
Aspergillus dar, so dass es wahrscheinlich ist, dass sie bei 
zureichendem Vorrath von guten Nährstoffen wie Zucker") wohl 

1) Wehmer I.e. 

2) Bei meinem Versuche betrug der Zuckergehalt nach Beendigung der Versuche wenig- 
stens 1.5 g in je 50 cc der Kulturflüssigkeit, d. h. ca o%- 



158 N. ONO : WACHSTHUMSBESCHLEUNIGUNG 

intakt geblieben wäre. Ferner wurde von Welimer^) constatirt, 
dass weder Licht noch tote Pilzmasse allein die Zersetzung der 
Säure hervorzurufen im Stande sind, wohl aber die Lebensthä- 
tigkeiten der Pilze. Es bleibt daher nur die Annahme übrig, 
dass die Säure durch Steigerung des Betriebsstoffwechsels lebhaft- 
erer Zersetzung unterworfen worden sei. 

Die dritte Möglichkeit ist schliesslich die, dass diejenigen 
Stoffe (Kohlenhydrat u. s. w.), welche bei normaler Wachsthums- 
energie durch Stoffwechsel z. Th. als Oxalsäure auftreten, bei 
der infolge der Peizwirkung über Norm gesteigerten Wachthums- 
thätigkeit nicht als jene Form abgesondert werden, sondern sich 
gerade in den integrierenden Theil des Pilzkörpers umwandel- 
ten, kurz, dass sie als Baustoff verwendet werden. 

Von einer anderen Seite müssen wir also dieses Problem an- 
greifen, um zu entscheiden, ob die eine oder andere von diesen Mög- 
lichkeiten für unseren Fall zutrifft. Die Ermittelung der Kolilen- 
säureausscheidung, des ökonomischen Coëfficienten') u. a. wird wohl 
an diesen Punkt anschliessend oder w^enigstens rathgebend sein. 

Im Folgenden gebe ich die Resultate meiner Bestimmungen 
ökonomischer Coëfïicienten bei den Kulturen mit Zusatz von 
ZnSOj, bei denen die Erntezunahme stets am auffallendsten war. 

Die Bestimmung des Coëfïicienten fand in folgender Weise 
statt : 

Die Kulturflüssigkeit wurde zunächst durch andauerndes 
Kochen mit verdünnter Salzsäure vollkommen invertirt. Dann 
verdünnte ich diese bis zu etwa ^o "/o Zuckergehalt. Eine Burette 
wurde mit der betreffenden Lösung gefüllt. 

In einem Kolben mischte ich genau 10 cc Fehling'sche 

1) Wehmer I.e. 

2) Man vergleiche hierüber H. Kunst mann, Ueber das Verhilltniss zwischen Pilzernte 
und verbrauchter Nahrung. 1895. (Leipziger Dissertation). 



EINIGER ALGEN UND PILZE. 



159 



Lösung mit etwa 40 cc Wasser und braclite es zum Sieden ; darauf 
fügte ich die oben genannte Lösung hinzu, bis schliesslich durch 
vollkommene Keduction des Kupfers zu Kupferoxydul die 
Flüssigkeit farblos geworden w^ar. 

Da unsere Fehling'sche Lösung in 1000 cc 34.64g Kupfer- 
sulfat, 174g Kaliumnatriumtartrat und 120g Natriumhydroxyd 
enthielt, so sollte je 10 cc derselben durch 0.05g Zucker reducirt 
werden. 

Nun kann man leicht durch die Lesung der Burette die 
Zuckermenge in der Kulturfiüssigkeit kennen lernen. Die 
Differenz zwischen der ursprünglich vorhandenen Zuckermenge 
in der Kulturflüssigkeit und der zurückbleibenden ergiebt selbst- 
verständlich die verbrauchte Zuckermenge. 

Kulturen mit Zusatz von ZnSO^. 

Aspergillus niger. 
Kulturdauer 14 Tage. Zimmertemperatur. 



Gelialt an ZnSU, 
(Gew. %) 



l'ilzL'inte in 



Verbrauchte 
Zuckermeiige 



Ökonomischer 
Coefficient 
Verbrauch 



d.h. 



Erute 



I. 






0.262 


1.594 


6.1 


0.0037 


0.860 


2.429 


2.8 


0.0074 


0.875 


2.429 


2.8 


0.0148 


0.785 


2.380 


3.0 


0.0297 


0.773 

11 


2.340 

[. 


3.0 





0.386 


1.707 


4.4 


0.0037 


0.924 


2.463 


2.7 


0.0074 


0.928 


2.463 


2.7 


0.0148 


0.918 


2.448 


2.7 


0.0297 


0.837 


2.480 


2.8 



160 



N. ONO : WACHSTHÜMSBESCHLEUNIGUNG 



III. 




0.0037 
0.0074 
0.0148 
0.0297 



0.392 
0.910 
0.908 
0.844 
0.827 



1.819 
2.462 
2.456 
2.456 
2.446 



4.6 

2.7 
2.7 
2.9 
2.8 



Was sich nun aus diesem Resultate beurtlieilen lässt, ist, 
dass der ökonomische Coefficient in jedem Falle bei weitem 
grösser ist in Kontrolle d.h. in nicht zugesetzter Kultur als in 
zugesetzter. Dieses Verhältnis deutet also an, dass die Pilze bei 
Anwesenheit von Zinksulfat veranlasst wurden, mit einem ver- 
hältnismässig kleinen Verbrauch von Zucker eine bedeutend 
grössere Körpersubstanz aufbauen zu können. So scheint mir, 
wenigstens für Zinksulfat, von den oben besprochenen drei 
Möglichkeiten die dritte die wirkliche zu sein. 

IV. Specielle Besprechungen. 

ZnSO,. 

Unter den von Kichards geprüften Stoffen übt dieses Salz 
die stärkste Wirkung aus. 

Auch bei unseren Versuchen mit Algen wirkte ZnSO^ nächst 
FeSOi sehr günstig auf das Wachsthum ein. Schon bei Zusatz 
von einer minimalen Quantität, wie 0.0000169^, nahm die Ernte 
etwas zu, und dies war noch deutlicher bei 0.000069^ bis 0.00039^. 
Stieg die Concentration auf 0.00169^, so litten die Algen nicht 
unerheblich, ohne jedoch das Waclisthum ganz herabzusetzen 
(cf. Tabelle. Algen A. I-IV). 

Unsere Versuche mit Pilzen stimmen mit denjenigen von 



EINIGER ALGEN UND PILZE. 161 

Richards überein. Bei längerem Stehen wurde der Unterschied 
zwischen den Versuchs- und Kontrollekulturen recht über- 
raschend (cf. Tabelle. Pilze A. I-III). 

Sehr sonderbar trat einmal bei einer Versuchsreihe mit 
Dextrose es hervor, dass kein nennenswerther Unterschied in 
der Ernte sowohl, als auch in der Säurequantität sich erkennen 
Hess. Den Grund davon kann ich aber nicht erklären (Tabelle. 
Pilze. A. IV.) 

Die gelbliche Färbung von Nährflüssigkeiten, sowie die 
Biklung der bräunlichen Sporen in den Versuchskulturen, welche 
schon von Autoren besprochen wurden, waren hier bemerklich. 

Die Säurenmenge nach Beendigung der Versuche war in 
Versuchskulturen viel kleiner als in Kontrollen (Tab. Pilze. 
A. I-III.). 

FeS04. 

Bichards giebt au, dass dieses Salz erst bei ziemlich gros- 
sem Gehalte einen schädigenden Einfluss ausübt. 

Meine betreffenden Versuche mit Algen zeigten auch, dass 
dasselbe noch höhere Concentration im Vergleich zu anderen 
Schwerenmetallsalzen erträgt. So lag bei Hormidiwn das Optimum 
etwa bei O.00059o und sogar bei einer höheren Concentration 
wie 0.01269^ war der Ertrag noch etwas grösser als bei den 
Kontrollen (cf. Tab. Algen. B. I-IL). 

In einer mit Zusatz von FeS04 angestellten Penicillium- 
Kultur trat merkwürdigerweise das ziegelroth gefärbte Mycelium 
zu Tage. 

NiS04. 

Bei Algen ruft der Zusatz von NiS04 einen befördernden 
Einfluss hervor. Die optimale Dosis lag etwa zwischen 0.00006 



162 



N. ONO : WACHSÏHUMSBESCHLEUNIGUNG 



und 0.00012%, während 0.0028"% eine beschädigende Wirkung 
ausübte (Tab. Algen C. I. II.). 

Säureproduktion bei Pilzkulturen war hier im Gegensatz zu 
den meisten Fällen grösser mit der Erhöhung der Zusätzeprocente 
(cf. Tabelle Pilze. C. I-IIL). 

C0SO4. 

Bei Algen scheint dies auch einen begünstigenden Einfluss 
auszuüben, doch lag der optimale Punkt etw^as niedriger als bei 
NiSOi ; Optimum etwa bei 0.00012?^ (cf. Tab. Algen D. I. IL). 

Säureerzeugung war wie gewöhnlich kleiner und zwar sehr 
regelmässig in Versuchskulturen. 

CuSO,. 

CuSOj wurde von Richards nicht untersucht. Im Jahre 
1897 constatirte Günther^), dass Kupfersalze in grösseren 
Mengen das Wachsthum der Pilze retardirten, in geringeren 
Mengen dagegen besseres Gedeihen mit sich bringen. Auch 
bei meinen mit Aspergillus und Pemcillium angestellten Ver- 
suchen beobachtete ich dieselbe Erscheinung. Hattori") fand 
auch in seinen Untersuchungen über die Giftwirkung der Kup- 
fersalze eine ähnliche Thatsache. 

Hier werde ich zwei Beispiele angeben ; für näheres verweise 
ich auf die tabellarische Zusammenstellung (Tab. Pilze. E.). 





Aspergillus niger. 






Gelialt an CuSO, 





0.0015% 


0.003% 


0.006% 


0012% 1 


Ernteertrag in g 


0.273 


0.307 


0.313 


0.324 


0.345 



1) E. Günther, Beitrag zur mineralischen Nahrung der Pilze. Erlangen 1897. 

2) H. Hattori, Ucber die Einwirkung des Kupfersulfates auf einige Pllanzen. 
Manuskript. 



EINIGER ALGEN UND PILZE. 



163 



Pénicillium glaucum. 



Gehalt an CuSO^ 


0.0015 0/ 


0.0039^ 


0.0069^ 


0.0129^ 


Ernteertrag in g 


0.213 


0.320 


0.338 


0.359 


0.410 



Bei Algen konnte ich dagegen keine Wachsthumsbeförde- 
rung nachweisen, wie aus der Tabelle ersichtlich ist. Schon bei 
0.00001^ steht die Ernte etwas zurück (Tab. Algen E.). Ob 
bei noch weiterer Verdünnung die wachsthumsbegünstigende 
Concentration erreicht sein könnte, lasseich vorläufig unbestimmt. 

Die Säurenmenge in Pilzkulturen war kleiner in Versuchs- 
kulturen (Tab. Pilze. E.). 

HgCl. 

Es ist von gewissem Interesse, dass dieses heftige Gift in 
genügender Verdünnung auch das Wachsthum der Pilze befördert. 
Schulz^) giebt an, dass Kohlensäureentwickelung der Hefe in 
Gegenwart einer kleinen Menge des Stoffes gesteigert wird. Der 
optimale Zusatz dabei ist etwa 1/500 000. 

Was Schimmelpilze betrifft, so findet man in der bisherigen 
Litteratur nur die Rede von dem schädigenden Einfluss des 
betreffenden Stofis. Raulin") betrachtet z. B. dies mit AgNOg, 
PtsCIe zusammen als das giftige Salz für Aspergillus. Er gibt 
1/512 000 als die Grenze der Giftwirkung. In seinem Experi- 
mente mit 1/819 200 konnte er jedoch keinen wachsthums- 
beschleunigenden Einfluss beobachten. Meines Wissens liegt 
uns zur Zeit kein Versuch vor, welcher die letztgenannte That- 
sache in positivem Sinne zeigt. 



1) H. Schulz, I.e. 

2) Raulin I.e. p. 134. 



164 



N. ONO : WACHSTHUMSBESCHLEUNIGUNG 



Meinem Versuclie nach (cf. Tab. Pilze. F.) tritt schon bei 
Verdünnung von 0.00179^ oder 1/60 000 ziemlich gute Ent Wicke- 
lung von Aspergillus ein. Stieg die Concentration auf 1/30 000, 
so kam die Entwickelung zum Stillstand. Die Grenze für Gift- 
wirkung liegt zwischen 1/60 000 und 1/30 000. 

Das Optimum war sowohl bei Peincillium als auch bei 
Aspergillus etwas unter 0.00139^). 

Hier gebe ich zwei Beispiele (cf. Tab. Pilze. F. I-VI). 






Aspergillus niger. 



Gehalt an HgClg 





0.0003 9^ 


0.0007^ 


0.00139^ 


0.00270/ 


Ernteertrag in g 


0.261 


0.355 0.354 


0.509 


0.451 



Pénicillium glaucum. 



Gehalt an HgCl^ 





0.0003 9^ 


0.0007 9^ 


0.0013^ 


0.0027«/ 


Ernteertrag in g 


0.183 


0.249 


0.213 


0.311 


0.246 



Säureproduction ist hier wie bei den meisten Fällen kleiner in 
Versuchskultui-en als bei Kontrollen (Tab. Pilze. F.). 

Auf Algen übte dieses Salz keinen beschleunigenden Einfluss 
aus, sondern wirkte nur giftig ein. Schon bei 0.000059^ war der 
schädigende Effekt deutlich zu erkennen. Doch weitere Ver- 
dünnung durfte ich nicht ausführen, da bei solchen hohen 
Verdünnungen einige Fehlerquellen als maasgebend auftreten 
(Tab. Algen F.). 

LiNOs. 

Von Kichards wurde LiCl zur Untersuchung herausgezogen 



EINIGER ALGEN UND PILZE. 165 

und in diesem Salze ein eine beträchtliche Wachsthurassteigerung 
hervorrufender Stoff gefunden. Bei meinem Versuche benutzte 
ich LiNOg mit gleichem Eesultate (Tab. Pilze G.) 

Säureproduktion war hier auch kleiner in Versuchskulturen 
als in Kontrollen. 

Algen zeigten auch eine besseres Gedeihen in zugesetzten 
Kulturen (Tab. Algen H.). 

NaFl. 

Dieser Stoff übte auch eine beschleunigende Einwirkung auf 
Algen aus. Der optimale Punkt liegt etwa bei 0.000039^, stieg 
die Concentration zu 0.00016?^ bis 0.00089^, so nahm die Ernte 
etwas ab, war noch grösser als bei Kontrollen. Erst bei 0.00429^ 
steht der Ertrag im Vergleich zur Kontrolle etwas zurück 
(Tab. Algen G.). 

Bei Pilzen beförderte dieser Stoff das Wachsthum (cf. Tab. 
Pilze H.). Seine Wirkung auf die Sporenbildung wurde schon 
im vorstellenden Capitel behandelt. 

Die Säurenmenge in der Kährflüssigkeit war wie gewöhnlich 
kleine) in zugesetzten als in Kontrolle-Kulturen. 

Arsen. 

Von Arsenverbindungen ist arsenige Säure giftig, doch ver- 
tragen höhere und niedere Pflanzen viel Arsensäure^). 

Da Arsenigsäureanhydrid nur schwer löslich ist, so bediente 
ich mich des arseuigsauren Kaliums. 

Bei Penicilliu7n-l\.u\iuY war kein bedeutender Unterschied 
der Ernte sowohl als auch in der Säureproduction bemerklich. 



1.) O. Loew, System der Giftwirkuugen. 



166 N. ONO : WACHSTHUMSBESCHLEUNIGUNG 

Merkwürdig war bier eine eigentliche Geruchsentwickelnng in 
zugesetzten Knlturen^). 

Auf Algen scheinen die genannten Salze etwas AVachsthums- 
begünstigend zu wirken. (Tab. Alg. I). 



VII. Schlussbemerkungen und Zusammenfassung 
der Resultate. 

Aus dem Vorstehenden geht zunächst hervor, dass die 
chlorophyll führenden niederen Organismen wie Algen in ihrem 
Gedeihen günstig beeinflusst werden durch einen geringen Zusatz 
von einigen Stoffen, welche für sich nicht Nährstoffe sind, ja 
sogar giftig wirken. In dieser Reaktion verhalten sich die 
Algen gerade wie die Pilze. Nur ist zu bemerken, dass die 
optimale Dosis für Algen viel kleiner als bei Pilzen ist, eine 
Thatsache, welche vielleicht vom oekologischen Standpunkte aus 
ihren Aufschluss haben wird. Von den geprüften Stoffen konnte 
ich nur bei Quecksilberchlorid und Kupfersulfat die besprochene 
Keaktion nicht constatiren, indem ich bei ihnen, soweit meine 
Versuche reichten, stets Giftwirkung beobachtete. Daraus muss 
aber nicht geschlossen werden, dass den beiden Stoffen die 
nämliche Eigenschaft nicht zukommt, da man bei ihnen unter 
Umständen doch noch jene wachsthumsbegünstigende Einwirkung 
wohl erwarten kann. 

Bei Pilzen konnte ich die früheren Versuche Richards' 
hauptsächlich bestätigen, dazu prüfte ich mit positiven Resultaten 
einige bisher noch nicht untersuchte Stoffe. 

Die Verzögerung oder Verspätung der Sporenbildung bei 
unseren Versuchen ist nicht als infolge einer üppigen vegetativen 

1) Schon von Gasio (Jaliresber. über Gährungsorganismen 1893) erörtert. 



EINIGER ALGEN UND PILZE. 167 

EntwickeliiDg verursachte Correlationserscheinurig, vielmehr als 
durch Keizstofle bewirkte Hemmung zu betrachten. 

Was nun die Art und Weise der Reizwirkung anbelangt, 
so bemerke ich folgendes : 

Wenn es sich hier zunächst um zeitliche oder andauernde 
Hyperaesthesia handelt, so muss Bau- und Betriebsstoffwechsel 
gleichzeitig gesteigert werden. 

Wenn aber dagegen durch Zusätze der Reizstoffe die Thätig- 
keiten seitens des Organismus so gesteigert werden, dass sie mit 
kleinerem Energieaufwand die Nährstoffe in sich aufnehmen und 
sich bauen, kurz, ökonomisch arbeiten können, so kann der dyna- 
mische Stoffwechsel nicht so erheblich beeinflusst bleiben. Um 
daher in dieser Hinsicht eine richtige Auffassung zu gewinnen, ist 
ein Einblick in den Betriebsstoffwechsel von Wichtigkeit. Einige 
von meinen Versuchen in dieser Richtung zeigten andeutungsweise, 
dass Betriebsstoffwecbsel nicht parallel mit Baustoffwechsel 
gesteigert werden ; doch sind zur Zeit meine diesbezüglichen 
Versuche leider unzureichend, um in bezug auf diesen Punkt 
Allgemeines zu sagen. 

Zum Schluss seien im Folgenden die wichtigsten Resultate 
kurz zusammengestellt : — 

1. Das Gedeihen der niederen Algen wird durch Einfüh- 
rung gewisser giftiger Stoffe in höchst verdünnten Zuständen 
begünstigt. Hierzu gehören ZnSO^, NiSO,, FeSO,, CoSO,, NaFl, 
LiNO,, K2ASO3. 

2. Die Erntezunahme bei Alsfcn muss auf die veo-etative 
Vermehrung der Individuenzahl zurückzuführen sein, da keine 
nennenswerthe Veränderung der Körpergrösse bemerkbar war. 

3. Die geeignete Dosis ist bei Algen bedeutend kleiner als 



168 N. ONO : WACHSTHUMSBESCHLEUNIGUNG 

bei Pilzen. Schon der Zusatz von 10~^ Gr. Mol. Salz wirkte in 
den allermeisten Fällen schädigend. 

4. In CUSO4 und HgClo fand ich, soweit unsere Studien 
ausreichen, keine beschleunigende Wirkung auf Algen, wohl 
aber begünstigend bei Pilzen. 

5. Bei Pilzen tritt durch Zusätze von HgClo (Optimum etwa 
bei 0.00139^ und CuSO^ (Optimum etwa bei 0.012^0 die Wachs- 
thumsbeschleunigung ein. 

6. Die Säurequantität in Kulturen mit Zusatz von ZnS04, 
C0SO4, HgCl.2, NaFl, CUSO4 war stets kleiner als in Kontrol- 
kulturen. Nur verhielt NiSO^, soweit meine Versuche ein 
Urtheil gestatten, sich diametral entgegengesetzt. 

7. Die geprüften Stoffe (speciell ZnS04 und NaFl) neigen 
dazu, die Sporenbildung der Pilze direkt zu hemmen, wenigstens 
das Auftreten der Sporen zu verspäten. 

8. Die oekonomischen Coefficienten in ZnSO^-Kultur sind 
in der Kontrolle, d. h. in der nicht zugesetzten Kultur, bei 
weitem grösser als in der zugesetzten. 

Juni, 1899. Botanisches Institut 

Kaiserl. Universität 
zu Tokyo. 



EINIGER ALGEN UND PILZE. 



169 



TABELLARISCHE ZUSAMMENSTELLUNG. 



BEMERKUNGEN : 

Versuche mit Algen — Der Entwickelungsgrad ist entweder mit Erntegewicht oder mit 
den relativen Werth gebenden Ziffern bezeichnet. 

Versuche mit Pilzen — In Colonne „Aciditüt" ist die Quantität des Decinormal Alkalis 
in cc gegeben, welche 10 cc der Nährflüssigkeit neutralisirte (Die ursprüngliche Aciditüt 
wurde natürlich vorher subtrahirt). Daraus ermittelte ich die Säurenraenge in je 50 cc Nillir- 
(iüssigkeit, berechnete sie als Oxalsäure inid in Colonne „als Oxali^äure umgerechnet" angab. 

In Kolonne „Säure pro lg Pilzsubstanz" ist das Verhältnis — j^J""'"*^ gegeben. 

Obwohl bei einigen Pe/u'aY/iiijft-Kulturen die Säureumenge gegeben sind, durfte ich doch 
nicht die ^ "''"^ '^ ermitteln, da bei Pénicillium das Titration unzuverlässig sclieint. 



I. Versuche mit Algen. 



A. I. Kulturen mit Zusatz von ZnSO^. 
Ernteertrao- in " Protococcxis sp. — augestellt 5. Oct. 



Zimmer-Temperatur. 



'öS 


Gram Mol. 





i X 10-« 


l X 10-5 


l X 10-^ 


Kultur- 
dauer 


Gew. o/o 





0.000014 


0.00014 


0.0014 


I 
II 
III 


0.010 
0.016 
0.012 


0.018 
0.023 
0.019 


0.018 
0.018 
0.021 


0.002 
0.009 
0.006 


23 Tage 
26 

9 



A. IL Kulturen mit Zusatz von ZnSO^. 
Ernteertrag in g. Protococcus sp.— angestellt 11. Oct. Zimmer-Temperatur. 





Gram 
Mol. 





J.xlO-^ 


i X 10-5 


10-5 


1x10-^ 


Kultii- 
dauer 


Gew. o/o 





0.000012 


0.00006 


0.0003 


0.0014 


I 

II 
III 


0.035 
0.032 
0.030 


0.043 
0.040 
0.040 


0.038 
0.036 
0.037 


0.042 
0.040 
0.042 


0.024 
0.023 
0.020 


44 Tage 
46 „ 
50 „ 



170 



N. ONO : WACHSTHUMSBESCHLEUNIGUNG 



Ernteertrac; in 



A. III. Kulturen mit Zusatz von ZnSO^. 

Chroococcum — angestellt 24. Dec. j^^ Treibhaus 16-20° C 



's 


Gram 
Mol. 





2^x10-5 


^xlO-5 


10-5 


i X 10-^ 


Kultur- 
dauer 


Gew. o/o 





0.000012 


0.00006 


0.0003 


0.0014 


I 

II 
III 


0.006 
0.009 
0.010 


0.015 
0.021 
0.024 


0.026 
0.022 
0.026 


0.022 
0.024 
0.026 


0.006 
0.010 
0.009 


71 Tage 



A. IV. Kulturen mit Zusatz von ZnSO^. 
Ernteertrag in g. Pro^ococcus— angestellt 5. Feb. -^^ Treibliaus 16-20° C. 






Gram 
Mol. 





2*5X10-5 


1x10-5 


10-5 


-1 X 10-^ 


Kultur- 
dauer 


Gew. 0/ 





0.000012 


0.00006 


0.0003 


0.0014 


I 

II 
III 


0.008 
0.009 
0.011 


0.017 
0.018 
0.015 


0.019 
0.023 
0.018 


0.019 
0.016 

0.017 


0.007 
0.009 
0.005 


65 Tage 



E. I. Kulturen mit Zusatz von FcSO^. 
Ernteertrag in g. ITonnUUum n;7e/(S— angestellt 7. Oct. Zinnner-Temperatur. 



■US 

'S 




Gram 
Mol. 

Gew. 0/ 





2^5x10- 


1x10-^ 


10-^ 


è+10-^ 


Kultur- 
dauer 





0.0001 


0.0005 


0.0025 


0.0126 


I 

II 
III 


0.038 
0.028 
0.031 


0.072 
0.083 
0.070 


0.074 
0.079 
0.082 


0.080 
0.062 
0.074 


0.042 
0.041 
0.039 


65 Tage 



B. IL Kulturen mit Zusatz von FeSO^. 
Ernteertrag in g. Honnidium ?i!7e»s- angestellt 10. Nov. jj., TreibJiaus 16-20° C. 






Gram 
Mol. 





2^5x10-^ 


1 X 10- ' 


10-^ 


1 X 10-=^ 


Kultur- 
dauer 


Gew. 0/ 





0.0001 


0.0005 


0.0025 


0.0126 


I 

II 
III 


0.025 
0.023 
0.027 


0.050 
0.067 
0.064 


0.052 
0.049 
0.058 


0.054 
0.042 
0.046 


0.041 
0.032 
0.032 


79 Tage 



EINIGER ALGEN UND PILZE. 



171 



Ernteertrag in g. 



C. I. Kulturen mit Zusatz von NiSO^ 
Chroococnnn — angestellt 2. Oct. 



Zimmer-Temperatur. 





Gram 
Mol. 





2^5 X 10"' 


l X 10-^ 


10-^ ix 10-^ 

1 


Kultur- 
dauer 


Gew. o/g 





0.000012 


0.00006 


0.00028 


0.0014 


I 

II 
III 


0.012 
0.011 
0.013 


0.012 
0.015 
0.018 


0.025 
0.024 
0.020 


0.021 
0.020 
0.022 


0.004 
0.006 
0.007 


64 Tage 



C. II. Kulturen mit Zusatz von NiSO^. 
Ilormidium niiens — angestellt IG. Nov. 



in Treibhaus 16-24° C. 



's 


Gram 
Mol. 





JgXlO-^ ix 10-= 


10-5 


i X 10-^ 


Kultur- 
dauer 


Gew. 0/^ 





0.000012 


0.00006 


0.00028 0.0014 


I 

II 

III 


4 

4 

3-4 


5 

5 
5 


4-5 

4 

4-5 


4 
4 
4 


1 
1 
1 


70 Tage 



N.B. Die Zitïtr zeigt dau Entwickfluugsgrad, 

D. I. Kulturen mit Zusatz von CoSO^. 
Hnrmidium niiens — ^angestellt 9. Dec. 



in Treil.liaus lG-24° C. 



'S 
'03 


Gram 
Mol. 





1 xlO -^ -ixlO-^ 10-^ 

1 


ix 10-^ 


Kultur- 


Gew. o/g 


. 


0.000012 


0.00006 ! 0.0003 0.0014 

1 


dauer 


I 
II 

III 


4.5 

4 

4 


5 4.5 4 2.5 
5 4.5 ^ 3.5 ' 2 

5 4.5 4 2 


70 Tage 



Ernteertrag in g. 



N.B. Die Zitier zeigt deu Entwickelungsgrad. 

D. II. Kulturen mit Zusatz von CoSO^. 
Prnfococcus — angestellt 19. Mai. 



Zimmer-Temperatur. 



■1^ 

Co 

'03 
O 


^^;"7 JgxlO ^ ixlO-' 10-^ !ixio-^ 

Mol. i --• 1 ^ 1 -^ 


Kultur- 
dauer 


Gew. o/o 1 0.000012 0.00006 


0.0003 0.0014 


I 

II 
III 


0.012 î 0.024 
0.009 0.C30 
0.010 0.024 


0.026 
0.025 

0.020 


0.020 
0.022 
0.018 


0.010 
0.007 
0.006 


32 Tage 



172 



N. ONO : WACHSTHUMSBESCHLEUNIGUNG 



E. I. Kulturen rait Zusatz von CuSO^. 
Stigeodonium — angestellt 14. Oct. 



Zimmer -Temperatur. 



ö 

1—1 
Co 
f-i 

<v 
Ü5 


Gram ^ 
Mol. 


2^x10-' 


1 X 10-^ 


10-=^ 


i X 10-^ 


Kultur- 
dauer 


Gew. o/o 





0.00001 


0.00005 


0.00025 


0.0012 


I 

II 
III 


5 

5 
5 


4 
4 

4 


2 
2-3 

2 


1 
1 
1 


1 
1 
1 


27 Tage 



N.B. Die Zitier zeigt deu Entwickeluiigsgrad. 

E. IT. Kulturen mit Zusatz von CUSO4. 
Ohroococcum — ans;estellt 13. Dec. 



in Treibhaus 16-20° C. 



'Öl 



Gram ^ 
Mol. 


2^x10- 


ix 10-^ 


10-^ 


ix 10-^ 


Kultur- 
dauer 


Gew. 0/, 0.00001 

1 


0.00005 


0.00025 


0.0012 


I 

II 
III 


5 

4-5 

5 


4 

4-5 

4 


3 
3 
3 


2 
2 
2 







71 Tage 



N.B. Die Ziiïer zeigt den Entwickelungsgrad. 

F. I. Kulturen mit Zusatz von HgCl^. 
Frotococru^ — angestellt 25. Dec. 



in Treibhaus 16-20° C. 





Gram ^ 
Mol. 


2's X 10-"' 


ixlO-^ lO-'^ 


l X 10-^ 


Kultur- 
dauer 


Gew. 0/0 





0.00001 


0.00005 


0.00025 


0.00124 


I 

II 
III 


5 
5 

5 


4 
3 
3 


1 















43 Tage 



Ernteertrag in 



N.B. Die Ziffer zei>ft deu Eutvviokeluuffsgrad 

G. T. Kulturen mit Zusatz von NaFl. 
Protococcus — angestellt 24. Dec. 



in Treibhaus 16-20° C. 



.4-3 

's 
'S 




Gram 
Mol. 





125 X 10-' 


2^ X 10-^ 


ix 10-^' 


10-^ 


Kultur- 
dauer 


Gew. 0/, 





0.00003 


0.00016 


0.0008 


0.0042 


I 

II 
III 


0.012 
0.012 
0.010 


0.018 
0.025 
0.027 


0.018 
0.018 
0.015 


0.018 
0.015 
0.015 


0.011 
0.015 
0.011 


76 Tage 

7) 



EINIGER ALGEN UND PILZE. 



173 



Ernteertras; in 



H. I. Kulturen mit Zusatz von LiNOj. 

P;-otocüccMS-angestellt 16. April. Zimmer-Temperatur. 



.2 

C5 


Gram ^ 
Mol. 


2^5X10-* 


-1x10- 


lO- 


ixlO- 


Kultur- 
dauer. 


Gew. 0/^ 1 0.00003 


0.00014 


0.0007 


0.0034 


I i 0.010 l 0.020 
II ! 0.009 ' 0.020 
III 0.010 0.018 


0.017 
0.020 
0.016 


0.012 
0.015 
0.011 


0.009 

0.010 
0.008 


24 Tage 

5> 



Ernteertrag in 2;. 



I. I. Kulturen mit Zusatz von KjAsO^. 

B'o/ococcws— angestellt 24. Dec. j^ Treibhaus 10-20° C. 



a 

(V 




Gram 
Mol. 





4x10-^ 


^5x10- 


1x10- 


10- 


Kultur- 
dauer. 


Gew. o/g 





0.00002 


0.0001 


0.0005 


0.0024 


I 

II 
III 


0.011 
0.008 
0.009 


0.015 
0.017 
0.015 


0.012 
0.020 
0.018 


0.011 
0.012 
0.014 


0.008 
0.007 
0.009 


54 Tage 

55 



II. Versuche mit Pilzen. 



Geerntet 18. Jan. 



A. I. Kulturen mit Zusatz von ZnSO^. 
Aspergillus niger — angestellt 24. Dec. '98. 

Kulturdauer 25 Tage. Temperatur 16-20° C. 



Gehalt in 


Ernteertrag 
in g. 


S 


iure 


Säure pro lg. 
Pilzsnhstanz. 


Gram Mol. 


Gew. % 


Acidität 


als Oxalsäure 
umgerechnet 





■ 


0.216 


15.7 


0.495 


2.245 


1 X 10- 


0.003 


0.863 


12.9 


0.406 


0.470 


JxlO- 


0.007 


0.938 


13.2 


0.416 


0.443 


i X 10- 


0.014 


0.944 


12.9 


0.406 


0.430 


10- 


0.028 


0.951 


13.0 


0.409 


0.430 



N.B. Asparagin als N-Quelle. 



174 



N. ONO : WACHSTHUMSBESCHLEUNIGUNG 



Geerntet 20. Jan. 



A. IL Kulturen mit Zusatz von ZnSO^. 
Af^crgillu? niger. — angestellt 24. Dec. '98. 
'99. Knlturdauer 27 Tage. Temperatur lG-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. f. 


Acidität 


als Oxalsäure 
umgerechnet 




1x10-« 

1 X 10-« 

1 X 10-« 

10-« 




0.003 
0.007 
0.014 
0.028 


0.181 

0.868 
0.870 
0.858 
0.821 


14.8 
11.8 
11.1 
12.1 
14.1 


0.467 
0.372 
0.350 
0.381 
0.444 


2.580 
0.428 
0.420 
0.444 
0.541 



N.B. Asparagin als N-Quelle. 



A. in. Kulturen mit Zusatz von ZnSO^. 

Aspergillus niger — angestellt 24. Dec. '98. 

Geerntet 20. Jan. '99. Knlturdauer 2? Tage. Temperatur lG-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. f. 


[Acidität 


als Oxalsäure 
umgerechnet 




1 X 10-« 

1x10-« 

1x10-« 

10-« 



0.003 
0.007 
0.014 
0.028 


0.187 

1.017 
1.336 
1.939 

1.204 


17.8 

12.5 
11.8 
12.5 
11.2 


0.561 
0.394 
0.372 
0.394 
0.353 


3.000 
0.387 
0.278 
0.419 
0.293 



N.B. Asparagin als N-Quelle. 



Geerntet 11. Jan. 



A. IV. Kulturen mit Zusatz von ZnSO^. 
Aspergillus niger — angestellt 20. Febr. 

Kulturdauer 20 Tage. Temperatur 16-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro ^g. 
Pilzsubstanz. 


Gram Mol. Gew. % 


Acidität 


als Oxalsäure 
umgerechnet 




1x10-« 

1 X 10-« 

1 X 10-« 

10-« 




0.003 
0.007 
0.014 
0.028 


0.634 

0.641 
0.635 
0.627 
0.585 


7.2 
7.2 
7.2 
7.2 
7.2 


























N.B. Pextrose anstatt Rohrzucker. 



EINIGER ALGEN UND PILZE. 



175 



Geerntet 25. April. 



B. I. Kulturen mit Zusatz von FeSO^. 
reniclUium jf/au«i?)i— angestellt IL April. 

Kulturdauer 14 Tage. Temperatur 16-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. % 


Acidität 


als Oxalsäure 
unigerechnet 



l X 10-3 
i X 10-3 
10-3 




0.007 
0.014 


0.191 
0.180 

0.233 


3.5 
3.3 
3.9 
3.5 
3.4 
















•2 X 10" 3 n of^R o 181 


1 











N.B. NH^NOi N-Quelle. In allen eisenhaltigen Kulturen waren die Pilzuiassen schön ziegelroth 
gefärbt. Schon bei 0.007 /{ deutliche rothe Färbung bemerklich. 



Geerntet ;">. März. 



C. I. Kulturen mit Zusatz von NiSO^. 

Aspergillus nUja- — angestellt 9. Febr. 

Kiilturdaucr 22 Tafre. 



Temperatur. 16-20=" C. 



Gelialt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. % 


Acidität 


als Oxalsäure 
umgerechnet 



|xl0-3 
l X 10-3 
i X 10-3 

10-3 




0.003 
0.007 
0.014 
0,028 


0.250 

0.297 
0315 
0.401 
0.295 


11.0 

11.1 

10.7 
14.6 
14.0 


0.346 
0.350 
0.337 
0.460 
0.441 


1.464 
1.179 
1.069 
1.147 
1.493 



N.B. NHiNOj als N-Quelle. 



Geerntet 4. März. 



C. II. Kulturen mit Zusatz von XiSO^. 
Aspergillus niger — angestellt 9. Febr. 

Kulturdauer 2.> Tage. Teiii])eratur 10-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg, 
Pilzsubstanz. 


Gram Mol. 


Ge^'. % 


Acidität 


als Oxalsäure 
umgerechnet 



1 X 10-3 
1x10-3 

1 X 10-3 

10-3 




0.0035 
0.007 
0.014 
0.028 


0.288 
0.316 
0.307 
0.387 
0.329 


11.1 

10.7 
11.1 
16.9 
16.7 


0.350 
0.337 
0.349 
0.532 
0.526 


1.216 

1.067 
1.130 
1.375 

1.500 



N.B. NH^NOa als N-Quelle. 



176 



N. 0]>Î0 : WACHSTHUMSBESCHLEUNIGUNG 



Geerntet 6. März. 



C. m. Kulturen mit Zusatz von IsiSO^. 
Aspergillus niger— angestellt 9. Febr. 

Kulturdauer 35 Tage. Temperatur 16-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 


Gram Mol. Gew. % 


Acidität 


als Oxalsäure 
umgerechnet 


Pilzsubstanz. 




1x10- 

1 X 10- 

* X 10- 

10-^ 



0.003 
0.007 
0.014 
0.028 


0.324 
0.310 

0.329 
0.362 
0.341 


9.8 
10.0 
11.9 
15.2 
17.3 


0.308 
0.315 
0.375 
0.479 
0.544 


0.951 
0.016 
1.140 
1.323 
1.695 



N.B. NH4NO3 als N-Quelle. 



C. IV. Kulturen mit Zusatz von XiSO^. 
Aspergillus niger — angestellt 22. April. 



Geerntet 4. Mai. 



Kulturdauer 12 Tage. 



Temiieratur 16-20° C. 



N.B. NH4NO3 als N-Quelle. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. % 


Acidität 


als Oxalsäure 
umgerechnet 



1x10- 
1 V 10^^ 




0.003 
n nn7 


0.262 
0.390 
0.404 
0.364 
0.315 


















1 X 10-^> 0.014 
1 0-3 n no.Q 
















\j.\j^'^ 









Geerntet 4. Mai. 



C. V. Kulturen mit Zusatz von NiSO^. 

Aspergillus niger — angestellt 22. April. 

Kulturdauer 12 Tage. 



Temperatur 16-20° C. 



Gehalt 


in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. % 


Acidität 


als Oxalsäure 
umgerechnet 



1 X 10-' i 
1 X 10- 
i X 10-' 

10-' 



0.003 
0.007 
0.014 
0.028 


0.214 
0.311 

0.300 
0.307 ! 
0.296 




1 































N.B. NH4NO3 als N-Quelle. 



EINIGER ALGEN UND PILZE. 



177 



C. VI. Kulturen mit Zusatz von NiSO^. 
A'tpergiUus niger — augestellt 22. April. 



Geerntet 4. 


Mai. 


Kulturdauer 


12 Tage. 


Temperatur lG-20^ C. 


Gehalt in 


Ernteertrag 

'' in g. 




Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mul. 


Gew. % 


Acidilät 


als Oxalsäure 
umgerechnet 



1 V 1 n— ' 





0.278 
' 0.340 

0.325 
j 0.308 

0.324 


1 








g X ±\J W.^WJ 

1 v> 1 n— 3 <^ rio~ 


1 




5 X lU 

1 V 1 0--^ 


n mj. 


! 




1 n-^ ri no.Q 




l 













N.B. ÎSH4NO3 als N-Quelle. 



D. I. Kulturen mit Zusatz von CoSO^. 
Aspergillus niger — angestellt 17. Febr. 



Geerntet IG. 


März. 


Kulturdauer 


27 Tage. 


Temperatur 16-20° C. 


Gehalt in 


Ernteertrag 
in g- 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. % 


Aeidität 


als Oxalsäure 
umgerechnet 








0.297 


9.0 


0.283 


0.953 


J^xlO- 


0.0017 


0.439 


10.3 


0.324 


0.738 


-I- X 10- 


0.0035 


0.565 


12.3 


0.387 


0.685 


1x10-^' 


0.007 


0.751 


11.2 


0.353 


0.470 


1- X 10-^ 


0.014 


0.872 


8.7 


0.274 


0.314 



N.B. NH4NO3 als N-Quelle. 



Geerntet 20. März. 



D. II. Kulturen mit Zusatz von CoSO^. 
Aspergillus niger — angestellt 17. Febr. 

Kulturdauer 31 Tage. Temperatur 16-20° C 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. % 


Aeidität 


als Oxalsäure 
umgerechnet 




J^xlO- 

ixlO- 

1 X 10-^ 

i X 10-^ 



0.0017 
0.0035 
0.007 
0.014 


0.280 

0.423 
0.582 
0.745 
0.815 


10.0 
12.1 
15.5 
16.7 
8.4 


0.315 
0.381 
0.491 
0.548 
0.265 


1.125 

0.900 
0.844 
0.735 
0.313 



N.B. NHiNO^ als N-Quelle. 



178 



N. ONO : WACHSTHUMSBESCHLEUNIGUNG 



Geeintot 17. März. 



D. III. Kulturen mit Zusatz von CoSO^. 
Aspergillus niger — angestellt 17, Febr. 

Kulturdauer 28 Tage. Temperatur 16-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro Ig. 
■Pilzsubstanz. 


Gram Mol. Gew. % 


Acidität 


als Oxalsäure 
umgerechnet 



-lg X 10- 0.0017 
1 X 10- 0.0035 
1 X 10- 0.007 
1 X 10- ' 0.014 


0.267 

0.393 
0.561 
0.742 
0.770 


10.6 
13.0 
15.5 
14.8 
10.0 


0.334 
0.409 
0.488 
0.466 
0.315 


1.251 
1.041 

0.870 
0.628 
0.409 



(xcerntct ."0. März. 



N.B. NH4NO3 als K-Quelle. 



D. IV. Kulturen mit Zusatz von CoSO^. 
Pénicillium glaucum — angestellt 20. März. 

Kulturdauer 10 Tage. Teniperalur l()-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gfiw. % 


Acidität 


als Oxalsäure 
umgerechnet 




J-xlO- 

1x10- 

l X 10- 

Ï X 10— 




0.0017 
0.0035 
0.007 
014 


0.108 
0.186 

0.225 
0.317 
0.366 


5.4 
5.0 
5.4 
5.9 
5.9 


0.170 
0.157 
0.170 
0.186 
0.186 
















N.B. NHjXOs als N-Quelle. 



Geerntet 1. April. 



D. V. Kulturen mit Zusatz von CoSO^. 
Pénicillium glaucum — angestellt 20. März. 

Kulturdauer 11 Tage. Temperatur 10-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. % 


Acidität 


als Oxalsäure 
umgerechnet 




Axio- 

1x10- 
1x10- 
1 X 10- 




0.0017 
0.0035 
0.007 
0.014 


0.242 

0.469 
0.354 
0.482 
0.772 


4.6 
5.7 
5.9 
5.7 
5.9 


0.145 
0.179 
0.186 
0.178 
0.186 















N.B. Nil 4 NO 3 als N-Quelle. 



EINIGER ALGEN UND PILZE. 



179 



Geerntet 13. April. 



D. VI. Kulturen mit Zusatz von CoSO^. 

Pénicillium rjaucum — angestellt 20. März. 

Kulturdaner 2P> Tage. Temperatur 16-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. % 


Acidität 


als Oxalsäure 
umgerechnet 


! 0.363 

1 vw'in— '5 n AAT7 rv QAQ 


6.4 
6.2 
.5.9 
.5.7 
6.6 


0.202 
0.195 
0.186 
0.179 
0.207 






T6 X J-^ 

ix 10- 
1 v^ 1 n— "> 


0.0035 ' 0.520 

r\ nr\-r r\ an 




1 w 1 n— 3 n AI /1 r> OQO 




^X lU 


v.vxt 


\j,^^<j 





Geerntet 31. März. 



N.B. NH4NO3 als N-Quelle. 



E. I. Kulturen mit Zusatz von CuSO^. 
Asp&'gillus niger — angestellt 21. März. 

Kulturdauer 10 Tage. Temperatur 16-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. Gew. % 


Acidität 


als Oxalsäure 
umgerechnet 




JgXlO- 0.0015 

1 X 10-"^ 0.003 

1 X 10- ' 0.006 

l X 10- 0.012 


0.307 
0.305 

0.297 
0.311 
0.360 


6.2 
5.2 
5.2 
4.7 
5.1 


0.195 
0.164 
0.164 
0.138 
0.160 


0.635 
0.538 
0.552 
0.444 
0.444 



N.B. NH4NO3 als N-Quelle. 



Geerntet 5. April. 



E. II. Kulturen mit Zusatz von CuSO^. 
Aspergillus niger — angestellt 21. März. 

Kulturdauer 15 Tage. Temperatur 16-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. % 


Acidität 


als Oxalsäure 
umgerechnet 



JgXlO- 
1 X 10- 
l X 10- 
1 X 10- 



0.0015 
0.003 
0.006 
0.012 


0.273 ! 9.7 
0.307 10.0 
0.313 10.3 
0.324 10.1 
0.345 1 9.4 


0.309 
0.315 
0.324 
0.318 
0.296 


1.121 

1.003 
1.035 
0.967 

0.858 



N.B. NHiNOg als N-Quelle. 



180 



N. ONO : WACHSTHUMSBESCHLEUNIGUNG 



Geerntet 5. April. 



E. in. Kulturen mit Zusatz von CuSO^. 

Aspergillus niger — angestellt 21. März. 

Kultnrdauer IG Tage. Temperatur 16-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro Ig. 
Pilzsuhstanz. 


Gram Mol. 


Gew. % 


Acidität 


als Oxalsäure 
umgerechnet 



J^xlO-' 
1 X 10-' 
1x10-' 
l X 10-' 



0.0015 
0.003 
0.006 
0.012 


0.218 

0.252 
0.352 
0.358 
0.343 


9.9 
10.1 
9.9 
9.3 
9.6 


0.312 
0.318 
0.312 
0.293 
0.302 


1.431 
1.265 

0.886 
0.818 
0.880 



N.B. NH4NO3 N-Quelle. 



Geerntet 9. Febr. 



F. I. Kulturen mit Zusatz von HgCl„. 
Aspergillus niger — angestellt 1. Febr. '99. 

Kulturdauer 8 Tage. Temperatur 16-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säui-e 


Säure pro lg. 


Gram Mol. 


Gew. % 


Acidität 


als Oxalsäure 
umgerechnet 


Pilzsubstanz. 



J^xlO-' 
1 X 10-' 
|xlO-' 
1x10-' 



0.0017 
0.0034 
0.0067 
0.0135 


0.126 12.6 
0.180 13.3 

! 


0.397 3.176 
0.419 2.328 






i 














' 



N.B. NH^NOa »'s N-Quelle. 0.0017^ gut entwickelt. Sporen braun. 0.0034^ fast keine Eutwickeluna 
0.0067^ u. 0.0135^ keine Eutwickelung. 



Geerntet 9. Febr. 



F. II. Kulturen mit Zusatz von HgCl„. 
Aspergillus niger — angestellt 1. Febr. '99. 

Kulturdauer 8 Tage. Temperatur 16-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 


Grnm Mol. 


Gew. % 


Acidität 


als Oxalsäure 
umgerechnet 


Pilzsubstanz. 




3LX10-' 

1 X 10-' 
1 X 10-' 
1 X 10-' 



0.0017 
0.0034 
0.0067 
0.0135 


0.119 

0.188 


10.0 
12.7 


0.315 
0.400 


2.644 
2.128 














1 



N.B. NH^NOs als N-Quelle. Eutwickelung vie vorige. 



EINIGEK ALGEN UND PILZE. 



181 



Geerntet 9. Febr. 



F. III. Kultnren mit Zusatz von HgCls- 
Aspergillus niger — angestellt 1. Febr. '99. 

Kulturdauer 8 Tage. Temperatur 16-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. f. 


Aciditüt 


als Oxalsäure 
urugerechuet 




T^XIO- 

|xlO-^ 
i X 10-^ 
1 X 10-=> 



0.0017 
0.0034 
0.0067 
0.0135 


0.153 

0.160 


11.3 
13.7 


0.356 
0.431 


2.366 

2.568 










— " ' 1 













N.B. NHiNOa als N-Quelle. Eut Wickelung wie vorige. 



Geerntet 1. Mai. 



F. IV. Kulturen mit Zusatz von HgCl, 

Pénicillium (jlaucum — angestellt 20. April. 
Kulturdauer 11 Tage. 



Temperatur 16-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 


Gram Mol. 


Gew. % 


Aoiflitït ''^'^ Oxalsäure 
Aciüitat umgerechnet 


Pilzsubstanz. 



JgxlO- 
1x10- 
1 X 10- 
1 X 10- 



0.0017 
0.0034 
0.0067 
0.0135 


0.203 

0.243 
0.242 
0.473 
0.251 


4.5 
4.7 
5.1 
5.1 
4.9 


0.142 
0.148 
0.159 
0.159 
0.1.54 











N.B. NH4NO3 als N-OiiiUe. 



Geerntet 2. Mai. 



F. V. Kulturen mit Zusatz von HgClj. 
Pénicillium glaucum — angestellt 20. April. 

Kulturdauer 12 Tage. Temjjeratur 16-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstauz. 


Gram Mol. Gew. /i 


Acidität 


als Oxalsäure 
umgerechnet 




1x10-* 

ix 10-* 

1x10-^ 

lO-* 




0.0003 
0.0006 
0.0013 
0.0027 


0.222 


4.6 0.145 






0.264 ; 4.7 


0.148 
0.145 
0.142 






0.301 


4.5 





N.B. NH1NO3 als N-Quelle. 



182 



N. ONO : WACHSTHUMSBEÖCIILEUNIGUNG 



Geerntet 1. Mai. 



F. VI. Kulturen mit Zusatz von HgCl„. 
Pénicillium glaucum — angestellt 20. April. 

Kulturdauer 11 Tage. Temperatur lö-20° C. 



Gelialt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. fc 


Acidität 


als Oxalsäure • 
iiiugereehnet 


1 0.183 


5.0 
5.5 
.5.9 
5.5 

5.8 


0.157 
0.173 

0.185 
0.173 
0.182 






'S X J-^ 

l X 10-* 
10-* 


0.0006 0.213 
0.0013 I 0.311 
0097 i n Odfi 















Geerntet 18. April. 



N.B. NH1NO3 als N.Quelle. 



F. VII. Kulturen mit Zusatz von HgCl„. 
Aspa-gillus niger — angestellt 30. März. 

Kulturdauer 18 Tage. Temperatur 16-20° C. 



Gehalt in 


1 Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. % 


Acidität 


als OxaLsäure 
umgerechnet 








' 0.347 


8.8 


0.277 


0.800 


.1 X 10-* 


0.0003 


0.517 


9.6 


0.302 


0..586 


ix 10-* 


0.0006 


Î 0.513 


9.4 


0.296 


0.555 


* X 10-* 


0.0013 


0.552 


10.9 


0.343 


0.621 


10-* 


0.0027 


' 0.565 


11.3 


0.356 


0.630 



N.B. NH1NÜ3 als N-Quelle 



F. VIII. Kulturen mit Zusatz von HgCl„. 

Aspergillus niger — angestellt oO. März. 

Geerntet 18. April. Kulturdauer l'J Tage. Temperatur 16-20° ('. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. % 


Acidität 


als üxalsiivire 
umgerechnet 








0.341 


8.6 


0.271 


0.795 


ixlO- 


0.0003 


0.458 


9.4 


0.296 


0.646 


i X 10-* 


0.0006 


0.474 


9.4 


0.296 


0.624 


* X 10-* 


0.0013 


i 0.630 


12.5 


0.394 


0.625 


10-* 


0.0027 


1 0.429 


10.3 


0.324 


0.755 



N.B. NHiNOa als N-Quelle. 



EINIGER ALGEX UND PILZE. 



183 



F. IX. Kulturen mit Zusatz von HgCl, 
Ai^pergi/lus nirjer — angestellt 30 März. 



Geerntet 18. 


April. 


Knltnrdauer 


19 Tage. 


Temperatur 16-20° C. 


Gehal 


f in 


Ernteertrag 
in g. 




Säure 


Säure pro lg. 
Pilzsubstanz. 

1.015 


Gram Mol. 


Gew. % 


Acidität 


als Oxalsäure 
umgerechnet 








0.261 


8.4 


0.265 


1x10- 


0.0003 


0.355 


9.2 


0.290 


0.816 


1 x 10- 


0.0006 


Î 0.380 


9.4 


0.296 


0.779 


h X 10-' 


0.0013 


0..509 


12.1 


0.381 


0.742 


10- 


0.0027 


0.451 


11.1 


0.3.50 


0.776 



Geerntet 18. April. 



X.B. XH^XOs als N-Quelle. 



G. I. Kulturen mit Zusatz von LiNOj. 
Aspei-gilhis ni'(7e/'— angestellt 1. April. 

Kulturdauer 17 Tage. Temperatur lG-20° V. 



Gehalt 


in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. f. 


Acidität 


als Oxalsäure 
umgerechnet 








0.300 


9.0 


0.284 


0.946 


^xlO-^ 


0.004 


0.408 


9.4 


0.296 


0.725 


1x10- 


0.008 


0.428 


8.9 


0.283 


0.661 


1 X 10-- 1 


0.017 


0.348 


8.7 


0.273 


0.784 


1 X 10- ; 


0.034 


0.345 


8.6 


0.271 


0.782 



N.B. NHi NO3 als N-Quelle. 



Geerntet 21. Febr. 



H. I. Kulturen mit Zusatz von NaFl. 
Aspa-gillus niga — angestellt 7. Febr. '99. 

Kulturdauer 14 Tage. Temperatur 16-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. f. 


Acidität 


als Oxalsäure 
umgerechnet 



t^gXIO- 
1x10- 
ixlO- 
1 X 10- 


\ 0.199 8.8 
0.0025 0.325 9.4 
0.005 0.312 7.2 
0.010 0.246 1 6.1 
0.021 j 0.289 6.0 


0.277 
0.296 
0.227 
0.192 
0.189 


1.392 

0.911 

0.727 
0.880 
0.654 



N.B. NH4NO3 als N-Quelle. 



184 



N. ONO : WACHSTHUMSBESCHLEUNIGUNG 



Geerntet 23. Febr. 



H. II. Kulturen mit Zusatz von NaFl. 
Aspergillus niger — angestellt 7. Febr. '99. 

Kulturdauer 16 Tage. Temperatur 16-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. % 


Acidität 


als Oxalsäure 
umgerechnet 



Je X 10-2 
-|xl0-2 
1x10-2 
1x10-2 



0.0025 
0.005 
0.010 
0.021 


0.314 
0.336 
0.385 
0.316 
0.274 


10.0 

10.0 
7.6 
.5.9 
5.Ç> 


0.315 
0.315 
0.239 
0.186 
0.176 


1.000 
0.937 
0.621 

0.589 
0.642 



N.B. NHiKOs .als N-Quello 



Ge-rntet 25. Febr. 



H. 111. Kulturen mit Zusatz von NaFl. 
Aspergillus niger — angestellt 7. Febr. '99. 

Knlturdauer 18 Tage. Temperatur 16-20° C. 



Gehalt in 


Ernteertrag 
in g. 


Säure 


Säure pro lg. 
Pilzsubstanz. 


Gram Mol. 


Gew. f. 


Acidität 


als Oxalsäure 
umgerechnet 



î\xl0-2 
1x10-2 
^xlO-2 
1x10-2 



0.0025 
0.005 
0.010 
0.021 


0.270 
0.280 

0.285 
0.264 
0.265 


8.4 
10.7 
7.7 
6.6 
6.1 


0.265 
0.339 

0.242 
0.208 
0.192 


0.982 
1.207 
0.849 
0.788 
0.725 



N.B. NIIiNOa als N-Quelle. 



EINIGER ALGEN UND PILZE. 



185 



INHALT. 



I. Einleitung und Litteratur. 
II. Methodisches. 

III. Vorbemerkungen über die Versuchsobjekte. 

IV. Veränderungen in der Wachsthumsweise und die Correla- 

tion zwischen Fortpflanzung und Wachsthum. 

V. Einfluss der Eeizstoffe auf die Betriebsstoffwechsel. 
VI. Specielle Besprechungen. 
VII. Schlussbemerkungen und Zusammenfassung der Kesultate. 

TA BELLARISCHE ZUSAMMENSTELLUNG. 

I. Versuche mit Algen. 
IL Versuche mit Pilzen. 



186 



Erklärung der Tafel XIII. 



Kulturen von Aspergillus niger mit und ohne Zusatz von NaFl. 

(Photographiert 15 Tage nacli der Sporenaussaat.) 
I. Ohne Zusatz ; Kontrollkultur. 
II. Mit Zusatz von 0.0025^ NaFl. 

III. Mit Zusatz von 0.005^ „ 

IV. Mit Zusatz von 0.010^ „ 
V. Mit Zusatz von 0.021% „ 

(Für näheres vgl. S. 15.S und ferner Tabelle Pilz. H.) 



H 

^ 

^ 



CO 




Ammonium Amidosulphite. 

By 
Edward Divers :^nd Masataka Ogawa, 

Imperial University, Tokyo. 



The interaction of sucli familiar gases as ammonia and 
sulphur dioxide ceased to attract with any effect the attention of 
investigators sixty years ago and more. Yet comparatively 
nothing had then been definitely made out about the nature of 
the product, and even the few statements concerning it in some 
of the best treatises on chemistry have but little experimental 
foundation. The history of t]\e subject is briefly given on p. 193. 

Non-union of dry sulphur dioxide and ammonia. 

Even when comparatively well-dried, sulphur dioxide and 
ammonia unite at once and with great energy when brought 
together ; yet they can remain mixed without combining, pro- 
vided suflicient care has been observed to exclude moisture. It 
has not been necessary, however, in order to demonstrate this 
striking phenomenon, to have resort to the elaborate precautions 



188 E. DIVEES AND M. OGAWA : 

adopted by Brereton Baker, in his famous experiments upon the 
non-union of hydrochloric acid and ammonia {J. Cli. Soe.y 1894, 
65, Gil ; 1898, 73, 422), and we have only dried the gases 
during their flow through the tubes. As we were able to dry 
sulphur dioxide better than ammonia, because common phos- 
phorus pentoxide could be used for the purpose, we have had 
success in mixing the gases without their combining only by giving 
this gas precedence. The preparation flask with its tubes having 
been heated and then kept for a while in the desiccator, was 
placed in ice and salt while a slow current was sent through it 
of sulphur dioxide, which had passed through drying- tubes of 
sulphuric acid and then of phosphorus pentoxide. The outlet- 
tube dipped into mercury. Ammonia, dried first by the cold of 
a freezing-mixture and then by long tubes of freshly fused and 
crushed potassium hydroxide (but no Stas's mixture), was now 
also passed into the flask slowly. The result was that the in- 
terior of the flask remained clear for some minutes, the mixed 
gases only combining on their escape through the mercury into 
the air. But the ammonia having, it is presumed, gradually 
brought enough moisture with it through passing more rapidly 
along the tubes than at first, the walls of the flask became sud- 
denly coated with an orange-coloured deposit, while the mercury 
rose high in the exit tube. 

Proportions m which sulphur dioxide mid ammonia combine. 

The proportions, in which ammonia and sulphur dioxide 
combine, or appear to combine together, depend largely upon 
the extent to which the temperature is allowed to rise, tlie heat 
of union being considerable. They vary also according as one 



AMMONIUM AMIDOSULPHITE. 189 

or Other of the gases is used in excess, unless the temperature is kept 
very low. But the variation of the proportions and the apparent 
condensation of additional sulphur dioxide by a sufficiently am- 
moniated product, that may be observed, are results clearly due 
to the secondary changes going on (p. 192). The simple union 
of ammonia and sulphur dioxide, which can be secured by 
keeping down the temperature by suitable means, especially with 
the ammonia in excess, is that of two volnmes of the former to 
one of the latter (p. 191). But since this union cannot be 
made at the ordinary temperature without being immediately 
followed by a decomposition, in which ammonia is evolved, the 
union of the two gases can appear to take place in other pro- 
portions than the above. It is pretty certain that, by proceeding 
slowly enough and using strong cooling agencies, secondary ac- 
tion could be almost entirely prevented and the statement just 
made be verified, even when working with the gases alone. We 
have not gone very near to getting such a result in this way, 
but then we have, for good reasons, not striven much to over- 
come the difficulties. Our experimental work, which will be 
further on referred to (p. 195), has shown that two much more 
nearly than one volume of ammonia can be made in this way 
to unite with one volume of sulphur dioxide, the only propor- 
tions which Rose met with in his experiments (p. 193), and that 
the presence of much ammonium amidosulphite in the j^i'oduct 
can be established with certainty. 

Preparation and analysis of ammonium amidosulphite. 

In order to get the primary product of the union of sul- 
phur dioxide with ammonia in its unchanged state, ether was 



190 E. DIVERS AXD M. OGAWA : 

made use of as the medium of the union, in order to keep the 
temperature under control. The ether, freed from alcohol and 
water by sodium, was contained in a small flask, fitted with 
inlet and outlet tubes, which was to serve, not only for the pro- 
duction of the new substance, but for its isolation and its weigh- 
ing for analysis. The flask was put in a bath of ice and salt, 
with the outlet-tube dipping into ;i trough of mercury, and then 
the ether was saturated with dried ammonia. Having shut off' 
the ammonia, a very slow current of sulphur dioxide was sent 
into the solution while the flask was continuously shaken, not 
only in order to diffuse the heat, but to prevent the product 
from caking on to the bottom of the flask and shutting in ether. 
The mouth of the tube conveying the sulphur dioxide soon 
became filled with a yellow pasty mass (p. 102), and had to be 
kept open by a platinum rod, manipulated through the rubber 
tubing above, but the precipitate itself was quite white and pow- 
dery. In spite of the external cooling, the heat of combining 
was sufficient to cause ammonia gas, saturated with ether-vapour, 
to escape through the mercury sealing the exit-tube, and when 
this escape became slight, the passage of sulphur dioxide stopped. 
With the use of about 20 c.c. ether, there had then formed well 
over a gram of the substance. In order to secure this undecom- 
posed, a second flask was put in connection with the preparation- 
flask, and ammonia again passed to the saturation point. The 
ammoniated ether was decanted off through the connecting-tube 
into the second flask, which was then detached, the whole opera- 
tion being carried out in the freezing- mixture. The current of 
ammonia was renewed over the ^precipitate in the flask, and 
continued for hours, until all the ether adhering to the precipi- 
tate had been carried away, the flask being all the while still 



AMMONIUM A3IID0SULPHITE. 191 

in the freezing-mixture. There was no other way of completely 
drying the salt, and even this way was not sufficiently successful 
when the salt had been allowed to cake together. The ammonia 
could not be replaced by air or hydrogen for drying the salt, 
nor could the flask be kept out of the freezing-mixture, so long- 
as ether still moistened the salt, without the latter taking an 
orange-colour. When dry and in an ammouical atinosi^here, 
the salt is more stable, but cannot long l)e kept at the ordi- 
nary temperature without getting discoloured through decom- 
position. 

Analysis. — The stopper carrying the gas-tubes having been 
replaced by a plain one, and air allowed to displace most of the 
ammonia gas, the flask was at once weighed and left for a time 
inverted with open mouth dipping into 100 c.c, or more, of 
water in a beaker. When the salt in it had become damp, it 
was washed into the water, and its very dilute solution distilled 
with alkali for its ammonia. The residue was divided into two 
measured portions, one of which was acidified and heated to 
150° under pressure for some hours and then redistilled with 
alkali for additional ammonia, of which only a trace was got 
(0.001 per cent, of the salt). The other part of the solution was 
treated with bromine, and next with hydrochloric acid and 
chlorate, after which barium sulphate was precipitated with the 
usual precautious. The results of the analysis were : — 

Found : 

S0,(NH3)o: 

The slight excess of ammonia indicated is safely attributable 
to the means taken to preserve the salt till it was analysed. 



Ammonia 


Sulphur diox. 


35.09 ; 


64.91 per cent, 


34.69 ; 


65.31 „ 



192 E. DIVERS AND M. OGAWA : 

Its properties, constitution, and name. 

The new salt is white and apparently crystalline, and ap- 
pears to be slightly volatile in a current of ammonia. It is 
very deliquescent and decomposes, losing ammonia, in the air. 
It dissolves in water, giving out heat and a hissing sound, and 
if dissolved by ice or enough ice-cold water, furnishes a solution 
answering all tests for pure ammonium sulphite. In this respect 
it is quite unlike ammonium amidosulphate or carbamate, since 
even the latter salt gives at first no precipitate with calcium 
chloride, which at once precipitates all sulphite from the new 
salt. When the salt is much decomposed, its solution gives 
other reactions besides those of a sulphite. In anhydrous alcohol 
it dissolves freely, evidently as ethyl ammoniumsulphite ; it is 
also slightly soluble in dry ether. It soon begins to change and 
then assumes an orange-colour, even at the common temperature. 
At 30-o5° it decomposes into a liquid and a solid part, both 
more or less orange-coloured, and into ammonia, the liquid part 
undergoing further change into solid matters (p. 197) 

Constitution. — The salt is more probably an amido- than an 
imido- compound, NH4N(S02NH4)2 (analogue of normal ammo- 
nium imidosulphate), because it can be obtained only when the 
temperature is kept down and the ammonia is in excess. It is 
still more probably a sulphuryl rather than a thionyl compound, 
because of its feeble activity as a reducing agent and of its very 
easy passage into ammonium sulphite or ethyl ammoniumsulphite. 
It has accordingly to be formulated as NH2"SOo*NH4, and not 
NHo-SO-ONH,. 

Name. — Since the salt represents ammonium sulphite» 
NIl40*S02*NH4, in which the ammonoxyl is replaced by amido- 



AMMONIUM AMIDOSULPHITE. 193 

gen, it is properly called ammonium aniidosulphite. Berglund's 
name of amidosulphonate now in use, for amidosulpliate is 
evidently based on a misconception. The name, amidosulphinate, 
in analogy with amidosulphonate, must be rejected on the same 
grounds, and because the salt has not the characteristic reducing 
action and the constitution of sulphinates. It does not seem 
possible, even were it desirable, to construct a term for the first 
amide of sul2:)hurous acid that would correspond to that of sul- 
phamic acid, the synonym of amidosulphuric acid. 

Nature of the decomposition by heat of the aynidosulphite. 

History. — Experiments made earlier than ours on the union 
of sulphur dioxide with ammonia gave the products of decom- 
position of ammonium amidosulphite instead of the salt itself. 
Doebereiner in 1826 (Schw, Jahrb., 17, 120), described the pro- 
duct of the union as a brown-yellow vapour quickly condensing 
to a bright brown solid mass, which the smallest quantity of 
water converts into (colourless) ammonium sulphite. Rose 2:)ub- 
lished three papers on ' anhydrous sul^^hite of ammonia ' in 
1834, 1837, and 1844 {Pogg. Ann., 33, 235; 42, 41 Ö ; 61, 397), 
in the second correcting statements made in the first, and 
modifying in the third the views he had expressed in the earlier 
papers. The outcome was that he had ascertained that the pro- 
duct of the union is always one and the same single substance, 
in whatever proportions the dry gases are taken ; that it is 
composed of equal volumes of the gases, is either yellowish-red 
and smeary, or red crystalline, very deliquescent and very 
soluble in w^ater without evolving ammonia ; that it yields a 
neutral solution, which is at first yellowish but soon, becomes 



194 E. DIVEES AND M. OGAWA : 

colourless, and gives, when recently prepared, the reactions 
mainly of a mixture of ammonium sulphate and trithionate, but 
to a small extent those of a sulphite also ; and, lastly, that when 
the solution is of certain concentration it gives a transient red- 
dish coloration with hydrochloric acid. 

Forchhammer {Comyt. reiid., 1837, 5, 395) found that, be- 
sides the orange- coloured substance, crystals of ammonium sulphate 
are produced by the union of the gases, which can sometimes be 
seen apart from the other product in some spots of the mass, 
though often indistinguishably mixed up with it. (That the 
crystals observed in the product were those of sulphate, could 
only have been a supposition of Forchhammer's). The mass 
when moistened is alkaline and evolves ammonia, yielding other- 
wise the reactions recorded by Rose. Absolute alcohol dissolves 
out of it a substance which takes a rose colour, soon disappear- 
ing. Indirectly, he represented the mass to be derived from two 
mois, ammonia to one mol. sulphur dioxide, as did also Doe- 
berenier. 

The views advanced as to the nature of the orange body 
have been, that it is a compound of ammonia with an isomer 
of sul^^hurous anhydride, which changes at once with water 
into ammonium sulphate and trithionate, just as ammonium pyro- 
sulphite slowly changes in hot solution (Kose) ; that it is amido- 
gen sulphide, S(NH2)2, mixed with ammonium sulphate (Forch- 
hammer) ; that it is, partly, thionamic acid, NHs'SO'OH, partly, 
ammonium thionamate, both volatile, being its colour due to 
impurity (H. Watts) ; and that it is ammonium pyrothionamate, 
NHo'SaO^-NHj (Joergenssen). 

Interaction of the gases. — We have repeated Hose's experi- 
ments of measuring over mercury the volumes of the gases which 



AMMONIUM AMIDOSULPIIITE. 195 

interact, in which he found that always equal volumes combiue, 
whichever gas may be taken in excess. The results somewhat 
approached this when no steps were taken to restrain the rise in 
temperature due to the union of the gases ; but when the gas- 
tube was immersed in a cooling-mixture and the ammonia was 
in excess, the volume of this gas consumed was much greater 
than that of the sulphur dioxide. This method of investigating 
the matter is, however, inapplicable, because the ammonium 
amidosulphite, which is formed, partly decomposes with free 
evolution of ammonia. By letting the dried gases come together 
in a vessel agitated in a freezing-mixture and keeping the am- 
monia in excess, a solid mass is obtained which consists largely 
of the amidosulphite, behaving as such in water, though mixed 
with other substances, and quantitative analysis of which shows 
that much more thiin three mois, ammonia to two mois, sulphur 
dioxide have gone to its formation. If, instead of examining it 
at once, it is kept for a long time in a gentle current of dry 
nitrogen or hydrogen, at a temperature of 30° to 35°, it no longer 
contains amidosulphite or gives any sulphite to water, and contains 
not much more than one atom of nitrogen to one of sulphur. 
Thus, Rose's results are explained and, at the same time, shown 
to be of no direct significance. 

Products of the decomjjosition. — Both Hose and Forchham- 
mer found ammonium sulphate to be a principal constituent of 
the product of the interaction of the gases. A sufficiently high 
temperature having been reached, this will have been the case ; 
furthermore, the solution of the even less heated product slowly 
becomes acid and full of sulphate. But when the temperature 
has not been allowed to exceed 30°, or even 40°, the quantity of 
sulpliate in the product is so small that it may almost be dis- 



196 E. DIVERS AND M. OGAWA : 

regarded. Along with sulphate, trithionate was considered by- 
Rose to make up most of the product, for the aqueous solution 
of the mass always gives a strong reaction with silver nitrate 
which might be that of trithionate and, in the case of his pro- 
duct, gave other reactions of a trithionate. But when the pro- 
duct has been carefully prepared and is free from amidosulphite, 
its solution gives the silver reaction without the others belonging 
to a trithionate. Thus, the solution may be acidified and left 
for hours without yielding more than mere traces of sulphur 
dioxide and sulphur ; to get these in quantity, the solution had to 
be strongly heated under pressure. Besides this, the absence of 
sulphate in the solution is of itself almost enough to disprove 
the production of trithionate, since, as Hose himself represented, 
sulphate and trithionate are complementary products of the de- 
composition. 

Heating pure ammonium amidosulphite gives the same re- 
sults as heating the coloured product of the union of sulphur 
dioxide and ammonia as gases. Rose's assertion that the product 
is a single substance, even in appearance, is certainly incorrect, 
according to our experience. By the union of the gases in a 
receiver kept well cooled, the product is deposited as a soft, 
waxy, yellow coating on the walls of the vessel and on the gas 
tubes. Its colour varies in different parts from nearly white to 
orange-red somewhat irregularly but generally so as to be whiter 
near where the ammonia enters, the whiteness not being due to 
moisture in the gases, as Rose assumed. When the product gets 
to 30-35°, whether by its own heating or by external heat, it is 
decomposed at first into an obscurely crystalline white solid and 
a much smaller quantity of a coloured, effervescing liquid, partly 
draining to the bottom of the vessel ; after a time all becomes 



AMMOXIUM AMIDOSULPHITE. 197 

solid again and tenaciously adherent to the glass. When i)ure 
ammonium amidosulphite is similarly heated in a dry inactive 
gas, it colours, softens, shrinks together, vesiculates, gives out 
ammonia, and becomes a mass like that derived directly from 
the union of the gases. With very gradual heating, the tem- 
porarily liquid product is much less coloured than in the other 
case, its colour being evidently caused by the presence of a 
red matter dissolved in it, wliicli gives indications of being- 
volatile. 

This orange-red substance is never formed in more than 
very small qua.ntit3^ It gives a yellow colour to the aqueous 
solution of the whole product, which, however, slowly fades away. 
Alcohol, carbon bisulphide, and other menstrua dissolve it out 
from the salts, leaving them white ; but the solutions are not 
pure. The yellow solution in water or alcohol takes a transient 
pink colour Avhen mixed with dilute hydrochloric acid, and the 
alcoholic solution an indigo-blue colour with concentrated am- 
monia. The residue left on evaporating the carbon- bisulphide- 
solution becomes explosive when heated above 150°, and may 
then have become nitrogen sulphide, but before being heated it 
is not this substance. 

Except the very little sulphate already mentioned, there is 
no as-yet known substance present in the residue of the decom- 
position of the amidosulphite by a gentle heat, so far as we can 
discover. Alcohol of 00 per cent, dissolves out something, but 
only very sparingly. By evaporation of the solution in a vacuum 
desiccator, a very deliquescent salt is obtained in crystals, having 
a composition that may be expressed by 9NH3, 8SO2, assuming 
the presence of 2.5 per cent, moisture. The composition of the 
whole crude residue does not differ much from this. The alco- 



198 E. DIVERS AXD M. OGAWA : 

liolic solution, cooled and charged with ammonia, gives minute 
scaly crystals in small quantity. This substance, dried in a 
current of ammonia, has a composition expressed by (NH3)3S203, 
nnd dried in the sulphuric-acid desiccator, that of (NH3)2S203. 
These three substances all give the silver-nitrate reaction of the 
aqueous solution of the whole residue, and on boiling with dilute 
hydrochloric acid give very little sulphur and no sulphur dioxide. 
At higher temperatures, whether dry or in solution, they yield 
sulphur, sulphur dioxide, and sulphate. Two potassium deriva- 
tives of these salts have also been prepared. Neither the crude 
residue nor any of the above substances yields all its nitrogen as 
ammonia when distilled with alkali, unless it has been first 
heated with hydrochloric acid under pressure. 

From the mother-liquor of the above mentioned S2O3 salt a 
substance was got which in composition and behaviour appeared 
to be sulphamide a little impure. Neither sulphamide nor amido- 
sulphate can be found in the fresh aqueous solution of the whole 
residue, but, by heating the solid residue itself to a higher 
temperature, imidosulphate is obtained in considerable quantity, 
besides sulphur and sulphate, and imidosulphate is a known 
product of first heating and then dissolving in w^ater, either 
amidosulphate or sulphamide. A proof-spirit extract and also a 
wood-spirit extract of the residue yield ammonium amidosulphate 
on evaporation, no doubt generated by hydration. An aqueous 
solution of the less heated residue, treated with excess of barium 
acetate and filtered, gave barium thiosulphate in crystals, on 
evaporating it over the water-bath. 

During the heating of ammonium amidosulphite at a tem- 
perature of 30° to 35°, besides much ammonia, small quantities 
of water and of sulphur dioxide are evolved, the former mainly 



am:moxium amidosulphite. 199 

in tlie early stage and the latter in the late stage of the de- 
comiDosition. This remarkable production of water, though 
always evident, was fully established by cooling the escaping 
gases and testing the water thus collected. The presence of 
sulphur dioxide later in the ojîeration was shown by the gases 
fuming on their escape into the air and then forming a small 
white deposit, slowly turning orange, and reacting as ammonium 
pyrosuljihite. In the interaction of sulphur dioxide with ammo- 
nia, and in the decomposition of the amidosulphite, no liberation 
of nitrogen could ever be discovered. 

To sum up the results of our incomplete work upon the 
decomposition of ammonium amidosulphite by a graduated and 
gentle heat, ammonia and a residue consisting of a substance 
(or substances), which behaves as a thio-amido-sulphouic com- 
pound, are the principal products ; in much less quantities, water 
and an orange-red substance are also produced, and, generally if 
not always, a very little sulphate ; and, as secondary products, 
apparently sulphamide and certainly amidosulphate and thiosul- 
pliate are obtainable, as well as imidosulphate, sulphur, and 
much sulphate. It seems of interest to point out that we here 
record the first production known of amidosulphate from am- 
monia and sulphur dioxide, which, hitherto, has been derived 
either from ammonia and sulphur trioxide or from a nitrite and 
sulphur dioxide. 

We hope in a future paper to be able to report the com- 
pletion of this investigation. 



Products of heating Ammonium Sulphites, 
Thiosulphate, and Trithionate. 



By 



Edward Divers and Masataka Ogawa, 

Imperial University, Tokyo. 



What lias been published upon the effects of heating am- 
monium sulphites and thiosulphate is but little in accordance 
with the results of experiments we have had to make upon these 
salts and upon the hardly known trithionate, in connection with 
an investigation of the decomposition by heat of ammonium 
araidosulphite. We therefore make known what we have ascer- 
tained. 

Preparation of the salts used. 

Ammonium sulphite, (NH4)2 SOo, OHo. — Statements are con- 
flicting as to whether this salt can be got from its solution by 
evaporation (Muspratt, Phil. Mag., 1847, iii, 30, 414 ; Marignac, 
Jahresb., 1857, 17 ; Forcrand, Compt. rend., 1885, 100, 245 ; 
Hartog, Compt. rend., 1887, 104, 1793 ; Eoehrig, J. pr. Ch., 1888, 
37, 227). We find that a concentrated solution, charged with 



202 E. DIVERS AND M. OGAWA : PEODUCTS OF HEATING 

ammonia, can be quite successfully made to deposit the salt by 
cold evaporation in a potash desiccator, but to get such a solution 
the moderately strong solution of ammonia, which must be used, 
has to be kept very cold while passing in the sulphur dioxide. 
Dilute solutions fail to yield the salt on evaporation because too 
much of it suffers decomposition. Much better than evaporating 
is to take advantage of the lessened solubility of the salt in 
presence of much ammonia. Ammonia solution, sp. gr. 0.895, 
containing therefore about 28 grm. ammonia in 100 c.c, is to be 
treated in a flask with sulphur dioxide, while it is kept in mo- 
tion in a mixture of ice and salt, and with the tube conveying 
the sulphur dioxide not dipping into the solution. The formation 
of a very little orange-coloured matter in the neck of the flask 
cannot be avoided, but this can be easily removed afterwards. 
When the solution has become thick with crystals, no more 
sulphur dioxide is to be added, although very much ammonia 
still remains. Even at the common temperature the crystals do 
not sensibly dissolve in presence of this ammonia. The salt, 
drained on a tile under close cover, can be dried either by 
filter paper or by only short exposure in the desiccator over 
potassium hydroxide or carbonate, salted just before with am- 
monium chloride. It is equivalent in quantity to about one- 
fourth of the ammonia taken. By long exposure in a dried 
atmosphere the salt becomes anhydrous without loss of ammonia. 
Exposed to the air, it is apparently deliquescent but in reality 
it evolves ammonia and thus becomes the very deliquescent 
pyrosulphite. 

Anhydrous aminonium sulphite is readily obtained from the 
hydrated salt by long enough exposure in the desiccator ; it is 
very hygroscopic. 



AMMONIUM SULPHITES, THIOSULPHATE AND TRITHIONATE. 203 

Äinmonium pyrosulphite, (NH4)2S205. — When, in the process 
just given for preparing the normal sulphite, the passage of 
sulphur dioxide is not stopped when the solution is full of crystals, 
these gradually dissolve up and the solution becomes greenish- 
yellow. Then, as it gets charged with sulphur dioxide, in the 
cooling mixture, the pyrosulphite crystallises out from it, in quan- 
tity equivalent to a little over one-fifth of the ammonia taken, 
being thrown out of solution by the sulphur dioxide. The salt 
can be obtained dry and pure in the same way as the normal 
sulphite, except that sulphuric acid, to which a little solid alkali 
sulphite has been added, is used in the desiccator, though it is 
very deliquescent and changeable when not carefully preserved 
from moisture. This salt is also easily obtainable by evapora- 
ting its aqueous solution, but hardly free from sulphate, and not 
without some decomposition, through loss of sulphur dioxide and 
through oxidation. It is much more soluble than the normal 
sulphite. 

Ammonium thiosulphate. — An old solution of calcium thio- 
sulphate, obtained by boiling lime and sulphur together in water 
and leaving the solution until much of the pentasulphide had 
been oxidised by the air, was decanted from insoluble matters, 
mixed with ammonium carbonate in some excess, filtered, and 
then freely exposed to the air for some time at 50-60°. In this 
way a very concentrated solution of ammonium thiosulphate was 
obtained, free from sulphate and other salts. The solution of 
this very soluble salt was then dried up to a crystalline mass in 
the desiccator. The well-dried crystals have been found by Lock 
and Kluess {Ber., 1889, 22, 3099) to be anhydrous. 

A?nvioiiium trithionate. — This salt has apparently not hitherto 
been prepared by any one. Being exceedingly soluble in water. 



204 E. DIVEE8 AND M. OGAWA : PKODUCTS OF HEATING 

it cannot be prepared by Plessy's excellent process for the po- 
tassium salt (Ann. Ch. Phys., 1844, iii, ii, 182), or by its 
slight modification by Hertlein (Z. phys. Ch., 1896, IQ, 287). 
We therefore made the pure potassium salt by Plessy's method, 
precipitated the potassium from it by hydrofluosilicic acid, neu- 
tralised quickly with ammonia, and precipitated the ammonium 
trithionate by absolute alcohol and dried it in the desiccator. 
This very deliquescent and changeable salt cannot be kept long 
in good condition, but it was used by us when freshly prepared 
and while still almost free from sulphate. 

Effects of heating the salts. 

The process. — The salts were heated in an oil-bath, in a 
subliming vessel consisting of a test-tube, 15 cm. long and about 
15 mm. in internal diameter. The tube was closed by a caout- 
chouc stopper, and a very slow current of dried nitrogen through 
the tube was maintained during the heating and cooling. The 
salt, usually about 4 grm., was contained in an open slender 
bottle, about 6 cm. long, having a platinum wire attached to it 
for lowering it into and lifting it out of the subliming tube. The 
tube was immersed in the oil to the level of the mouth of the 
bottle inside, so as to cause all dry sublimates to collect in the 
tube above this level. When, as in the case of the hydrated 
normal sulphite, the heating was divided into stages, the bottle 
was transferred between these to a second subliming-tube. The 
heating of the oil was conducted very slowly, so that the tem- 
peratures mentioned which were those of the oil, may be accept- 
ed as being very nearly those of the salts at the time. 

In describing the effects of heating them, the salts are taken 



AMMONIUM SULPHITES, THIOSULPIIATE AND TRITHIONATE. 205 

in the inverse order of that followed above, in accordance with 
usage. This is done because of the nature of the products. 

Ammonium trithionate. — This salt is hardly affected until the 
temperature is above 150°, and at 160-170° it steadily decom- 
poses into sulphur dioxide and a residue of ammonium sulphate 
and unfused sulphur. The non-fusion of the sulphur is remark- 
able and only to be referred to the presence of minute quantitities 
of impurities. It all dissolved readily in carbon bisulphide, and 
crystallised out on evaporating the solvent. 

It can hardly be doubted but that ammonium tetrathionate 
(and peniathionate, if it can exist) would decompose in the same 
way as trithionate. Ammonium hyposulphate (dithionate) has been 
shown by Heeren {Pogg., 1826, 7, 55), and more definitely by 
Kluess {Ann., 1888, 246, 194) to first become anhydrous, if not 
already so when heated, and then to decompose at about 130° 
into sulphur dioxide and a residue of ammonium sulphate. 

Anwioniuvi thiosulphaie. — Zeise, in 1824 {Gm. Hbh) found 
this salt to be converted by heat into water, ammonia, and a sub- 
limate of sulphur, much thiosulphate again and sulphite, and a 
little sulphate. This result must have been obtained by rough 
heating. A much more weighty statement is that made by 
Spring {Ber., 1874, 7, 1159), namely, that the dry salt can be 
sublimed unchanged, intermediate dissociation being admitted. 
We have found it to decompose very slowly at 150°, the main 
products being a sublimate of anhydrous normal sulphite and 
a residue of sulphur unfused, as in the case of the trithionate. 
But, also very small quantities of hydrogen sulphide and am- 
monia passed off in the current of nitrogen, and the sublimate 
contained a very little of a salt having some of the properties 
of trithionate and which did not strike the violet colour with 



206 E. DIVERS AND M. OGAWA : PRODUCTS OF HEATING 

ferric chloride given by u thiosulphate. Analysis of the subli- 
mate and of that part of the salt which remained mixed with 
the sulphur when the progress of the decomposition was arrest- 
ed after only half of it had been decomposed, gave results that 
showed the former to be essentially anhydrous normal sulphite, 
and the latter unchanged thiosulphate : — 





Ammonia 


Sulphur 


(ßüußO, 


29.31 


27.59 per cent 


Sublimate 


27.54 


27.55 „ 


(NH,),SA 


22.27 


43.24 „ 


Residue 


20.69 


42.31 „ 



The main decomposition of the thiosulphate is in full agree- 
ment wit1i the relation of thiosulphates to sulphites. Very in- 
teresting is the production of a little ammonia and hydrogen 
sulphide, in connection with the relation of trithionate to thiosul- 
phate as its thio-anhydride (Spring) :— 2(Nfi,)2S203=2NH3 + 
SH2 + (NH4)2S30fi. When ammonium thiosulphate is rapidly and 
more strongly heated, ammonia is lost and sulphur sublimes; then 
as a matter of course and of no significance, thiosulphate and 
even trithionate are produced on adding water to the mixed 
sublimates. 

Ammonium pyrosulphiie. — We did not get this exceedingly 
deliquescent salt into the tube ready for heating before it had 
condensed some moisture, and to this we attribute part of the results 
obtained. Change went on slowly in the salt at 130° and some- 
what faster at 150°. At first there was little else than a slight but 
steady evolution of sulphur dioxide, and this continued though very 
feebly, to the end and while a sublimate forming. The sublimate 
was pyrosulphite in one experiment ; in another, it was this salt 
mixed with a very little anhydrous sulphite. But there was a 



AMMONIUM SULPHITES, TIIIOSULPHATE AND TRITHIONATE. 207 

considerable residue, more than one-tliird of the weight of the 
salt taken, consisting of sulphate, trithionate, sulphur, and ap- 
parently some tetrathionate. There was no sulphite or thiosul- 
phate. The tetrathionate, the sulphur, and the sulphur dioxide 
were very probably derived from decomposition of trithionate 
by moisture. From a consideration of the results it seems almost 
necessary to assume that perfectly dry pyrosulphite sublimes un- 
changed (witli no doubt intermediate dissociation), and that the 
presence of a little moisture causes it to decompose partly into 
sulphate and trithionate. 

Anhydrous ammonium sulphite volatilises at about 150°, 
yielding a sublimate of the same salt, or rather, a pseudosub- 
limate, for the salt surely dissociates when heated. 

Hydrated ammonium sulphite. — According to Muspratt, this 
salt all volatilises when heated, no sulphate being produced, and 
yields water, then much ammonia, and finally a sublimate which, 
judging, from its properties, is ammonium pyrosulphite. We 
observed the following effects of gradually heating it in a very 
slow current of dried nitrogen. At about 90°, the salt moistened 
and escape of ammonia became quite evident, and at a little 
above 100° distillation of water also took place ; both water and 
ammonia continued to escape in noticeable quantities for 2Y2 hours 
longer, when the temperature for some time had been 120°; up 
to this, a very little sublimate only had formed and matters were 
now almost at a standstill. The quantity of the salt heated was 
about 4 grm., and this had now lost one-fifth of its weight, the 
residue having the composition expressed by (]SrH3)io(SO..)G(OH2)7, 
equivalent to a mixture or combination of the three salts, hydrat- 
ed sulphite (39.49^), anhydrous sulphite (34.1^), and pyrosul- 
phite (26.59^), dividing equally among themselves the sulphur 



208 E. DIVEES AjSTD M. OGAWA : PRODUCTS OF HEATING 

dioxide. Some repetitions of the experiment gave almost the same 
results. Calculation and the results of one experiment gave the 
following numbers : — 





Ammonia 


Sulphur dioxide 


(NH3\o(SO,\(OH,), 


25.00 


56.47 per cent 


Found 


24.6.5 


56.20 „ 



If, in the formation of this complex, no longer losing 
material quantities of ammonia and water, only these products 
had been given off, the residue should have been 8472 per cent, 
of the hydrated normal sulphite, whereas it proved to be little 
more than 79 per cent., in consequence of volatitisation of some 
of the (dissociated) salt, made manifest by the production of a 
little sublimate. 

After renewing the heating in a fresh subliming-tube, allow- 
ing the temperature to rise slowdy from 120° to 150°, the residue 
had almost all disappeared in two hours, while an abundant dry 
sublimate had deposited. For some time during this heating, 
sulphur dioxide steadily escaped, but practically ceased to do so 
long before sublimation was finished. The residue left when 
sulphur dioxide was no longer coming oft', proved on analysis to be 
normal sulphite again, but only half hydrated, 2(NH4)2S03, OH2. 
The sublimate, also, now and at the finish, consisted of 
normal sulphite, apparently anhydrous though found to be a 
little hydrated because it is very hygroscopic and had unavoid- 
ably some exposure to the air while it was being scarped out of 
the tube into the weighing bottle. 

Hydrated ammonium sulphite, therefore, becomes by gra- 
dual heating to 120° converted one-third into the anhydrous 
salt, and one-third into pyrosulphite, by loss of w^ater and am- 
monia ; and then the nearl}^ stable complex of these salts with 



AMMONIUM SULPHITES, THIOSULPHATE AND TRITHIONATE. 209 

the other third of the original salt becomes converted into the 
nearly anhydrous normal sulphite, between 120° and 150°, sul- 
phur dioxide and water escaping. The presence of water is es- 
sential to the occurrence of both changes ; dry ammonium 
pyrosulphite partly sublimes as such at 150° and partly changes 
into sulphate and trithionate, as already described. Heating in 
the open tube, and more rapidly, Muspratt's results will be got, 
for then weter is more quickly expelled, and some pyrosulphite 
can deposit as a sublimate. 



Potassium Nitrito-hydroximidosulphates and the 
Non-existence of Dihydroxylamine Derivatives. 

By 

Edward Divers, ^r. D., D. Se, F. R. S., Emeritus Prof., 

and 
Tamemasa Haga, D. Se, F. C. S., 

Professor, Tokyo Imperial University. 



Like potassium nitrate (this Journal, 7, 56), potassium 
nitrite forms double salts with the potassium hydroximido- 
sulphates (sulphonates), the non-recognition of whose existence 
has allowed mistaken notions to arise about the nature and the 
products of the sulphonation of nitrous acid. 

Potassium nitrite and 2/3 normal hydroximidosulphate, KNOo, 
HON(S03K)2. — The sparing solubility of 2/3 normal potassium hy- 
droximidosulphate in water is hardly affected by the presence of 
potassium nitrite and when a sufficient quantity of the salt has 
been dissolved by heat it crystallises out again almost pure on 
cooling the hot solution, even though the water has also dissolved 
in it as much as one-sixth of its weight of the nitrite. When 
the solution of the nitrite is stronger than this there crystallises 



212 E. DIVERS Aîs^D T. HAGA : 

out instead of the liydroximidosulpliate itself a combination of it 
with a molecule of the nitrite. The same double salt is also 
formed in the cold when the liydroximidosulpliate is triturated 
and digested with such a solution of the nitrite. Precautions 
being taken against the hydrolysis of the unstable hydroximido- 
sulphate this salt can be dissolved at 70° in as little as 3.8 times 
its weight of a 22 per cent, solution of nitrite and by cooling the 
solution the double salt be got in crystals iu quantity equivalent 
to about 12/13 of that of the hydroximidosulphate. 

While the hydroximidosulphate itself crystallises in hard 
rhombic jmsms with 2OH2, its compound with the nitrite is in 
silky asbestus-like fibres which are anhydrous. The compound 
salt is also not deliquescent although potassium nitrite alone is 
very deliquescent. There is nothing else in its properties where- 
by to distinguish it from a mixture of its component salts. It 
can be recrystallised from a hot solution of potassium nitrite of 
a strength of 10 per cent, or more nitrite. It is neutral to lit- 
mus and very soluble in water but its solution soon deposits 
crystals of the 2/3 normal potassium hydroximidosulphate unless 
it is very dilute. In any case the hydroximidosulphate can be 
precipitated and thus separated from the nitrite by the addition 
of barium hydroxide. Like a simple hydroximidosulphate (this 
Journal, 7, 40), the solid salt digested with a highly concen- 
trated solution of potassium hydroxide is converted into sulphite 
and nitrite. When acidified its solution becomes yellowish for 
a short time and then effervesces from the escape of nitrous 
oxide, a result of the hydroximidosulphate being a sulphonated 
hydroxylamine, for hydroxylamine and nitrous acid decompose 
together into nitrous acid and water, the other product in the 
present case being potassium acid sulphate only. It decomposes 



POTASSIUM NITRITO-HYDROXIMIDOSULPHATES. 213 

explosively when heated — more so than does the hydroxiraido- 
siilphate by itself — giving off almost colourless gases and white 
fumes, just as might be expected and just as does a dry mixture 
of its constituent salts in corresponding proportions or a mixture 
of nitrite with a little sulphite. 

The compound salt can be purified from other salts or from 
alkali when these are present by recrystal Using from strong 
enough potassium nitrite solution. But from its own mother- 
liquor it can bo separated only by draining on the tile and not 
by washing. Such draining however is very effective because of 
the felted fibrous form of the salt, its non-deliquescent natuie, 
and the hygroscopic character of a solution of potassium nitrite. 
The analysis of the salt was made in the usual way described 
in our previous papers on hydroximidosulphates and other 
sulphonated-nitrite derivatives. By boiling its solution with an 
acid most of its sulphur appears as ordinary sulphate, but not 
quite all ; so that in estimating the sulphur the solution must be 
hydrolysed for some hours at 150"' under }>ressure. The results 
of analysis were :— 





Pütas^iimi 


Sulphur 


Foimd, 


33.14 


17.9.5 per cent. 


KsHNoS^O,, 


33.10 


18.06 „ 



There are other ways in which the potassium iiitrito- 
2/3 normal hydroximidosulphate may be formed all cousistino- 
essentially in producing the hydroximidosulphate by sulphonatins; 
a small portion of the potassium nitrite in a concentrated solution. 
Thus the following mode of working will give good results with 
certainty but it may be wideh^ deviated from with due conside- 
ration and precaution provided only that a concentrated solution 



214 E. DIVERS AND T. HAGA : 

of nitrite be employed. Potassium nitrite, 30 grams ; potassium 
liydroxide, 10 grams ; water, 50 to 100 grams are to receive a 
current of sulphur dioxide freely until crystals begin to form, 
the containing flask being all the time agitated in a cooling 
bath of ice and brine. The sulphur dioxide is now to be entered 
more slowly for some time longer and then stopped. After let- 
tins: the flask stand for half an hour the solution should be full 
of the desired salt which is then drained dry on the tile. Its 
mother-liquor is alkaline to litmus but not to rosolic acid (pre- 
sence of sulphite, absence of alkali) ; the well-drained salt itself 
is only faintly alkaline to litmus, if at all so. The double salt 
is also produced when to an ice-cold nearly saturated solution 
of potassium nitrite a similar solution of potassium pyrosulphite 
is very slowly added until crystallisation begins after which the 
solution is allowed to stand for some time. Thus prepared, the 
compound salt is liable to be contaminated with a little nitrilo- 
sulphate and sulphite. The experiment just described was made 
first by Raschig but he attached to it a significance unlike that 
here presented. Discussion of his views will be found towards 
the end of this paper. 

There is yet another way in which this potassium nitrito- 
hydroximidosulphate can be produced which it is of interest to 
mention because it illustrates the decomposibility of potassium 
5/6 normal hydroximidosulphate into the normal and 2/3 normal 
salts. AVhile the 2/3 normal salt dissolved in 16 per cent, or 
richer solution of the nitrite crystallises out only in combination 
with nitrite, the öjß normal salt can be dissolved in a nitrite 
solution of even 50 per cent, and yet for the most part crystal- 
lise out again uncombined. But generally with this strength of 
nitrite solution a little fluffy or cotton-like lustreless matter also 



rOTASSIUM NITKITO-HYDROXIMIDOSULPHATES. 215 

separates. If now to this fluffy matter suspended in its cold 
mother-liquor carefully decanted from every particle of the crys- 
tals of the '3/6 normal salt a liot solution of this 5/() normal salt 
in 50 or even 40 per cent, nitrite be poured in, a relatively large 
quantity of the fluffy matter is obtained and not the hard 
prisms of the 5/6 normal salt. Under the microscope the fluffy 
matter proves to be crystalline and when drained on the tile it 
exhibits a silvery lustre while on analysis it proves to be the 
nitrito-2/3 normal hydroximidosulphate only slightly impure from 
the presence of a little 5/6 normal hydroximidosulphate and nitrite. 
Thus in place of the potassium 33.10 and sulphur 18.06 per cent, 
we found in it 33.79 and 18.35 respectively, together with an 
alkalinity equal to 1.09 per cent, potassium. Dissolved up in 
hot 12 per cent, nitrite solution it recrystallises as the pure 
double salt. It is thus apparent that in a very concentrated 
solution of nitrite containing the ojß normal salt dissolved there is 
unstable equilibrium between tlie tendency to yield HON (SO;; K).^, 
KON(SO,K)2,OH2 again and that to form HON (SOo K),, 
KONO. 

Sodium nitrite forms a compound with sodium 2/3 normal 
hydroximidosulphate which has not been further examined prin- 
cipally because of its high solubility in sodium-nitrite solution. 

Potassium nitrite and normal hydroximidosul'phate KNOo, 
2K0N(S0oK)., 40Ho.— This compound salt is only obtainable 
from a strongly alkaline solution. For when the normal hydrox- 
imidosulphate is dissolved in a hot concentrated solution of the 
nitrite only the h\'6 normal hydroximidosulphate crystallises out 
on cooling just as it would do in the absence of nitrite. In 
order to crystallise out either the normal hydroximidosulphate 
(this Journal, 7, 30) or its combination with nitrite free alkali 



216 E. DIVERS AND T. HAGA : 

must be present in some quantity in the solution. The presence 
of too much alkali causes a little of it to separate with the 
normal salt, taking the lAace apparently of the water of crystal- 
lisation of this salt (this Journal 7, 02), and similarly to 
separate with the normal salt in its combination with nitrite, 
then also seeming to lessen the capacity of the normal salt to 
take up nitrite. The double salt is readily obtained by dis- 
solving normal hydroximidosulphate nearly to saturation in a 
hot (70°) solution consisting of 33-66 parts nitrite and 3-5 parts 
hydroxide to 100 parts water and cooling. Usually it forms 
lustrous silky fibres like those of the 2/3 normal double salt but 
radiating from points to form voluminous soft spherical masses. 
When the solution is more strongly alkaline the double salt 
separates as nearly opaque spherical granules with sometimes 
long fibres growing out from them. Under the microscope these 
granules are seen to have also a radiating fibrous texture and to 
represent the soft voluminous spheres highly condensed. Probably 
these always begin their growth from a minute granular nucleus. 
The double salt can only be purified for analysis by pressing it 
on the porous tile, when the soft spheres become a felted lustrous 
cake and the hard white granules crumble down like masses of 
wax. Analysis of the tw^o forms has given us the following 
results : — 





Potssm. 


Alk. potssm. 


.Sulpluir 


hjilky ; fuuiicl, 


35.21 


9.92 


16.23 per cent 


K,N3ÖAr.,4.4 0Ii,, 


35.14 


10.01 


16.43 „ 


Granular ; loimd, 


33.99 


9.20 


15.90 „ 


KvNgS.O,,;, COH,, 


33.89 


9.G8 


15.85 „ 



The varying amount of water is only the recurrence of what we 



POTASSIUM NITRITO-HYDROXIMIDOSULPHATES. 217 

have recorded concerning the normal potassium hydroximido- 
sulphate by itself. The double salt is exceedingly alkaline, its 
alkalinity we estimated by means of decinormal acid and litmus. 

Like the previously described double salt it is but little 
soluble in concentrated nitrite solution and freely soluble in water 
which decomposes it into its constituent salts and also decom- 
poses one of these, the normal hydroximidosulphate, into alkali 
and crystals of the oj6 normal salt. When heated it decomposes 
suddenly but gently and without fusing or scattering, and evolves 
slight red fumes only. It was by this behaviour quite distin- 
guishable from the 2/8 normal double salt and also from any 
other hydroximidosulphate which, simple or combined with 
nitrite, contained less than its K; to S^. By dissolving the 
nitrito-2/3 normal hydroximidosulphate in a hot concentrated 
solution of nitrite containing sufficient alkali the nitrito-normal 
hydroximidosulphate can be readily obtained by cooling the 
solution. 

Potassium nitrite and j^otassiimi Ö/6 normal hydroximidosul- 
phate. — We have obtained three compounds of the 5/6 normal 
salt with nitrite, one being TKNO., 2EIK5(NS20.)2, 30Ho. By 
using an almost saturated solution of potassium nitrite con- 
taining a little potassium hydroxide and dissolving in it by heat 
the Ö/6 normal hydroximidosulphate there is obtained a compound 
in minute fibrous crystals very lustrous when dry and decomposed 
by water but recrystallisable from a saturated nitrite solution. 
The same compound salt can be obtained also by dissolving the 
nitrito-normal hydroximidosulphate in hot almost saturated solu- 
tion of nitrite. 

Heated it proves to be mildly explosive. Its composition 
approaches that indicated by the formula given above. For 



218 E. DIVERS AND T. HAGA : 

analysis it was only air-dried on the tile ; in the desiccator it 
would prohably have lost its 3 per cent, of water (^SOHo) and 
then approached in composition Fremy's sulphazite. 





Pots8m. 


Sulphur 


Alk. Potssm. 


Original salt, found, 


36.81 


14.35 


4.80 per cent. 


Eecrystallised, ,, 


36.68 


14.47 


4.51 „ 


Ki;H2NuSA2, 30H,, 


36.87 


14.20 


4.34 „ 



A second double salt, SKNOo, K5H (NSo07)2, OH2, was got by 
dissolving one mol. 5/6 normal salt and 1.4 mol. potassium hy- 
droxide in a hot 65 per cent, nitrite solution and cooling. In 
appearance it resembled the other compound salt. Its analysis 
gave : — 

Potssm. Sulphur Alk. potssm. 

Found, 36.17 15.07 4.51 per cent. 

Calc, 36.81 15.06 4.61 

A third double salt, anhydrous, 7KNO2, 3K5H(NSo07)2, was 
not prepared synthetically but by treating an almost saturated 
solution of the nitrite with alkali and sulphur dioxide, and ad- 
ding alkali again after the sul phonation, imitating a process of 
Fremy's. Then, filtering the heated solution from much crys- 
talline 5/6 normal hydroximidosulphate mixed with a little of its 
combination wnth nitrite, we got the mother-liquor, when quite 
cold, almost filled with tiny j^risms of a compound answering to 
the above formula : — 

Potssm. Sulphur Alk. potssm. 

Found, 36.94 16.37 4.96 per cent. 

Calc, 36.99 16.51 5.05 

This salt was quickly resolved by water into nitrite and 



POTASSIUM NITRITO-HYDROXIMIDOSULPHATES. 219 

crystals of the very sparingly soluble 5/6 normal hydroximido- 
SLilphate. 

The varying proportions in which potassium nitrite and the 
5/6 normal hydroximidosulphate unite would possess but little 
interest were it not for the fact that they have evidently been 
severally met witli and taken to be salts of specific constitution 
by Fremy and b}^ Raschig. 



No7i- existence of Dihydroxylaminesul'phonateB. 

Fremy believed in the existence of less sulphonated deriva- 
tives of potassium nitrite than his sulphazite (see next paper) it- 
self less sulphonated than his sul^^hazotates (hydroximidosulphates) 
and attributed his failure to find them to the fact of their possess- 
ing exceedingly high solubility. Claus held much the same 
views and believed that by adding to an aqueous solution of 
potassium nitrite an alcoholic solution of sulphur dioxide in not 
too large a quantity he had obtained an impure crystallisation 
of a salt, ON SO3K (Ber. 1871, 4, 508) : he did not prove this 
to be the case, but what he did publish about his product is 
sufficient to show us that he had got the compound of potassium 
nitrite with 2/3 normal hydroximidosulphate we have described 
in this paper. A repetition of his experiment gave us this double 
salt together with much ethyl nitrite. Raschig regarded Ciaus's 
preparation as essentially the same as one of his own salts to 
which he gave the constitution of basic dihydroxylamine sul- 
phonate derivatives with the following formulae : — 

/OK HOv /SO,K 

HON< and ^^ >NON< _^ 

\SO3K (SOoK)/ \0K 



220 E. DIVERS AND T. HAG A : 

These he prepared by partial siilphonation of the nitrite in 
known ways. They both yielded crystals of a hydroximidosul- 
phate when dissolved in a little water and differed in no essen- 
tial particular from nitrito-hydroximidosul pates. From hot 
solutions of nitrite and a hydroxiraidosulphate we obtained by 
cooling an apparently homogeneous crop of crystals of almost 
the same composition and properties as one or other of Raschig's 
salts. Raschig gave two ways for preparing the salt having the 
second of the formulae just given and in these ways we have 
obtained the nitrito-2/3 normal hydroximidosulphate already de- 
scribed in this paper, but mixed with a little potassium sulphite. 
This impurity accounts for the alkaline reaction of Raschig's 
preparation and the presence in it of a little more than Ko to S2. 

He got the other salt (K., to S) only once and in the form 
of white crusts when working unsuccessfully for hydroximidosul- 
phate in Claus's way, the other main product being imidosulphate, 
that is hydrolysed nitrilosulphate as he himself pointed out. 
We have obtained — also by sulphonating nitrite, following Fremy 
— a product qualitatively like Raschig's salt though quantitatively 
a little different from it, and at the same time like the second 
salt compounded of nitrite and 0/6 normal hydroximidosulphate, 
described by us on page 218. The percentages found by Raschig 
were potassium, 36.84, and sulphur, 15.50. 

When Raschisr's salt was dissolved in water and acidified it 
gave nitrous oxide as the only gaseous product while ours gave 
also some nitric oxide. This fact might have served lo render 
incorrect the application of our formula to his salt but for the 
evidence there is that this was mixed wâth a little sulphite which 
would have reduced any nitric oxide. Its mother-liquor on 
further evaporation gave, we are told, so much sulphite along 



POTASSIUM NITKITO-HYDROXIMIDOSULPHATES. 221 

with tlie next crop of the salt itself as to cause its rejection. 
The presence of sulphite in less quantity in the first crop of 
crystals will have been masked by the oxidising action of the 
nitric oxide in becoming nitrous oxide. That sulphite was pre- 
sent in Kaschig's preparation well accords also Avith the fact that 
potassium hydroxide added in excess precipitated potassium sul- 
phite, for, although hydroximidosulphate is itself decomposed by 
the most concentrated solutions of potassium hydroxide into 
sulphite and nitrite, this decomposition is slo^v and the sulphite 
only deposits after some time. Raschig's preparation when dried 
on a tile was only a powder, that is, presumably, was not ob- 
viously crystalline, a point which also indicates an impure salt. 
Since the potassium and sulphur are iu the same ratio in the 
two salts, quantitative analysis would hardly have made its 
presence known.'-' Inspection of Raschig's formuhe is of itself 
sufficient to prevent their getting accepted as in accordance with 
the facts. For from these formuhe both salts should be strongly 
alkaline, while in reality one is neutral. Above all it is hardly 
credible that dissolution in cold water should suffice to cause 
monosulphonated nitrogen to become disulphonated. 

Raschig held his two salts to be identical with Fremy's 
potassium sulphazite and sulphazate respectively ; but the nature 
of Fremy's salts will be found, we believe, more precisely given 
in the paper following this. The point we would here insist 
upon is that Raschig's preparations, judged by their chemical 
behaviour, have no claim to be considered as dihydroxy lamine 
derivatives, being in every way indistinguishable from synthe- 

*0f the 3KN0„ of our formula (p. 218) only one mol. can give nitric oxide and only 
to the extent of two-thirds of its nitrogen ; the other tliird becoming nitric acid. Kaschig's 
analysis indicates the presence of only .S/4 mol. active nitrite. Tlie quantity of iiydrated 
sulphite required to be present is therefore only 5.2 per cent, of the mixed salts. 



222 E. DIVERS AND T. HAGA : 

tically prepared compounds of nitrite and hydroximidosulpliates. 
Diliydroxylamine salts have as yet only a hypothetical existence 
and are likely to remain so. For the double linking of the 
oxygen atom with the tervalent or quinquevalent nitrogen atom 
seems always experimentally to make or break itself in a single 
act, notwithstanding its bipartite character. 



Kaschig in his researches on Fremy's sulphazotised salts got, 
besides those we have just discussed, two other salts of undeter- 
mined constitution, both of which were most probably also 
nitrito-hydroxiraidosulphates. They may therefore be noticed 
here although Kaschig did not represent them to be dihydroxyl- 
amine derivatives. Yet they were evidently closel}^ like the 
other two in properties. One was isomeric with potassium liypo- 
nitrososulphate (Pelouze's salt) and also with his (K^ to S) ' dihy- 
droxylamine ' salt, allowing for different hydration, and the other 
was isomeric with potassium Ö/6 normal hydroximidosulphate. 
Each could be obtained but once and they only call for any 
detailed notice because of the theoretical imj)ortance given to 
them as isomerides of other salts. The first referred to above 
was mistaken by Raschig for Pelouze's salt (hyponitrososul- 
phate) but that salt it certainly was not (this Journal, 9, 85). 
It was got by dissolving nitric oxide in solution of potassium 
sulphite and hydroxide and evaporating to a small volume till 
crusts formed. If we assume that air or nitric peroxide was not 
excluded there were the conditions present for getting a nitrito- 
hydroximidosulphate, for, as we show in a paper which will 
shortly follow this, nitrous fumes passed into potassium sulphite 



POTASSIUM NITEITO-HYDIIOXIMIDOSULPHATES. 223 

solution generate hydroximidosulphate freely together with 
nitrite. 

The other salt isomeric with 5/(3 normal hydroximidosulphate 
was obtained in Raschig's attempt to form 2/3 normal salt by 
passing surphur dioxide into a solution of potassium nitrite and 
hydroxide and letting stand for a day. These, too, are conditions 
for geting nitrito-hydroximidosulphate. Now, both products 
agreed in being decomposed by water in such a way as to yield 
hydroximidosulphate and in other ways behaved as compounds 
of nitrite with one of these salts. The behaviour of the one 
isomeric with hyponitrososulphate was indeed exceptional in 
that when dissolved in water containing a little alkali it gave 
the 2/3 normal hydroximidosulphate when according to our cal- 
culation it should have given the 5/6 normal salt, while it also 
gave in hot alkaline solution u little nitrous oxide which only 
hydroxyamidosulphate is known to give. These peculiarities 
we may attribute to partial hydrolysis having occurred in the 
very unstable salt before these experiments were made. 

The calculated formula for the isomeride of hyponitrosul- 
pliate as a nitrite compound is oKNOo, K.^H (NS^Oy),, 2OH2, and 
such a. compound we liave described on page 218 ; that for the 
isomeride of the 5/6 normal hydroximidosulphate treated as being 
a nitrite compound is 3KN0,, 6K,HNS,0;, 5K5H(NS207)2, which 
in water should give crystals of KoHNSoO;, 2OH2. This com- 
pound salt we have failed to get but its occurrence can be 
readily accepted as possible. Its assumed existence affords a much 
more satisfactory explanation of the nature of this salt of Ras- 
chig's than that we were able to offer in our paper on hydrox- 
imidosulphates already referred to. 



224 





Potssra. 


Sulphur 




Nitroso isomer, found, 


35.72 


14.40 


per cent. 


Calculated, 


36.05 


14.75 


i> 


Ü. & H's salt, found, 


36.17 


15.05 


}> 


Oximido isomer, „ 


33.04 


21.23 


f> 


Calculated, 


32.91 


21.54 


V 



For the present, the existence of isomesides of Pelouze's salt 
and Fremy's basic sidphazotale must be regarded as no longer even 
probable. 



Identification and Constitution of Fremy's 

Sulphazotised Salts of Potassium, his 

Sulphazate, Sulphazite, etc. 

By 

Edward Divers, M. D., D. ^c, F. R. S., Emeritus Prof., 

and 
Tamemasa Haga, D. i^c, F. 0. S., 

Professor, Tokyo Imperial University. 



A safficieDtly concentrated solution of potassium nitrite and 
hydroxide submitted to the action of sulphur dioxide gave 
Freray minute silky needles of a salt which he provisionally named 
jjotassium sulphazate. With slightly diminished concentration of 
the solution he generally obtained instead the brilliant, often 
hard, rhombic prisms of j^otassmm basic suljohazotate (o/6-normal 
hydroximidosulphate, this journal, 7, 15). But sometimes 
there was obtained neither of these salts before the solution 
became transformed into a starch-like jelly through the form- 
ation of a salt which he named potassium metasulphazate, or 
else became filled with spangles of yet another salt called by him 
potassium metasulphazotate. When the solution was a little too 
dilute to give any of these and when too much alkali had not 
been added, there usually appeared peculiarly pointed crystals of 
the salt he named potassium neutral sul])liazotate (2/3-normaI hydr- 
oximidosulphate Raschig) and, lastly, with still greater dilution 



226 DIVERS & HAGA : IDENTIFICATION AND CONSTITUTION OF 

the minute brilliant needles of his potassium suljohammonaie 
(nitrilosulphate Berglund). Still other salts he believed to be 
produced in the first stages of the reaction between the nitrite 
and sulphur dioxide, one of which he named potassium sulph- 
azite; but this he did not obtain directly, finding a reason for 
this in the exceeding solubility of this early formed salt. He 
prepared it — but only in quite small quantity and as crystalline 
warty granules — by the action of water upon the ' sulphazate ' 
whereby this was converted into ' basic sulphazotate ' which de- 
posited and a solution that on evaporation yielded the ' suIjdIi- 
azite.' These two salts could together in solution be changed 
back into the ' metasulphazotate ' while the ' sulphazite ' and 
the ' sulphazate ' could similarly often be changed into the ' meta- 
sulphazate ' again. These two ' meta ' salts he regarded therefore 
as perhaps merely double salts of the others. The ' sulphazite,' 
the ' sulphazate,' and the ' sulphazotates ' he treated as being 
members of a series of salts in which there Avere to two atoms of 
nitrogen from one up to eight atoms of sulphur, — three in the 
" sulphazite, four in the ' sulphazate', and five in the ' sulph- 
azotates.' With this conception of the nature of these salts, based 
on his analyses, it was easy to understand the decomposition of 
the ' sulphazate ' into the ' sulphazite ' and the ' sulphazotate.' 
But this and other of Fremy's interpretations of the facts ob- 
served by him have lost all importance and particular interest 
through the progress of chemistry since his memoir was pub- 
lished and only his account of the facts requires consideration 
now. 

Subsequent work by others and ourselves in the same field 
has shown that Freniy in the account he gave of the preparation 
of his many salts went two little into details as to the conditions 



FEEMy's SULPIIAZOÏISED SALTS OF POTASSIUM. 227 

under which they were obtained, — apparently because he was 
not able to be more precise. When Claus attempted to get 
Fremy's salts he obtained only masses of minute crystals of salts 
of whose individuality and nature he could make out little be- 
cause of the impossibility of dissolving them all up undecomposed. 
In his experiments the ' sulphammonate ' (nitrilosulphate) was 
always formed in considerable quantity eitlier as a first or second- 
ary product and by its presence prevented any satisfactory inves- 
tigation of the other salts. In Fremy's working, this most easily 
formed salt came only as the final product of the sulphonation 
and therefore gave him no trouble. Claus emphatically displayed 
his scepticism as to Fremy's results ; yet in nearly every point 
in which he difi'ered from Fremy as to the facts we find Fremy 
to have been right. When Raschig repeated Fremy's work — 
hut with the modifications in procedure introduced by Clans — he 
got results similar to, though less unsatisfactory than, those Clans 
had obtained. He made an approach to Fremy's work in so far as 
that he often got very little nitrilosulphate ; nevertheless he too 
failed in his attempts to prepare the ' sulphazate ' in Fremy's way. 

In perhaps all essential points we can lay down the method 
to repeat Fremy's experimental work successfully. But in some 
cases a little uncertainty obtains owing to the fact that the very con- 
centrated and complex solutions which yield Fremy's salts are apt 
to deposit what is virtually the same salt in different forms as well 
as at times salts quite distinct from each other under only slight and 
obscure variations in the circumstances attending their formation. 

Sulphazate. — This is Fremy's first salt directly obtained in 
his sulphonation of the nitrite. In getting it he took approxi- 
mately 5 mois, potassium nitrite to 2 mois, potassium hydroxide 
and a little water and into the solution passed sulphur dioxide 



228 DIVERS & HAGA : IDENTIFICATION AND CONSTITUTION OF 

until it became almost filled with silky needles very soluble in 
water. So far it is easy to follow Freray with a full measure of 
success if only the water used is limited to perhaps twice the 
weight of the nitrite and that the heating effect of the nitrite is 
counteracted by cooling. Claus and after him Raschig failed but 
then inexplicably to us they did not start with Fremy's propor- 
tions of nitrite to hydroxide, though even with the proportions 
they took, success was possible with care. The salt thus formed 
by Fremy was not tested and analysed by him until after it had 
been changed (but without his having recognised the fact) by 
the further treatment to which he submitted it. Before its 
change it is potassium nitrito-2/ö normal hydroximidosulphate 
described in the preceding paper, a neutral salt decomposed by 
water into its constituent salts. Fremy's finished ' sulphazate ' 
was strongly alkaline and very caustic and when decomposed by 
water gave nitrite and the öjQ normal hydroximidosulphate — 
not the 2/û-nornial salt. Also the analysis he gave of it fur- 
nished numbers such as the original product could not have given 
him. Instead of potassium, 33.10, sulphur, 18.06, and nitrogen, 
7.9 per cent., he got potassium, 34.90, sulphur, 19. öö, and 
nitrogen, 4.9. We can learn wliat his after-treatment was by 
reference to other parts of his paper where he s])eaks of the care 
necessary (when sulphonating the nitrite) to maintain the alka- 
linity of the solution by adding potassium hydroxide from time to 
time and of dissolving sulphazotised salts for examination in water 
containing this alkali. Certain it is he must have added some 
potassium hydroxide to the solution after getting it to crystallise, 
as a precaution to preserve the salt. Now the effect of this ad- 
dition is to change the composition of the product without much 
aflecting its silky asbestus-like appearance. The change in com- 



) 



fkemy's sulphazotised salts of potassium. 229 

position is to deprive it of much of its nitrite and to convert the 
2/3-normal into more nearly normal hydroximidosulphate — to 
replace, therefore, potassium nitrite by potassium hydroxide. 
Accepting Fremy's mean numbers as accurate, what he analysed 
had the composition, 

llKoNSoO^OHo; KoHNSA, SOH,; 2(KoHNSA, KNO,). 

Putssm. Sulphur Nitrogen Alk. potssm. 
Found, 34.9 19. ,5.5 4.9 per cent. 

Calc, 34.9 19..51 4.9 9.36 „ 

But his analyses have no claim to receive such close treat- 
ment, his nitrogen seemingly being always much too low ; and it 
is sufficient to say of his ' sulphazate ' that it was the silky 
asbestus-like nitrito-2/o normal hydroximidosulphate more or less 
converted into the also silky asbestus-like normal hydroximido- 
sulphate, an account of it with which Fremy's description of its 
other properties entirely agrees. With dilute acids it gave slowly 
nitrous oxide unmixed with nitric oxide. Fremy specially points 
out that no sulphazic acid or any other sulphazates could be ob- 
tained from the potassium salt. There is, therefore, nothing to 
justify belief in this compound being the salt of a particular 
single acid, the sulphazic. 

Sulphazite. — What Fremy named potassium sulpkazite he only 
once obtained, and then not by direct sulphonation of t]ie nitrite, 
in the form of white mammillated crystalline crusts from a solu- 
tion thickened by the other salts contained in it. That is, to 
say, his sulphazate when dissolved in a little water containing 
some potassium hydroxide deposited crystals of basic sulphazotate 
{5/6 normal hydroximidosulphate), and left a mother-liquor which 
on cold evaporation till syrupy yielded the sulphazite. It showed 
great analogy with his sulphazate but was distinguished from it 



230 DIVERS & HAGA : IDENTIFICATION AND CONSTITUTION OF 

by having little tendency to hydrolyse and by at once evolving 
some nitric oxide when its solution was mixed with a dilute acid. 
Water decomposed the sulphazite, but into what products was not 
ascertained. 

We have sufficiently realised Fremy's expectations that his 
sulphazite might directly result from sulphonating the nitrite with 
subsequent addition of alkali. The substance obtained in this 
way did not differ greatly in composition from his : 

Potassm. Sulpliur 

Fremy's salt, .38.1G 16.27 per cent. 

D. & H's salt, 3G.94 1C..37 „ 

and agreed with it in chemical properties, so far as is known. 
At the same time it was indistinguishable from a compound of 
nitrite with 5/6 normal hydroximidosulphate, and has been de- 
scribed by us as such in the preceding paper (p. 218) in which 
it stands as the third of these double salts and in which its 
preparation is given. Other experiments of various kinds have 
yielded us such ' mammillated crusts' as Fremy got, which, 
though only in rough agreement in percentage composition with 
his sulphazite, behaved like it and proved to be impure double 
salts of nitrite w^ith o/ß normal or more nearly normal hydrox- 
imidosulphate. We are therefore convinced that his sulphaz- 
ite wns only such a double salt. 

Metasulphazate'"'' — In Fremy's experience it sometimes hap- 
pened, when passing sulphur dioxide into solution of nitrite and 
alkali of a concentration intermediate to that giving sul2:)hazate 
and that giving basic sulphazotale, the solution set to a starch-like 
jelly instead of crystallising. He obtained a similar jelly by cooling 



* Often, misprinted metasulphazolate in the French original, but not in the German 
translation. 



fremy's sulphazotised salts of potassium. 231 

a concentrated solution of sulphazate and sulphazite ; also by- 
boiling a solution of sulphazate and then cooling it. When 
strongly compressed the jelly became a transparent wax-like mass. 
Heated in this waxy state to 50°-60° it suddenly changed into a 
solution of sulp)hazite and minute crystals of basic sulphazotate. 
In all other respects it proved to be intermediate in properties 
to sulphazate and sulphazite. No other metasulphazates could be 
prepared from it, so that Fremy was disposed to regard it as 
being a doublesalt oî sulphazate ?ind sulphazite. Its constitution 
must therefore have been that of nitrite combined with normal 
or ö/ß normal hydroximidosulphate in such proportions and with 
such additions perhaps of alkali as prevented crystallisation. 

We have not had Fremy's success in getting this salt in 
form of jelly and wax but have met with just such phenomena 
when forming barium sodium hydroximidosulphate, BaNaNSoOy, 
OHo, as will be found described in our paper already frequently 
referred to. We have however obtained a salt, or homogeneous 
mixture of salts, of the same composition as the metasulphazate, 
but with the form of the silky radiating fibrous crystals of the 
nitrito-normal hydroximidosulphate, from wdiich it differed only 
in showing deficiency of nitrite, that is, it w^as equivalent in com- 
position to a mixture of the normal salt and its nitrite com- 
pound, both of which crystallise with the same habit. We give 
below Fremy's numbers, our own, and those calculated for the 
expression, 3(KN0o, 2K3NSA, 40Ho); K.NSA; 30Ho. 

Potssm. Sulplinr Nitrogen Alk. potssra. 

Found (Fremy), 35.10 16.74 4.81 i^er cent. 

,, (D. & H.), 35.10 16.68 10.47 ,, 

Calculated, 35.06 16.74 5.23 10.23 „ 

We got the salt by dissolving the hydroximidosulphate in 



232 DIVEKS & HAGA : IDENTIFICATION AND CONSTITUTION OF 

hot concentrated nitrite solution containing alkali. To 100 cc. 
water there were present 45^ /o grm. nitrite and 1% grm. potassium 
hydroxide ; for 66 mol. nitrite there were dissolved 10 mol. 
anhydrous normal hydroximidosulphate. But for the salt being 
in beautiful asbestus-like fibres, there was nothing to distinguish 
it from the jelly and the wax-like 7}ietasulphazate, which, therefore, 
we do not hesitate to class as a nitrito-hydroximidosulphate. 

Basic sulphazolaie, which Fremy considers next, has been 
shown by us already [loc. cit.) to be the 5/6 normal hydroximi- 
dosulphate, and not the salt of a distinct acid, the sulphazotic. 
It is liable to contain a small excess of potassium when crys- 
tallised from a strongly alkaline solution. x\ solution of the 
normal salt readily deposits it, as does also that of the nitrite 
compound of the normal salt. 

Neutral sulphazotate was shown by Raschig to be the 2/3 
normal hydroximidosulphate. The potassium sulphazotates were 
distinguished by Fremy from the salts previously described by 
him by their ability to form other sulphazotates by double de- 
composition. Fremy's analytical results in the case of the 7ieu- 
tral sulphazotates are hopelessly out of accord with its constitution 
and properties, though those for the basic sulp)hazotate are satis- 
factory enough. 

Sulp)hazidate, produced by the hydrolysis of the sulphazotate, is 
hydroxyamidosulphate (Claus). Sulphazilate and metasulphazilate, 
oxidation products oi sulphazotate are 0N(S0oK)2 and 0N(S03K)o, 
and have been studied by Claus, Raschig, and Hantzsch. 

31etasulphazotate. — Sometimes Fremy got a salt in the form 
of spangles {paillettes), in appearance like minute crystals oî basic 
sulphazotate, but differing from these in not being hard under pres- 
sure. This salt he named, therefore, metasulphazotatc. According 



fremy's sulphazotised salts of potassium. 233 

to liini it is also obtainable by mixing (hot) solutions of the (basic) 
sidphazotale and sidphazite. It is very soluble in water, very 
alkaline and unstable unless the water contains alkali. In pure 
water it becomes basic sidphazotale and sulphaziie ygain. It 
shoAvs the greatest analogy with metasulphazate and is distin- 
guished in the same way as this salt from basic sulphazotate. It 
may be a compound of basic sulphazotate and sidphazite. So far 
Fremy. It will be evident that there is nothing in its history 
or properties to distinguish it, except its occurring in the form 
of sparkling particles and even that can be met with in the 
basic sulphazotate suddenly precipitated ; we have also got other 
of the sulphazotised salts in what may be called spangles, 
though not this particular salt. In the preceding paper, page 214, 
we have described an impure form of nitrito-2/3 normal hydrox- 
imidosulphate obtained by dissolving the 5/6 normal salt in a hot 
concentrated solution of nitrite, but still not so very concentrated 
as to give the nitrito-5/6 normal double salt. This preparation 
is lustreless while in its mother-liquor, but w^hen dried on the 
tile has a line silvery lustre. It has when dried in the desic- 
cator exactly the composition of Fremy's metasulphazotate and is 
much less alkaline than the metasulphazotate and is much less 
alkaline than . the metasulphazate. It may be formulated as 
K3NSA; 9(KN0.,,K2HNSA,1V20H2). 

Potsstu. Sulphur Nitrogen Alk. potssm. 

Found (Fremy), 33.8 18.6 3.5 percent. 

„ (D. &H.), 33.79 18.35 1.09 „ 

Calculated, 33.68 18.37 7.63 1.12 „ 

Sulphammonate and sulphamidate are respectively nitrilo- 
sulphate and imidosulphate (Berglund). 



On a Specimen of a Gigantic Hydroid, 

Branchiocerianthus imperator (Allman), 

found in the Sagami Sea. 



By 



M. Miyajima, Rigakushi. 
Science College, Imperial University, Tokyo. 



With Plates XIV cO .YT^. 



On the morning of January 1, 1899, quite a commotion was 
produced in the Marine Biological Station at Misaki b}' the 
bringing in of a very beautiful and gigantic Coelenterate (PI. XIY). 
It had been caught, on the previous day, by a fishing " long- 
line," from a depth of about 2ö0 fothoms near Okinose, a 
submarine bank 18 kilometers south of ^lisaki. It was an object 
which was calculated to raise enthusiasm in a naturalist. A large 
disc surmounted a Ions; stalk which evidentlv fixed the animal 
on the sea-bottom. A circle of numerous graceful tentacles haiûg 
down from the margin of the disc, while on its upper surface 
arose an oral tube, surrounded at its base by bushy dendritic 
appendages and having a second circle of slender tentacles around 
its upper edge. The total height of the animal was 700 milli- 



236 M. MIYAJIMA : 

meters and the prevailing colour transparent scarlet. It was 
agreed on all sides that it was a New Year's gift from Otohime'-' 
and that it should be known in Japanese as Otohime no 
Hanagasa. 

The specimen, when brought in, was entirely fresh but was 
not living. It was placed in 29o formalin to preserve, if pos- 
sible, something of its beautiful colour. At first the attempt 
seemed successful, but after a while the colour began to fade 
gradually, until now the specimen is completely bleached to pale 
white, For histological examination, pieces of the fentacles and 
the dendritic appendages were fixed in the suhlimate and in 
Perenyi's fluid. 

The specimen was handed over by Prof. Mitsukuri to me 
to work out its finer structure. 

It was evident from the first that the specimen was very 
similar to the form only a short time before described by Mark 
('98) as BranchioceriantJms urceolus. I started, therefore, with 
an idea that I was dealing with an Actinian. 

As I proceeded in my investigation, however, it became plain 
that this idea was not tenable, and the conclusion was finally 
reached that the aniuinl was very closely allied to Corymorjjha, 
and that it belongs probably to the species obtained by the " Chal- 
lenger " at about the same locality and named by Allman ('85) 
Monocaulus Imperator, notwithstanding many discrepancies between 
his description and the specimen. This conclusion was communi- 
cated through Prof. Mitsukuri to Dr. Mark and a request was also 
sent to him, that during his opportune stay in Europe, he should. 



"•* Otohime " is a beautiful goddess who is supposed to have her palaces at the bottom of 
tlie sea. " Hanagasa " is the flower-sun-sliade or ornamental jiarasol. Thus Otohime no 
Hanayasa means " the ornamental parasol of Otohime." 



BRANCHIOCERIANTHUS IMPERATOR (aLLMAx). 237 

if possible, examine the original specimen of Monocaulus Imperator 
in the British Musenm. To the results of his examination of the 
specimens I shall return in the later part of this paper. 

Meanwhile an article was published in the Zoologischer An- 
zeiger by O. Carlgrex (*99) throwing doubt on Mark's Brcmchio- 
cerlanthus being an Actinian, and contending that it more 
probably is a Corymorpha or at least a form standing very close 
to Corymorpha. 

In June, 1801), a correction was published by Mark ('99) 
himself in the Zoologischer Anzeiger. His previous preliminary 
description had been based on external anatomy, and he now 
frankly admitted that further researches had convinced him of the 
fact that the animal in question must be more nearly related to 
the Hydroidea than to the Actinia, though its exact affinities he 
had not yet determined. In a postcript he mentions our conclu- 
sions which had been communicated to him, as mentioned, by 
letter, and thinks that both his and our specimens belong to the 
same genus and that our specimen is probably identical with the 
Monocaulus Imperator of Allman. 

Before going further I wish to express my deepest feeling 
of obligation to Prof. Mitsukuri for the supervision and advice 
which he has given during the progress of my work. 

Description. 

This hydroid is a solitary form consisting of a well marked 
hydranth and a hydrocaulus. Its most striking feature is a 
strongly expressed bilateral symmetry. The hydranth is disc- 
shaped and bears two sets of tentacles and a circle of dendritic 
gonosomes, all showing in their arrangement a well marked 



238 M. MIYAJIMA : 

bilaterality. The hydrocaulus, which is attached not to the 
center but to the edge of the hydranth, is nearly cylindrical and 
increases in diameter from the attachment of the hydranth 
towards the end which is fixed in the sandy sea-bottom. The 
total height of the animal attains 700 mm., as measured from 
the top of the oral tube to the attached base of the hydrocaulus. 

In the fresh condition the hydranth was rose pink and its 
tentacles, both oral and marginal, were deep scarlet in colour, 
while the gonosomes possessed light rufous colour. The hydro- 
caulus was light pink in colour, being quite pale in its middle. 

The general features and the colours are Avell shown in 
Fig. 1, PI. XIV, which was drawn from the preserved specimen 
by Mr. Nagasawa, artist of our Institute, making use also of 
the rough sketches I made at the time of the fresh object. 

Hydranth. 

The upper surface of the hydranth is flattened so that it 
may be described as an *' oral disc." The lower surface, how- 
ever, assumes a shallow funnel-shape, which passes downwards 
into the hydrocaulus. This disc has an oval outline, but differs 
from that of Branchiocerianthus urceolus, in having its sagiilal 
diameter less than its transverse, the two diameters] [being res- 
pectively 80 and UO mm. (Woodcut 2). 

At one end of the sagittal diameter is attached the hydro- 
caulus where the circle of the marginal tentacles is also inter- 
rupted. The plane of the disc is oblique to the long axis of the 
hydrocaulus (Woodcut 1, I), though not to the same degree as 
in Branchiocerianthus urceolus Mark. 

The edge where the hydrocaulus is attached I shall designate 
the lower, and the opposite the higher, edge. 



BRANCHIOCERIANTHUS IMPERATOR (aLLMAN). 



239 



Woodcut 1. 
1. 



re. 




vit 



Diagrams showing sagittal (1.) and transverse (II.) sections of the hydranth. 
B. hypostomal region of the disc ; C. orifice of the diaphragm {m) in the hy(h-anth ; (7 orifice 
of the diapliragm {m') in the hydrocaulus ; eg. central, tg. lateral, globule of the gonophore ; 
11. upper, IP lower cavity of the hydranth ; h. hypostorae ; k. intercalated cord ; HC' 
hydrocaulus; nt. marginal, ol. oral, tentacle; P. peduncle uf tlie gonosome; R. outer 
region of the disc, provided with the radial canals (/'.f.). 



240 M. MIYAJIMA : 

The liypostome (Woodcut 1, I, II, h), the superior prolonga- 
tion of the disc, is slightly conical, diminishing gradually in 
its diameter from the base towards the free end where the 
mouth opens. A little below the mouth the liypostome bears a 
brush-like group of about 180 filiform tentacles {ot.) which are 
arranged in three or more closely packed verticils, the outer 
tentacles being much larger than the inner. The outermost 
ones attain a length of 00-55 mm., while the innermost are so 
small and crowded that I could neither measure them well nor 
count their exact number. Below the oral tentacles the liypo- 
stome is slightly constricted, but there is no indication of syphono- 
glypli which is said to be present in the oral tube of Branchio- 
cerianthus iirceolus. The side of the liypostome turned towards 
the lower edge of the disc passes gradually to the disc, while 
on the opposite side it seems abruptly raised from the disc, 
so as to make an angle between. The liypostome is thus oblique 
to the disc proper which again is not perpendicular to the axis 
of the hydrocaulus. Hence we can show the relation of the three 
parts, the liypostome, the disc and the hydrocaulus, diagram- 
atically with three lines, of which two vertical ones, correspond- 
ing to the axes of the liypostome and of the hydrocaulus, meet 
with an oblique one representing the axis of the disc, forming 
obtuse angles betw^een them (Woodcut 1, I). 

The base of the hypostome (Woodcut 2, B.) occupies about the 
middle of the disc, but on the side turned towards the lower edge, 
its base gradually becoming lower and lower, may be said to 
stretch as far as the margin of the disc, while laterally and towards 
the higher edge it is distant from the margin 35 mm. and 22 mm. 
respectively. It thus assumes an ovoidal outline, the pointed end 
attaining the lower margin of the disc and passing directly to the 



BEANCHIOCEPvIANTHUS IMPERATOR (aLLMAN). 



241 



Woodcut 2. 




Diagram sliowing tlie upper surface of the disc. 
V, hiatus at the lower edge of the disc. Other letters as in Woodcut 1. 



surface of the hydrocaulus. The longer {i. e. up— clown) diameter 
of this ovoidal space measures 60 mm,, while the transverse at 
the widest middle portion is only 2-3 mm. This space is des- 
titute of the radial canals which are prominently seen in the 
remaining part of the disc. 

Around the margin of this hypostomal region there arises 
from the surface of the disc a row of dendritic gonosomes 
{p) which in shape strongly remind one of the heads of cauliflowers. 



242 M. MIYAJTMA : 

They number in all 9G and are arranged approximately in a 
single row, which, being interrupted at the lower edge of the disc, 
assumes the form of a horse-shoe. (Woodcut 2, P). At the two 
ends of the horse-shoe are situated the smallest gonosomes which 
stand at a distance of 15 mm, across from each other. The 
length of the stalk of the gonosomes varies from 20 mm. to GO mm. 
While the gonosomes nearer tlie lower edge of the disc are on 
the whole shorter than those nearer the upper edge, it is to be 
noticed that the larger and smaller gonosomes are placed alter- 
nately, indicating faintly the two circles in their arrangement, 
the larger gonosomes being placed in the outer, and the smaller 
in the inner, circle. 

The region of the disc outside the gonosomes is marked with 
numerous radial canals (Woodcut 2, R.) w^iich run from the 
base of the gonosomes to the margin of the disc. This region 
thus assumes the form of a wide horse-shoe, whose two arms 
gradually diminish in their breadth towards the lower edge of 
the disc until they terminate at that edge. Hence this region 
varies in breadth, measuring 20 mm. on the median line at the 
higher edge, and 35 mm. on the lateral region, while on the lower 
side both arms are practically zero. 

The radial canals (PI. XV, Fig. 1, r.c.) slightly swell out the 
surface of the disc thus giving the latter an undulating appear- 
ance. The canals are intercalated by solid cords (PI. XV, Fig. 1, 
i.e.) which appear on the surface of the disc as opaque lines. The 
canals and the intercalated cords are longest in the lateral region 
where they run obliquely across the disc, and are longer than 
the breadth of this region. The canals situated nearer the lower 
edge are smaller and shorter than those higher up, until at the 
both arm-ends they are practically nil. On the other hand the 



BRANCHIOCEEIANTHUS IMPERATOR (aLLMAn). 243 

canals on the higher side of the disc are not so long as those 
on the lateral, bnt run straight from the base of the gonosomes 
to the outer margin of the disc, the length of the canal being 
thus the same as the breadth of this region (Woodcut 2). 

The radial canals and the intercalated cords increase in their 
width towards the outer margin of the disc where the both struc- 
tures are broadest. Inwards, the radial canals open into that part 
of the hydranth-cavity where the cavities of the gonosomes stand 
in communication with the latter. Outwards, the canals terminate 
blindly on the margin of the disc. The intercalated cords 
enlarge suddenly near the margin of the disc and acquire a 
cavity which forms a part of that of the marginal tentacle (PL 
XV, Fig. 1). 

The name of marginal tentacles [mi) is given to the outer- 
most circle of filiform tentacles arranged like a fringe around the 
margin of the disc. The circle is not complete, there being a 
hiatus (v) at the lower edge of the disc where the surface of the 
hypostome passes directly into that of the hydrocaulus. The 
shortest tentacles arising from the 6th or 7th intercalated cord, 
counting from the lower edge, occu^^y the two ends of this in- 
complete ring. AVhether there were any smaller tentacles nearer 
the lower edge, I am not sure. There is no indication, so far 
as I can see, of any having existed. Towards the higher edge 
of the disc they increase successively in length until about the 
10th (counting from the lower edge) is reached, of which the 
length on both sides is about 200 mm. After this there seems 
to be no special arrangement of the tentacles, which vary from 
200 mm. to 300 mm. in length. They numbered 198 in all. 
The tentacles are flattened at their base, and are compressed so 
closely with one another that the basal portion appears to form 



244 M. M I YA JIM A : 

a part of the disc. Just above the flattened base the tentacle 
assumes the form of a tube, 4 mm. in diameter, and tapers 
gradually towards its free end. 

The hydranth (Woodcut 1, I. II.) contains a wide cavity 
which is separated by a thin membrane (???) into two parts, an 
upper (H) and a lower (H'). The superior prolongation of the 
upper cavity is that of the hypostome, which does not show 
any indication of the septal partition. The lower cavity is more 
spacious than the upper, and not only occupies the whole lower 
part of the hydranth but extends also through the entire length 
of the hydrocaulus. 

The membrane {m) separating the hydranth-cavity arises 
diaphragm-like just below the upper wall of the disc. In about 
the center of this diaphragm, directly below the mouth, is an 
ovoid orifice (Woodcut 1, II, C) which puts the upper and lower 
cavities in communication with each other. The orifice is 
11 mm. and 15 mm. respectively in its transverse and sagittal 
axes. 

That part of this diaphragm which corresponds to the part 
of the upper surface of the disc marked B in Woodcut 2, i.e. to 
the basal part of the hypostome, projects into the cavity of 
the hydranth like a shelf, with the aforesaid opening near its 
middle and with no attachment either above or below. Outside 
this portion, however, the diaphragm forms the floor of the 
radial canals mentioned above, so that it is suspended, so to sjDeak, 
by the numerous intercalated cords {vide supra) to the upper 
surface of the disc. At the margin the diaphragm is united to 
the outer wall of the hydranth (Woodcut 1). 

To show the somewhat complicated relations existing between 
the marginal tentacles, radial canals, intercalated cords, etc., a 



BEANCHIOCEEIANTHUS IMPERATOK (aLLMAn). 245 

series of sections (PI. XV Figs. 4-9) passing through the lines 1-1, 
2-2, 3-3, 4-4, 5-5, 6-6 in Fig. 1, PL XV is introduced. The 
first section (Fig. 4) through the outermost margin of the disc, 
which corresponds to the line 1-1 in Fig. 1, shows that the bases 
of the marginal tentacles (Lb.) and the blind ends of the radial 
canals {r.c.) are arranged alternately, the former projecting out 
above and below more than the latter. The upper projection 
corresponds to the enlarged end of the solid cord. The cavity 
of the tentacle is almost filled up by the spongy endoderm which 
lines the whole cavity of the animal, so that it remains as a 
narrow canal only in the upper and lower sw^oUen parts of the ten- 
tacles. On the other hand the radial canals contain a wide cavity 
which is clearly separated from that of the tentacle-base by the 
well developed mesoderm. In the next section (Fig. 5, through 
the line 2-2, Fig. 1) cut just inside the margin of the disc, the 
radial canals already assume their characteristic shape in cross- 
section, while the intercalated cords have already lost their cavity 
entirely. Bounded by the mesoderm the intercalated cord as- 
sumes in cross-section the form of a trapezoid. It is convenient to 
distinguish here three kinds of the mesoderm-lamellœ, the upper, 
basal, and the vertical. The upper lamelki {u.l.) is situated along 
the surface of the disc, the basal (b.L), in the lloor {i.e. in the dia- 
phragm), and the vertical [u.l.) connects these two lamelhe. 
When traced inwards, (PI. XV Figs. 6, 7) the intercalated cord 
becomes thinner and thinner, until it no longer shows in cross- 
section the form of a trapezoid, but assumes the shape of a tri- 
angle formed by two vertical and one basal lamella. AVhere the 
gonosome arises (Fig. 8), the vertical lamella does not reach the 
upper lamella ; hence the radial canals communicate here with one 
another and form the upper common cavity of the hydranth. 



246 M. MIYAJIMA : 

Within the circle of the gonosomes (Fig. 0) the upper himella 
stands entirely separated from the basal, on which the vertical 
lamella shows itself only as a ridge-like line which in cross- 
section is recognizable as a simple small knob. 

Preserved in formalin, the fine tissnes of the animal were 
unfortunately mostly gone. Luckily, however, the pieces of the 
gonosomes and the marginal tentacles, which were preserved in 
sublimate, etc., helped us to ascertain something of the histological 
character of the animal. 

The wall of the animal-body, I need hardly say, consists of 
the three layers, ecto-, endo- and mesoderm as in other Coelen- 
terates. 

The ectoderm, the outermost layer, has been entirely shed 
oti' from the s[)ecimen in formalin, but in the pieces fixed with 
sublimate was well preserved. This tissue is a shigle layer of 
cylindrical cells which in their preserved condition are more or 
less vacuolised. There are present a few nematocysts which are 
characteristic of the ectoderm of Coelenterata. 

The mesoderm is a very firm, supporting layer which is 
placed between the ecto- and endo-derm or two portions of the 
endoderm. This tissue was well enough preserved even in 
formalin so that the structure of the animal could be largely 
made out by this layer alone. 

The endoderm, the innermost layer, which lines the whole 
cavity of the animal, remained unfortunately only here and there 
in the specimen in formalin. From these patches it could be 
made out that the endoderm lining the hydranth-cavity is several 
cells thick (Figs. 3 & 10). The cells are irregularly formed and 
contain but a little cytoplasm which is pressed towards the wall 
with the nucleus. Consequently the wall of the cavity gives a 



BRANCHIOCERIANTHUS IMPEEATOR (aLLMAN). 247 

spongy appearance. I can not think that this appearance of the 
endoderm is caused by bad preservation, for the tentacles fixed 
with sublimate show also the same structure. In the preserved 
state, the endoderm forming the upper ceiling of the lower cavity 
of the hydranth has a thickness of o mm. 

In the cross and longitudinal sections (Figs. 11 & 12) of the 
marginal tentacle, the whole of the space inside the mesoderm 
is entirely filled up with a tissue which reminds one of the verte- 
brate notochord. It has the same structure as the spongy en- 
doderm of the hydranth-cavity already mentioned. Only near 
the base of the tentacle, this spongy tissue leaves in the center 
a small cavity which is separated by the mesoderm from the 
hydranth-cavity. Hence the cavity of the tentacle-base is of a 
limited extent, extending not farther towards the distal end, and 
communicating nowhere with the general cavity of the hydranth. 
A longitudinal section (Fig. 2) through the margin of the disc 
shows plainly the relations of the disc and the base of the tentacle 
(the mesoderm being drawn darker than other parts in the figures). 

The gonosomes (Fig. 1, p.) as already mentioned consist 
of the branched tubular stalks, upon which the gouophores 
are grouped in a crowded cluster. Each stalk branches dicho- 
tomously into about the lOtli or 12th order. Each branchlet 
terminates in a group of small globules, of which we recognize 
two kinds (Figs. 13 & 14). The one kind of which there is only 
one in each cluster is situated on the top of the terminal branch, 
while the others take a more lateral position. The former is 
larger than the latter, consisting of the irregularly shaped cells 
mostly vacuolised (Fig. 14, e.g.). In this kind of globule the 
mesoderm of the branch is no longer recognisable and the ecto- 
and endo-derm can not here be clearly distinguished. It seems. 



248 M. MIYAJIMA : 

however, reasonable to suppose that the centrally placed smaller 
cells which are continuous with the endoderm of the branchlet 
belong to that layer. The cells which presumably belong to the 
ectoderm and form the main part of this globule seem to be 
mostly distended. In this globule the nematocysts (Fig. 15, 7i.) 
are found in a large number ; hence the central globule may be 
regarded as the battery. 

The lateral globules (Fig. 14, !.(/.) are mostly spherical and 
consist of compactly packed cells rich in cytoplasm. The meso- 
derm prolonged from the branchlet distinctly separates the 
ectoderm from the central cell-mass. After examining many 
sections I was able to find a few globules which enable us to 
see that the clusters are true gonophores. In such globules 
(Fig. 16), one is able to see that the ectoderm cells at the tip 
are grouped into a mass forming the " bell-nucleus " which 
pushes the endoderm in as a cup. This part of the endoderm 
is arranged into a regular layer one cell deep and is easily dis- 
tinguishable from the remaining part. Owing to the section 
(Fig. 16) having been cut slightly obliquely, the cavity in the 
endoderm seems irregular and very limited. In leality, there 
is a wide cavity occupying the whole interior of the globule, 
which communicates with that of the branchlet. I could not 
detect gonophores developed any further than this in our speci- 
men. January is probably not the season in which the ripen- 
ing of the sexual products takes place. 

The terminal branch thus bears two sorts of globules, the 
one being a nematocyst- battery and the other a true sexual organ. 
Hence the dendritic gonosome of this animal is a peculiar organ 
wliich bears on a common stalk the sexual and defensive elements. 



BKANCHIOCERIANTHÜS IMPERATOR (aLLMAN). 249 

In Other hydroids these two elements are borne on separate stalks, 
as for example in Pennalia. 

Hydrocaulus. 

The under part of the hydranth is prolonged to a shallow 
funnel whose neck corresponds to the hydrocaulus. At about 
the point where the hydranth joins the hydrocaulus, there is 
a circular constriction (Woodcuts 1 & 2). Here the diameter of 
the hydrocaulus is only 9 mm. and from this part down to the 
base it increases in its diameter. Within the constriction is a 
diaphragm (Woodcut 1, I, II, m') separating incompletely the 
cavity of the hydranth from that of the hydrocaulus. In other 
words the circular constriction is the surface expression of the 
insertion of the diaphragm. In the midst of this partition there 
is an opening (Woodcut 1, I, II, C) which puts the two cavities 
above and below in communication. It is about 4 mm. in dia- 
meter and is almost circulai*. The plane of the diaphragm is not 
visibly oblique to the long axis of the hydrocaulus. In the speci- 
mens of 3Ionocaidus imperator in the British Museum, this dia- 
phragm is, according to Dr. Mark, distinctly oblique and the 
central opening is much elongated. 

The hydrocaulus is a hollow tube which has a total length 
of G50 mm. including the proximal end with hair-like appen- 
dages. The hydrocaulus, even when fresh, was collapsed and 
more or less longitudinally folded, so that the exact measure- 
ment of its diameter was almost impossible. Approximately, it 
was 15 mm. just below the constriction, 25 mm. at the middle, 
and 42 mm. at the terminal root. 

The outer surface of the hydrocaulus is smooth. In the upper 



250 M. MIYAJIMA : 

half of it there are visible from outside 15-20 longitudinal wavy 
bands (Fig. 17). They stand about 2-3 mm. distant from one 
another and run down to about the middle part of the hydro- 
caulus where they become obscure. From the surface they look 
remarkably like the mesenterial filaments of an Anthozoon. These 
wavy bands anastomose here and there with one another and 
give to the hydrocaulus of our specimen an appearance much 
resembling that of Corymorpha. Though the bands are in the pre- 
served state still visible, they were more conspicuous when fresh. 
These longitudinal bands show themselves in cross-section (Fig 18) 
as dense spots (:r) in the mesoderm, which have a great affinity for 
any staining agents. From the bad state of preservation of the 
specimen, in which the ectodermal and endodermal cells were 
mostly lost, I could not ascertain whether the wavy bands were 
the endoderm canals, a structure peculiar to Corymorpha^ or not. 
I think it, however, very probable that they existed, and gave 
rise to these band-like appearances. In the published accounts 
of Monocaulus imperator of Allman the endoderm canals were 
plainly described and figured. 

The mesoderm is very well developed, especially in the 
hydrocaulus where it reaches a thickness of about 0.2-0.3 mm. 
This remarkable layer shows itself in the form of a fibrillated 
membrane, which, wheu macerated with caustic potash, is separated 
into two layers, the outer longitudinal (Fig. ID, /./.) and the 
inner circular (Fig. 19, c.L). The former is thicker and stains 
less with any coloring matter than the latter. 

In our specimen there is no sudden bulb-like expansion at 
the lower end of the stalk, such as is described by Mark in 
Branchiocerianthiis urceohis or by Allman in Monocaulus imperator. 
The lowest and broadest part of the hydrocaulus is enclosed for 



BKANCHIOCERIANTHUS IMPEEATOR (aLLMAN). 251 

about 30 mm. from the base in a cliitinous sheath which gives 
an anchorage to the Hydrozoon. With the exception of the upper 
edge the sheath (Fig. 20 s) bears in most parts very numerous 
hair-like processes (ap) of brown color, which are so entangled that 
many foreign bodies {e.r/. Echinus spines, sand grains, dead 
shells) are wrapped up within them. The sheath and the root 
proper are united so closely that they are not to l)e separated from 
each other without tearing. In contrast to the pink-colored 
hydrocaulus the brown color of the sheath with its appendages is 
very conspicuous. 

At the lowest end of the hydrocaulus the wall is very deli- 
cate and has an opening, the margin of which is destitute of the 
hair-like appendages. 

Above the root the mesoderm possesses here and there 
irregular small depressions which are recognized by tolerable 
magnification from the surface as clear spots. Tliese depres- 
sions are also present in the wall of the root enclosed in the 
sheath. 

A portion of the root cut longitudinally (Fig. 21) shows 
that the sheath with its appendages is separate from tlie root 
proper, but has an organic connection with it. The hair- 
like appendage (ap.), which is seen to be a slender hollow 
process of the sheath, embraces in its interior the thread-like 
outgrow^th (o.) of the wall, which perforates the mesoderm and 
is connected directly with the endoderm of the inner cavity of 
the hydrocaulus. Hence it seems to me that the above- 
mentional small depressions in the mesoderm are certainly the in- 
dications of the wart-like processes of the wall of the hydrocau- 
lus as in Corymorpha. 



252 M. MIYAJIMA : 

Summary. 

1. The most striking feature in our specimen is its strongly 
expressed bilateral symmetry as shown by the excentric attach- 
ment of the hydranth to the hydrocaulus and by an interrup- 
tion of the circles of the gonosomes, radial canals, and marginal 
tentacles at the lower edge of the disc. Those who have read 
the above account will, I think, agree with nie in thinking that 
this bilateral symmetry is due, not to the primitive state of the 
body-organization, but rather to its elaboration and specialization. 
We must therefore regard this remarkable case of bilateral 
symmetry in a hydriform person as very different from that ex- 
pressed for instance in the planoblastof Corymorpha and Dicoryne, 
which is but temporary and occurs only at a certain period of 
development, or from the biradial symmetry as expressed in a 
few genera like Ilonobrachium and Lav by a reduction in the 
number of tentacles. 

2. The hydranth-cavity is divided into two parts, of which 
the upper is in its outer part again divided into many radial 
canals visible even on the surface of the disc. That remarkable 
structure is not, however, peculiar to our specimen. For 
example, the hydranth-cavity of Tubularia is divided similarly 
into two parts by a peculiar ring-shaped formation'^' observed by 
several authors. In Tubularia the lower cavity is narrower than 
the upper, so that the former forms a slender canal in the middle 
of the " Wulst." Gosta Grönberg (*98) described in the hydranth 
of Tubularia indivisa slender endoderm-canals which are the same 
in number to the proximal (marginal) tentacles and situated 
between every two tentacle-bases, running obliquely from the cora- 

*0. Hamann ('82) described that formation as " aboral Wulst," G. Grönberz ('98) as 
" Mesoderm-wulst." 



BKANCHIOCERIAMTHUS IMPERATOll (ALLMAN). 253 

mon cavity outwards and downwards. These canals, though not 
visible from the surface, may be regarded as corresponding to the 
radial canals in our specimen, since they both arise from the 
upper cavity of the hydranth and are arranged alternately with 
the marginal tentacles. 

o. The tentacles are filiform and arranged in two sets, oral 
(distal) and marginal (proximal), as is characteristic of the 
tentacles of Tubularidœ, Oorymorphidœ^ and 3Ionocaulidm. The 
cavity of the tentacle is mostly obliterated, being filled up with 
a cellular tissue — a condition very frequently met with in the 
tentacle of the Hydrozoa. 

4. The dendritic appendage is a true gonosome which bears 
in its summit the sexual elements. Our specimen seems to be 
immature, hence it could not be decided whether the gonophore 
is a planoblast or a sporosac. 

5. The hydrocaulus is marked with many wavy bands visible 
from the surface, and possesses a thin sheath with filamentous 
appendages at its lowest end. 

Considerations on the Systematic Position of our 
Specimen. 

Those who would compare the account given above of the 
structure of our specimen with that of Branchiocerianthus urceolus, 
Mark ('98) will not for a moment doubt that we have in these 
two cases essentially similarly constituted animals. It seems almost 
superfluous to call attention to the points of likeness : the hydro- 
caulus with the wavy bands in its upper half and with the sheath 
and filamentous appendages at its base, the hydranth surmounting 
the hydrocaulus, with its radial canals, dendritic gonosomes, and 
two sets of tentacles, all of which show a strongly expressed 



254 



M. MIYAJIMA : 



bilateral symmetry, being interrupted at the lower (or Mark's 
posterior) edge where the hydrocaulus is attached. 

That our specimen and Brachiocerianthus urceolus belong at 
least to the same genus, there can hardly be any doubt. Whether 
they belong to the same species is another question. It is per- 
haps premature to decide this point, at present, in as much as 
Mark has not yet publislied his full paper. Judging from his 
preliminary notice which deals exclusively with the external 
features, the following are the chief points of difference. 
a. The general shape. B. urceolus is stated to have an extremely 
graceful, symmetrical vase-like figure ivith flaring lips. The 
lateral margins of the hydranth-disc were in the natural 
state *' folded in symmetrically from either side, so as almost 
to touch at a point, a little below the middle of the oval. 
This bending in of the margins of the disk produces at the 
upper end of the animal a sort of eccentric funnel-shaped 



depression, 



ct 



A, 



B. 



'•' The fancied resemblance of 
the animal to a little pitcher, 
which this side view presents has 
suggested the specific name 
adopted — urceolus.'" (Mark *98 
p. 148). The pitcher or vase 
shape of the hydranth is thus 
due to two causes : (1) the fold- 
ing in of the disc-margin, and 
(2) the extreme obliquity of the 
axis of the hydranth to the axis 
of the hydrocaulus. In the an- 
nexed woodcut, a represents the 
axis of the hydrocaulus and b 



BEANCHIOCEEIAMTHUS I3IPERAT0K (aLLMAN). 255 

that of the hydranth. Thus these two axes make in B. ur- 
ceolus an extremely obtuse angle as in A, and thus lielp to 
produce the vase-shape. In our specimen the angle which 
these two axes make with each other (B) is much less obtuse, 
and moreover the folding in of the disc margin has not been 
noticed from the first, either in the fresh or preserved state. 
The disc lay flat and open as a disc, and never suggested the 
idea of a pitcher. 
b. The shape of the disc is oval in B. urccolus ; in our specimen 
it is more nearly circular. Moreover in the former, the 
sagittal or longitudinal diameter is greater than the trans- 
verse, while in our specimen it is the transverse diameter 
which is the greater of the two. The following measure- 
ments will make this point clear. 



Sagittal (longit.) diam. Trans, diaiii. Ratio of trans, 

in mm. in mm. diam. to sagittal. 

B. urceolus. 

Small specimen.,, 25 1.5 60% 

Large specimen . . .38 30 nearly 79^ 

iS'colle,^-- SO 00 112.5»/, 



c. The size : — 

T M c i\ 11 Maximum length Maximum length 

Length of the Jivdro- ,. ,, . " , ~ ., i * ? i 

'' 1 . " oi the marginal ot the oral tentacle 

caulus in mm. . i • ° 

ternacle in mm. in mm. 

B. urceolus 105-200 125 30-35 

Specimen of nnç) oqq 50-t"'i 

Science College ^^^ '^^^ '^^ '^'^ 



d. The lower end of the hydrocaidus : — Mark describes a bulb at 
the lower end of the hydrocaulus. In our specimen, there 
is no such sudden enlargement as deserves the name of a 
bulb, although that end is, as has been stated, the largest. 



256 M. MIYAJIMA : 

e. The radial canal: — Mark mentions that the radial canals of 
B. urceolus run *' from the base of the oral tube to the bases 
of the marginal tentacles, before reaching which many of them 
fork, each of the branches communicating with the lumen of a 
single tentacle " (Makk '98, p- 150). The case is very dif- 
ferent in our specimen in which (he radial canals do not 
fork at all and do not communicate with the lumen of the 
marginal tentacles. The latter, on the contrary, are the 
continuations of the intercalated cords. 

Whether these differences are to be regarded as only 
specific or due simply to the diöerences in size, age, etc., we 
must leave for the present an open question. I am inclined, 
however, to think that B. urceolus and our specimen are of 
different species. 

References have already been made several times in the course 
of the foregoing pages to the resemblance of our specimen to 
Jlonocaulus imperator of Allman, a gigantic hydroid dredged 
by the Challenger off Yokohama (stat. o27). The description 
given by Allman of this animal in his report of the Hydroidea 
of that Expedition Ç88) is not as exhaustive as is desir- 
able. He makes no mention of any bilateral symmetry in the 
animal, but we must remember that the specimens which he had 
before him were extremely badly preserved, as he is careful to 
mention, and that the figure of the animal wliich was made on the 
spot by the artist of the Expendition must necessarily have been 
made hurriedly, and as we can testify from our own observation of 
the fresh object, it is very easy to overlook such a feature as bilateral 
symmetry when the disc is lying in the midst of a mass of 
tentacles. Of course the best thing we could do under the cir- 



BRANCHIOCERIANTHUS IMPERATOR (aLLMAN). 257 

cumstances was to appeal to the original specimens. At the 
request of Prof. Mitsukuui, Prof. Mark, who was opportunely 
staying in Europe at the time, was kind enough to examine the 
type specimens of llonocaulus imper^ator, kept in the British 
Museum. The results of his observation were not entirely conclu- 
sive, as the specimens " have so long been in strong alcohol that it 
was quite impossible to make out anything very satisfactorily." 
He naturally made special efforts to ascertain the condition of 
the hydranth — whether it was radially or bilaterally symmetrical. 
In one specimen, he felt tolerably confident, though by no 
means sure, that there was an interruption narrower than in 
Branchiocerianthus u7'ceohis iîi the marginal tentacles. In another 
specimen the central opening in the diaphragm which divides the 
cavity of the hydranth from that of the hydrocaulus was found 
much elongated — a point which in his opinion pointed to bila- 
teral symmetry.'^' He also thought that there is much less ob- 
liquity of the hypostomal region to the axis of the Hydroid than 
in B. urceolus " for the wall of the hydranth between the con- 
striction and the base of the tentacles can be seen to be nearly 
the same height all around, or at least not markedly different on 
opposite sides." This last point is against the view that our 
specimen is identical with 3Ionocaulus imperator, for although 
the disc is much less oblique in our specimen than in B. ur- 
ceolus, as shown above, the hydrocaulus is attached at one end 
of the sagittal (longitudinal) diameter of the disc. But Prof. 
Mark adds, *' the specimens were so much wrinkled and folded 
that I have not much confidence in this conclusion." There is 



* In our specimen the opening wliich puts the hydranth- cavity and that of the hydrocaulus 
in communication is not elongated, but almost circular as already stated. 



258 M. MIYAJIMA : 

one curious point of difference between our specimen and 3fono- 
caulus Imperator. While the hydranth in the Challenger speci- 
men is much smaller than that of our specimen, the stalk is 
enormously longer, being said to reach the almost incredible 
length of 7 feet 4 inches. This is, however, stated to be when 
stretched, and is not the normal length. 

While it is not thus possible to establish absolutely the 
identity of our specimen with 3Iouocaulus imperator of Allman, 
there are on the whole strong probabilities in favor of this 
assumption. Those who read carefully Allman's description will 
notice that the points which he brings out distinctly in the 
structure of his species, such as a wide cavity extending through 
the entire length of the stalk, the presence of the stalk-mesoderm 
in the shape of a fibrillated membrane — a point which Allman 
emphasizes as *' the most striking feature in the histology of the 
Hydroid " — and so forth, are absolutely similar in our specimen. 
If we remember in addition that both came from practically the 
same locality, it is, I believe, within the scope of reasonable- 
ness to conclude that our specimen belongs to 3Ionocaulus imperator 
of Allmann. 

If this is really the case, we must examine other species in 
3ïonocaulus. The genus includes, besides Monoca^ilus impe7'ator, 
two other species ; 31. glacialis, (Sars) (for which Allman esta- 
blished the genus) and 31. pendula, (Agassiz). These two forms 
show, however, a radial symmetry, and now that 31. imperator 
is shown to have a bilateral symmetry, can not possibly be put 
in the same genus with the latter. 31. imperator must therefore 
be separated from the other two species and placed in a new 
genus. According to the rules of nomenclature, this new genus 



BRANCHIOCERIANTHUS IMPERATOR (aLLMAn). 259 

must take the name Branchiocerianthus '^'' first given by Mark, 
and our specimen then ought to be known as 

Branchiocerianthus imperator (Allman). 

P.S. We shall await with interest the full report on the 
specimen of 3Ionocaulus obtained by Prof. Ciirx in his recent 
deep-sea expedition ('09). 

December, 1899. Zool. Institute, Science College, 

Imperial University, Tokyo. 



*Maek ('99) mentions that Rhizonema carnea of S. F. Clarke ('76) may be a form same 
as, or closely related to, Branchiocerianlhus. Clarke's original description is very brief and 
it is impossible to determine wl'.ether Mark's suspicion is correct or not. At any rate, 
Clarke makes no mention of anv bilateral svmmetrv in the structure of the animal. 



260 M. MIYAJIMA : 



Literature. 
Agassiz, L. 

* 62, Contributions to the Natural History of the United States of 

America. Vol. TV. {Boston). 
Allman, G. J. 

* G4. On the construction and limitation of genera among the Hy- 

(h-(jidea. Ann. and Mag. of N. H., 3rr? series, No. 77. 
' 71. A monograph of the gymnoblastic or Tubularian Hydroids. 
{Lo7idon.) 
* 83- 88. Report on the Hydroids dredged by H. M. S. Challenger during 

the years 1873-76. 
Brauer, A. 

' 91. Ueber die Entstehung der Geschlechtsprodukte und Entwickelung 
von Tuhularia mescriihryaniUemum. Zeit für wiss. Zool. 
Bd. 52. 
Carlgren, 0. 

' 99. Branchiocerianlhus nrceoJns E. L. Mark, eine Hydroid ? Zool. 
Aug. Bd. XXII. No. 581. 
Chun, C. 

' 91. Coelenteraten, in Bronnes Klass. und Ord. des Tierreichs. 
' 99. Die deutsche Tiefsee-Expedition 1898/1899. {Berlin). 
Clarke, S. F. 

'■ 76. Report on the Hydroids collected on the coast of Alaska and the 
Aleutian Islands, by W. H. Dall, U.S. Coast-Survey and Party, 
from 1871 to 1874. 

DOFLEIN, F. 

' 96. Die Eibildung bei Tubularia ; Z. für Wiss. Zool. Vol. 62. 
Grönberg, g. 

' 98. Beiträge zur 'Kenntniss der Gattung Tubularia. Zool. Jalirh. 
Bd. 11. Heft 1. 



BRANCHIOCERIANTHUS IMPERATOR (ALL:\rAN). 261 

Hamann, 0. 

'82. Der Organismus der Hydro id-Polypen. Jen. Zeit. Bd. 8. 
Vol. 8. {Jena). 
HiNCKS, Th. 

' 68. British Hjdroid Zoophytes. 1868. 
Von Koch, G. 

' 73. Vorläufige Mittheiking über Cölenteraten. Jen. Zeit. Bd. 7. 
Mark, E. L. 

' 98. Preliminary report on Branchiocerianihus urceolus. Bidl. of 

the Mus. of Comp. ZooJ. Vol. XXXII. No. 8. 
* 99. " BrancMoceriantJms," a correction. ZooL Anz. Bd. XXII 
No. 590. 
Weismann, A. 

'83. Die Entstehung der Sexualzellen bei den Hydromedusen, 1883. 



Explanation of Plate. 
PI. XIV. 



The hydranth with the upper half and the lowest part of the hydro- 
caulus. Nat. size. 

PI. XV. 

Reference letters : B. hypostomal region of the disc ; Ect. ectoderm ; End. 
endoderm ; iZ upper, H' lower, cavity of the hydranth ; ic. intercalated cord ; 
ones, mesoderm ; mi. marginal tentacle ; p. peduncle of the gonosome ; R. 
outer region of the disc, provided with the radial canals ; re. radial canals ; 
t.h. lumen of the base of the marginal tentacle. 

Fig, 1. Surface-view of a portion of the disc with gonosomes (p) and 
marginal tentacles {nit.). Nat. size. 

a upper wall of the disc, h a part of the diaphragm in the hydranth. 

Fig. 2. Longitudinal section through the outermost margin of the disc 
Zeiss a X 4. 



262 



Fig. 3. Cross-section of the upper wall of the lower cavity of the 
hydranth. Zeiss DD x 2. 

Fig. 4-9. Serial sections of the upper part of the disc. Zeiss a x 4. 
Fig. 4. Cross-section through the line 1-1 in Fig. 1, 
Fio- 5 ^-2 

Fig. 6. „ „ „ 3-3 

Fig. 7. „ „ „ 4-4 

Fio- 8. ij-.^ 

Fio- 9. (^-(j 

Fig, 10. Cross-section of the radial canal and intercalated cord. Zeiss 

BBx2. 
Fig. 11. Cross-section of the marginal tentacle. Zeiss 83X2. 
Fig. 12. Longitudinal section of the same. Zeiss a^ x 2. 
Fig. 13. Terminal branches of a gonosome. Zeiss' ax 2. 
Fig, 14. Longitudinal section of a branchlet of the gonosome. Zeiss 
Fx2. e.g. Central, I.g. lateral, globule. 

Fig. 1.^. Central globule with nematocyst {n). Zeiss DD x 4. 
Fig. IG. Lateral globule in which the bell-nucleus {h.n.) and the endo- 
derm-cup (enc.) are fairly well recognizable. Zeiss F x 2. 

Fig. A ]iart of the wall of the hydrocaulus with the wavy bands 
Nat. size. 

Fig. 18. Cros.--section of the mesoderm in the hydrocaulus. Zeiss a x 2. 
h.J. outer, l(mgitudinally, <•./. inner, circularly, striated layer, .r. spot cor- 
responding to the wavy band. 

Fig. 19. Mesoderm of the hydrocaulus, macerated with caustic potash. 
Zeiss a x 2. 

Fig. 20. Surface-view of the root and the sheath. Nat. size. ap. hair- 
like appendage ; .s. sheath. 

Fig. 21. Longitudinal section of the root figured in Fig 20. Zeiss 
a X 2. 0. outgrowth of the endoderm: other letters as in F'vj:. 20. 



X 



\ii 





r 



rig. 1 



Fig. 3 



Jour. Sci.Coll. Vol. XIII. PI. XV. 

Fig. J 6 bn. 




Mutual Relations between Torsion and Magneti- 
zation in Iron and Nickel Wires. 



By 



H. Nagaoka, Rigaknliahtshi, 
Prufe.s^-ur uf Applied IMatheinatics, 

and 

K. Honda, Rigakushi, 
Po.<t-gradiiate in Physic.?. 

WÜh Plate XVI. 

The various effects of stress on the magnetization of ferro- 
magnetic metals are of such a complex character that no simple 
relation seems to exist among them. The strains caused by mag- 
netizing the ferromagnetics are of no less complex a nature, so that 
the co-ordination of these two classes of complicated phenomena 
is, up to the present, still a matter of doubt. Various isolated 
facts, such as the analogies between the change of magnetization 
by longitudinal pull and that of length by magnetization, the 
relation between the twist caused by the interaction of longi- 
tudinal and circular magnetizations and the circular or longi- 
tudinal magnetization produced by twisting a longitudinally or 
circularly magnetized wire respectively, were long considered as 
affording a clue to the explanation of these phenomena. So far 



264 H. NAGAOKA AND K. HONDA : 

as we are aware, no attemi^t lias yet been made to place all of 
these difierent phenomena on a, common footing. Some time ago^\ 
we hinted at the probable connections which exist between 
the twist caused by passing an electric current through a longi- 
tudinally magnetized wire and the change of volume and of 
length in ferromagnetic metals produced by magnetization. The 
said relation can also be extended to the explanation of other 
phenomena ; namely, the transient current produced by twisting a 
magnetized wire and the longitudinal magnetization caused by 
twisting a circularly magnetized wire. It is our object in the 
present paper to show that these different phenomena can be 
linked together in a common bond. 



§ 1. Twist produced by the interaction of circular 
and longitudinal magnetizations. 

The subject was first studied by G. Wiedemann'-'^ who esta- 
blished remarkable reciprocal relations with the longitudinal 
magnetization produced by twisting a circularly magnetized wire. 
Dr. Knott"^ found that the direction of twist in iron is opposite 
to that in nickel ; BidwelP' afterwards discovered that the twist 
in iron is reversed in high fields and takes place in the same 
direction as in nickel. Unfortunately some of the experiments 
were undertaken with wires which were longer than that of the 
coil, so that the magnetization was far from being uniform. It 
will suffice for qualitative tests, but we can not hojDe for any 



l)Nagaoka and Honda, this Journ. 13, p. 57, 1900; Phil. mag. 49, p. 341, 1900. 
2)G. Wiedemann, Pugg. Ann. 103, p. 571, 1858; 106, P- 1<J1, 1859; Elcktridtäi, 3. 
:]) Knott, Trans. Koy. Sue. Edinb., 32 Ü), p. 193, 1882/8;]; 35(2), p. 377, 1889; 36 (2), 
p. 485, 1891. 



TOESION AND MAGNETIZATION. 



265 



definite quantitative results. The position of maximum twist 
in nickel shows a large difference in the present from the corres- 
ponding experiment by Dr. Knott. 

The twist produced by longitudinal magnetization of a cir- 
cularly maojnetized wire was measured in the followin"' way. To 
the extremities of an iron or nickel wire 21 cm. Ions; were 

brazed stout brass wires, and a 
light plane mirror was attached 
to the lower one. The end of 
the lower brass wire was dipped 
in a mercury pool, while the 
upper brass Avire was clamped to 
a small tripod, which rested on the 
top of a magnetizing coil pro- 
vided with hole, slot, and plane 
arrangement. One end of the 
accumulator Avas connected Avith 
the tripod, while the other was 
led to a mercury pool. The 
wire hung vertically in the axial line of the coil, which was 
30 cm. long and gave a field of o7.1'7 C.G.^. units at the 
centre by passing a current of one ampere. The vertical com- 
ponent of the terrestrial magnetic field was compensated by plac- 
ing another coil in the interior of the magnetizing coil. The 
lower part of the wire to be tested Avas protected against air 
current by enclosing it in a Avide brass tube with a small win- 
dow^, just where the reflecting mirror was attached. The twist 
was measured by scale and telescope method, by which the 
deflection of 0.3" per. em. was easily read. The current Avas 
measured by Kelvin graded amperemeters, whose constants were 



Ö 




266 H. NAGAOKA AND K. HONDA : 

from time to time checked by means of aii ampere balance. 
The experiment was conducted in the following manner : — 

1. The circularly magnetizing- current was kept constant, 
and the amount of twist measured by varying the longitudinally 
magnetizing current. 

2. The longitudinally magnetizing current vvas kept constant, 
and the amount of twist measured by varying the circularly 
magnetizing current. 

Before each experiment, care was taken to demagnetize the 
wire completely either longitudinally or circularly by passing an 
alternate current of gradually diminishing intensity. 

Twid hjj varyiiKj the longitudinal field (Fig. 1). — The 
direction of twist in iron, so long as the longitudinal magnetizing 
field is not strong, is such that if the current is passed down the 
wire from the fixed to the free end and the wire is magnetized with 
north pole upwards, the free end, as seen from above, twists in the 
direction of the hands of a watch. By keeping the circular field 
constant, the amount of twist increases at first, till it reaches a maxi- 
mum in a field of about 20 units ; it then goes on diminishing till it 
ultimately changes the direction and continues to twist in the oppo- 
site direction witii increasing field. The field at which the twist is 
reversed increases with the circularly magnetizing field. In nickel, 
the direction of twist is opposite to that in iron, but the general 
feature is similar to iron, the only difterence being that even 
in strong longitudinal fields, the twist is not reversed. For wires 
of the equal thickness, the amount of twist in nickel is greater 
than that in iron — the maximum twist in iron wire of 1 mm. 
diam. by passing 6 amperes through it amounts to about 28" 
})er cm., while with nickel wire of 0.83 mm. diam. under similar 
conditions, the maximum twist amounts to about 200." 



TORSION AND MAGNETIZATION. 267 

Twist by varying the cii^cular field (Fig. 2). — Here we 
notice a slight dissimilarity between iron and nickel. In 
iron, the twist increases with the strength of the circular field, 
if the longitudinal field remains constant. Such is also the 
case with nickel in moderate and strons; fields. In low lonsi- 
tudinal fields, however, the twist does not continue to increase with 
the circular, but we notice a maximum as will be clear in the 
figure.- There is great experimental difficulty in increasing the 
circular field, inasmuch as the heating of the wire becomes very 
great and tlius materially deteriorates the result. 

The hysteresis accompanying the cyclical change of the cir- 
cular magnetization deserves special notice (see Fig. 3). If the 
longitudinal field besuch that with the increase of the circularly 
magnetizing force, the twist reaches a maximum, the curve of twist 
goes below the former course on weakening the circular magneti- 
zation. The twist, however, goes on slowly increasing, till it 
crosses the o??,-curve and then reaches a maximum, wlience it 
gradually diminishes and ultimately vanishes in negative field. 
The course after passing this point is exactly the reverse of that 
already described. The character of twist is exactly the same 
for iron as for nickel, when we take the opposite character of 
twist into account. The nature of the hysteresis is nearly the 
same when the longitudinal magnetizing field is made to vary, 
while the circular field remains constant. 

The results thus for obtained are in accordance with the 
experiments of Wiedemann and Knott ; we have only to notice 
the discrepancy as regards the position of maximum tw^ist in 
nickel. In Dr. Knott's experiment, the said point occurs in 
tolerably high field, while in the present experiment, it occurs 
nearly in the same field as in iron. It may partly be due to the 



268 



IT. NAGAOKA AND K. HONDA I 



difference in the method of measuring the twist and partly to 
the non-uniformity of the fiehl, as was often the case in most of 
the ohJer experiments. 

The observed angles of twist in iron and nickel are exhibited 
in the following tables, where C denotes the longitudinal current 
per sq. mm. in amperes, H the field strength in C.G.S. units 
and r the angle of twist per cm. expressed in seconds. 



Circular Field being Constant. 
Iron wire: fliam.=0.98 mm. 



c= 


= 1.06 


C 


= 3.38 


C = 


-6.95 


H 


T 


H 


r 


H 


r 


5.ß 


18.6" 


.5.6 


21.9'' 


.5.6 


16.4" 


12.4 


21.7 


11.3 


35.6 


11.3 


30.0 


26.0 


13.1 


22.6 


32.8 


22.6 


36.4 


41.3 


7.9 


36.2 


24.7 


49.2 


27.8 


90.5 


2.5 


79.2 


13.7 


78.0 


19.7 


112.0 


1.6 


97.3 


10.7 


96.2 


16.4 


21.5.0 


- 0.5 


191.0 


6.0 


132.0 


8.2 


492.0 


- 1.9 


442.0 


2.7 


455.0 


2.5 



Nickel wire: diam.=0.83 mm. 



C 


=2.45 


C = 


= 4.33 


C = 


= 6.05 


H 




H 


T 


H 


r 


4.9 


- 94.1" 


4.9 


- 97.6" 


4.9 


- 88.6" 


11.2 


-137.0 


11.2 


-164.4 


11.2 


-1.59.8 


24.3 


-125.5 


24.3 


-179.4 


24.3 


-196.4 


38.8 


-102.1 


38.8 


-156.6 


38.5 


-182.6 


.53.6 


- 8.5.1 


.53.4 


-1.34.1 


.53.2 


-1.59.2 


84.0 


- 62.9 


83.8 


-101.8 


84.0 


-119.7 


102.2 


- 54.5 


102.2 


- 88.0 


103.4 


-101.8 


225.0 


- 31.7 


222.0 


- 47.9 


226.0 


- 59.3 


415.0 


- 19.6 


389.0 


- 29.8 


414.0 


- 40.2 



TORSION AND MAGNETIZATION. 



269 



Longitudinal Field being Constant. 





Iron Wiro : diam. =1.05 ram. 




H= 


=2.67 


H =.5.92 


H = 


= 19.38 


C 


z 


C r 


C 


r 


0.35 


0.3" 


0.58 0.3" 


0.40 


0.0" 


0.80 


1.1 


1.03 0.6 


0.89 


0.0 


1.06 


2.0 


1.73 2.0 


1.36 


0.0 


1.80 


3.7 


2.20 3.0 


1.92 


0.3 


2.22 


5.7 


2.96 .5.1 


2.66 


0.4 


3.04 


8.8 


3.55 6.2 


3.48 


0.6 


4.35 


14.2 


4.78 9.1 


4.88 


1.4 


6.79 


21.5 


6.73 13.6 


6.25 


2.5 


7..52 


23.2 


7.34 14.7 


7.16 


2.8 



Nickel wire: diam. =0.83 mm. 



H-- 


= 5.7 


H = 


= 39.7 


H= 


= 96.3 


C 


r 


C 


r 


C 


r 


1.03 


22.1" 


0.95 


22.6" 


0.81 


8.5" 


1.89 


48.1 


2.06 


59.3 


1.76 


19.9 


2.82 


70.2 


3.34 


101.4 


2.82 


32.3 


4.50 


102.0 


4.10 


123.6 


4.29 


50.5 


7.25 


121.2 , 


.5.90 


153.6 


6.22 


73.2 


9.40 


116.4 


7.95 


172.8 


8.65 


99.6 


11.60 


107.4 


11.40 


178.8 


11. .50 


127.5 


13.38 


99.0 


13.10 


176.4 


13.22 


141.6 



§2. Circular magnetization produced by twisting a 
longitudinally magnetized wire. 

By twisting a longitudinally magnetized ferromagnetic wire, 
circular magnetization is developed. If, therefore, two ends 
of tlie wire are connected by a conducting w^ire, a transient 
current due to the circular magnetization appears in the cir- 
cuit at the moment when the twist is applied. Some years ago, 



270 



H. NAGAOKA AND K. HONDA : 



one^^ of us investigated the transient current for iron and 
nickel wires. It was then found that the current due to twist- 
ing was opposite in direction in tliese two metals and that it 
reached a maximum in moderate fields. As the magnetizing 
current was not very strong, no conclusive measurements were 
made as regards the nature of the transient current in strong 
fields. In order to make this point clear and see if any inti- 
mate relation with the Wiedemann effect could be traced, fresh ex- 
periments were undertaken by the. same method as before. We 
have to notice that the ferromagnetic wire was so placed in the 
axial line of the magnetizing coil that it lay in nearly uniform field. 
Some of the measurements of the transient current for iron 
and nickel wires are given in the following table and in Fig. 4. 



Iron 


wire: diaui.= 1.33 mm. 


Nickel 


wire: diam.=: 1.09mm. 




length =20.90 cm. 




length =20.80 cm. 


d= 


15° 


6 = 


50° 


6 = 


15° 


^=50° 


H 


Q X 10 


H 


Q X 10 


H 


Q X id 


H Q X lo' 


3.2 


25.4 


3.2 


30.7 


1.3 


- 5.8 


1.1 - 2.5 


5.3 


29.0 


4.8 


33.6 


3.7 


- 7.1 


2.7 - 6.2 


1.5.6 


24.9 


15.8 


42.0 


16.0 


-17.7 


.5.3 -18.5 


32.6 


16.3 


30.5 


38.4 


33.7 


-21.2 


10.2 -27.1 


54.5 


10.9 


48.6 


30.5 


.53.4 


-21.5 


25.6 -38.0 


87.4 


6.4 


86.0 


18.5 


81.2 


-20.1 


44.9 -41.7 


120.8 


2.5 


139.4 


8.1 


115.4 


-19.2 


67.6 -40.7 


165.6 


1.5 


183.8 


3.7 


1.57.0 


-16.4 


107.4 -40.0 


242.0 


- 0.1 


327.6 


- 3.2 


213.7 


-12.6 


160.2 -35.7 


495.4 


- 1.7 


447.4 


- 5.0 


319.6 


-10.3 


241.5 -29.1 


6.50.2 


- 2.9 


711.0 


- 5.1 


.530.3 


- QSo 


316.4 -26.7 


908.0 


- 3.3 


9.59.0 


- 5.2 


703.0 


- .5.4 


699.8 -13.9 


1298.0 


- 3.1 


1.521.0 


- 4.7 


1214.0 


- 3.9 


1150.0 -10.6 


1790.0 


- 2.5 


1872.0 


- 4.4 


1815.0 


- 2.0 


1853.0 - 8.9 



l)N:igauka, Journ. Sei. Cuil., Tôky.'), 4, p. 323, 1891. 



TOKSION AND MAGNETIZATION. 271 

Here 6 denotes the angle of torsion and Q the time-integral 
of the transient current expressed m C.G.S. units. The resistance 
of the whole circuit was 4.5 ohms. The nickel wire here used was 
made of the same specimen as the nickel prism used in our former 
experiments. 

As is w^ell known, the direction of the transient current, 
and therefore that of the circular magnetization, is opposite in 
iron and nickel. The current for constant amount of twist in- 
creases with the strength of the longitudinal field ; it, however, 
soon reaches a maximum, whence it gradually diminishes. In 
nickel, the transient current attains asymptotic values in strong 
fields without changing its direction, while in iron, it is reversed 
in a field of about 200 C.G.S. units, when the twist is small. 
The increase after the reversal is not pronounced, but becomes 
finally asymptotic. 

§ 3. Longitudinal magnetization produced by twisting 
a circularly magnetized wire. 

The longitudinal magnetization produced by twisting a cir- 
cularly magnetized wire presents the same character as the tran- 
sient current above described. The experiment is very difficult 
on account of the heating of the wire. To avoid the rise of 
temperature, the iron or nickel wire to be tested was covered with 
urushi (Japan lac) which has the special property of being a very 
good insulator while, at the same time, the melting temperature 
is comparatively high. The wire thus insulated was stretched 
in the axial line of a secondary coil whose diameter was 1.5 cm. 
and wdiose total number of turns was 540, and the current of 
cold water was kept fiowing about it to keep the temperature 



272 



H. NAGAOKA AND K. HONDA : 



of the wire uniform. Thus maintaining the electric current 
in the wire constant, it vvas twisted and the induced current 
in the secondary circuit due to the longitudinal magnetization 
thereby developed was measured by the ballistic method. 

Some of the results of observations are given in the follow- 
ing table and graphically shown in Fig. 5. 



Iron 


wire: diam.= 0.888 nnn. 


Nickel 


wire: diain.= 0.965 mm. | 




length = 20.74 


cm. 




leng 


bh =20.94 cm. 


d = 


15° 


d = 


50° 


6 = 


■15° 


/?=50° 


C 


Q X lo' 


C 


Qxio' 


C 


Q X 10 ' 


C Q X lo' 


0.21 


0.7 


0.21 


16.1 


0.21 


- 45 


0.14 -133 


0.85 


3.7 


0.69 


45.3 


0.66 


-111 


0.23 -294 


1.53 


20.5 


1.49 


89.1 


1.56 


-205 


0.90 -397 


2.36 


31.4 


2.19 


111.0 


2.40 


-239 


2.05 -448 


3.93 


36.5 


3.27 


125.6 


3.35 


-256 


2.87 -450 


4.72 


33.6 


4.65 


124.2 


4.36 


-265 


4.42 -452 


6.55 


29.2 


5.91 


115.4 


5.86 


— 272 


7.37 -448 


7.89 


21.9 


8.29 


109.6 


7.95 


-269 


10.34 -431 


12.82 


13.9 


12.48 


86.2 


10.89 


-256 


15.33 -407 


19.01 


10.9 


17.08 


65.7 


14.07 


-243 


20.85 -397 


24.29 


5.8 


24.37 


51.8 


20.18 


-219 


23.04 -378 


28.64 


5.8 


29.14 


43.2 


26.46 


-206 


26.13 -362 



C denotes the total current through the wire expressed in 
amperes ; and Q, have the same meanings as before. 

As will be seen from the figure, the quantity of induced 
electricity in the secondary circuit, and therefore the longi- 
tudinal magnetization developed, by twisting a circularly mag- 
netized iron wire attains a maxinnim, when the mean circular 
field is about 10 units. It then decreases, but in s})ite of the 



TOKSION AXD MAGNETIZATION. 273 

constant stream of water, tlie heating due to electric current 
prevented the experiment from being pushed to the point where 
the direction of the current is reversed. However, to judge from 
the course of the curve, the tendency is such that there is a 
reversal. In nickel, the direction of the induced current is 
opposite to that in iron, and the total quantity of the current 
attains a maximum, whence it continually diminishes, but not 
to such an extent that the current ultimately changes its 
direction. 

These experiments show that the twist produced by the 
combined action of the longitudinal and circular magnetizations, 
the circular magnetization produced by twisting a longitudinally 
magnetized wire, and the longitudinal magnetization caused by 
twisting a circularly magnetized wire are characterized by having 
various peculiarities, which are common to all of them. This 
can not be a mere chance coincidence ; we shall have to ascribe 
these allied phenomena to the same common cause. 

In the experiments of this and the last paragraphs, we were 
assisted by Mr. S. Shimizu, a post-graduate in physics, to whom 
our best thanks are due. 

§4. Theory. 

As already remarked in our last paper on magnetostriction, 
Kirchhoft^s theory can be extended to the study of the relation 
between torsion and magnetization, exactly, in the same manner as 
was done by Maxwell and Chrystal to explain the AViedemann 
effect. There we found tliat the mean circular magnetization called 
into play by twisting a ferromagnetic wire of radius R through 
angle w amounts to 



274 H. NAGAOKA AND K. HONDA 

1 



7^(0 



h" HR. {A) 



in field H, and that the mean longitudinal magnetization caused 
by twisting a ferromagnetic wire carrying an electric current C 
amounts to 



1 



?rW 



2 



h"C. (B) 



The reciprocal relation between these two phenomena is thus 
apparent at a glance. We shall next show how the same pheno- 
mena are reciprocally connected with the torsion produced by the 
interaction of the longitudinal and circular magnetizations. 

The stress components in a magnetic medium as given by 
Kirchhoff' are as follows : 

î;= -(^ + /.• + ^) ß'-+ 4(^ + /■• - ^y ) («^+ ir-+ f). 
^. = -(^ + I- + 4-) r+ |(-^ + /.• - z^-' ) (-+ /5^+f). 

-^,= i;= - y^^ + ic + ^ y«- 

Taking the axis of in the axial line of the wire, and two other 
axes in the plane perpendicular to it, we see that the component 



TOKSION AND MAGNETIZATION. 275 

magnetic forces iu a longitudinally magnetized wire traversed by 
an electric current are 

6C= — li Hill 6, ß= h COS ä, ï = li, 

where h denotes circular iield given by 

, •26V 

C being the current, r the distance of the point from the axis 
of the wire, R the radius, and the angle between r and the 
axis of X. 

The stress components in ferromagnetic medium acted upon 
by the forces above specitied are given by 

-^=-(^ +Ä;+ ^)^'*^^^'^+-^(^ + ^-^') (^^ + ^^). 

z, = -(-^ + k + -^Ç) H^ + |( i- + k -/.')( IP +ir y 

Z, =.Y, =i^_ + /.• + ^^^ h Hsln d, 

X^ = Y, -(-4!- + /^' + ''f ) /i' 6"i d cos 0. 
The moment about the axis of the wire is given by 

= - JJ (^^i- + k + ^\hllrdxdy. 



I 



276 H. NAGAOKA AND K. HONDA : 

Since -x~ and k are very small compared with k", the torsional 
couple twisting the wire amounts nearly to 

^CHR^= ^—^ X Cross section. (C) 



Since the amount of torsion of a cylindrical wire by a given 
couple is inversely proportional to the fourth power of its radius, 
it is evident that for given longitudinal current and held, the 
angle of twist is inversely proportional to the square of the 
radius. This inference was approximately verified in the present 
experiments. 

In deducing the three formulie (A), (B), (C), we can 
not, strictly speaking, put k" outside the sign of integration, 
because the strain coefficient depends on the field strength, which 
is not uniform in a wire traversed by electric current. Hence 
in these formulae, we shall have to use a mean value to obtain 
a close approximation. 

The mutual relations between twist and magnetization are 
embodied in the three formuhe above given. There we notice 
that the strain coefficient k" determines the nature of the three 
different phenomena studied in the above experiments. The 
fact that the coefficient k" is principally determined by the 
elongation in the ferromagnetic metal accounts for the close 
analogy between the said phenomena and the elongations due 
to magnetization. As the above result imports, the analogy is 
not exact, inasmuch as the elongation is also affected by terms 
depending on k', which depends mostly on the change of volume. 

In order to test the consequences of the theory as regards 
the twist produced by the joint action of circular and longitudinal 
magnetizations, we have calculated the twist by assuming the 



TOUSION AND MAGNETIZATION. 277 

values of h", calculated from the changes of volume and of 
length in iron and nickel ovoids. Graphically represented (Fig. 
6.), the fields of maximum twist by calculation coincide nearly 
with that given by experiments, and the reversal of twist in iron 
takes place in low^ fields as actually found by observation. The 
quantitative differences are, however, tolerably large in iron, 
but in nickel the amount of twist is nearly coincident with the 
experimental values. Calculating, in the same manner, the 
quantity of the transient current produced by twisting longitudi- 
nally magnetized wires, we find a close coincidence between the 
experimental and theoretical values in nickel, but the difference 
is tolerably large in iron. In using the strain coefficients, we 
must always bear in mind that these values are widely different 
according to the nature of the specimen ; especially with wires, 
we are not sure of its being magnetically isotropic. The apparent 
discrepancy would probably be lessened, if we could measure the 
twist as well as the strain coefficients on the same sj^ecimen. 
The remarkable qualitative coincidence as regards the existence 
of maximum twist and its reversal in iron are convincing proofs 
that the theory, so far as wo know at present, admits of con- 
necting various experimental facts in a common bond. 

As regards the mutual relations among the three different 
phenomena above enumerated, it will suffice to state that several 
of them have already been noticed by G. Wiedemann in his 
researches on the relation between torsion and mairnetism. He 
especially studied the relation between permanent torsion and the 
effect of magnetizing the twisted wire. The principal object of 
his researches was to expose the different aspects of the phenomena 
involved in the relation between torsion and magnetization in order 
to bring to light his ingenious theory of rotatory molecules. 



278 H. NAGAOKA AND K. HONDA : 

Elegant as it at first sight appears to be, Wiedemann's theory 
abounds with hypotheses which we are not always warranted in 
making. 

In his work on the applications of dynamics to physics and 
chemistry, J. J. Thomson has propounded a new method of 
investigating the mutual relations between the effects of various 
physical agencies. He showed that the existence of a certain 
phenomenon involves as a natural consequence that of another 
reciprocating with it. As an application of his method, he showed 
that if the wire be twisted by the interaction of longitudinal and 
circular magnetizations, a transient current will be produced 
simply by twisting a longitudinally magnetized wire and a 
longitudinal magnetization will be developed by twisting a cir- 
cularly magnetized wire. 

The peculiar feature of Kirchhoff's theory lies in the simple 
and natural way of elucidating the relations between the various 
kinds of strain caused by magnetization and the effects of stress 
on magnetization. Just as we can study the various elastic be- 
haviour of isotropic bodies by knowing the bulk- and stretch- 
moduli, we have to deal, in Kirchhoff's theory, with the strain 
coefficients k' and k" which play the rules of different moduli 
in the theory of elasticity. 

The reciprocal relations between the strain caused by mag- 
netization, and the effects of stress on magnetization, as found 
by actual experiments, will be found to be of paramount im- 
portance in arriving at a correct theory of magnetostriction. 
The strain accompanying the magnetization of ferromagnetic metal 
will be determined, when we know the effects of stress on mag- 
netization and vice versa. As regards the relations between twist 



TOKSION AND MAGNETIZATION. 



279 



find magnetizatioD, we may conveniently place tlieni under the 
following parallel statements : 



Strains produced by magnetization. 

(a) — (Experiment and theory). A 
longitudinallv magnetized wire is 
twisted hy circular magnetization. 

(h) — (Experiment and theory). A 
circularly magnetized wire is twisted 
by longitudinal magnetization. 

(o) — (Experiment and theory). Up 
to moderate fields, the twist produced 
hy the longitudinal and circular mag- 
netizations of an iron wire is op- 
posite to that in nickel. 

(d) — (Experiment and theory). 
The twist due to longitudinal mag- 
netization of a circularly magnetized 
iron or nickel wire reaches a maxi- 
mum in low fields. 

(e) — (Experiment and theory). In 
strong fields, the twist due to longi- 
tudinal magnetization of a circularly 
magnetized iron wire is reversed and 
takes place in the same direction as 
in nickel. 



Effects of Stress on magnetization. 

(a/) — (Experiment and theory) 
Twisting a longitudinally magnetized 
wire gives rise to circular magneti- 
zation. 

(?>'; — (Experiment and theory). 
Twisting a circularly magnetized wire 
gives rise to longitudinal magnetiza- 
tion. 

(c') — (Experiment and theory). Up 
to moderate fields, the transient cur- 
rent, or the longitudinal magnetization 
produced by twisting a longitudinally 
or circularly magnetized wire respec- 
tively, is opposite to that in nickel. 

(d/) — (Experiment and theory). 
The transient current produced by 
twisting a longitudinally magnetized 
iron or nickel wire reaches a maxi- 
mum in low fields. 

(e') — (Experiment and theory). In 
strong fields, the direction of the 
transient current produced by twisting 
a longitudinally magnetized iron wire 
is reversed and is in the same 
direction as in nickel. 



In his paper on the principle of least action, Helmholtz^' 
has placed the reciprocal relations of a dynamical system under 
three heads. Denoting the generalized co-ordinates, the veloci- 



1) Helmholtz, Crelle's. Journal 100, p. 137, 1886 ; Abk., 3, p. 203, 1895. 



280 



ties, the accelerations and the forces by p's, q's, q"s, and P's, 
the relations are g-enerally expressible by tlie equations 



(I) 
(•■2) 
(3) 



9P. 


^P, 


9a 


^Pa ' 


9P„ 


9P. 


Mo 


Ma ' 


^Pa _ 


9P. 


H\ 


9g':' 



It will be easily seen that the relations above cited belong to 
case (2). 

The greatest difficulty that we encounter in establishing the 
relations between the effects of stress on nuignetization and the 
strain caused by magnetization lies in the great difference of 
strain coefficients according to the nature of the specimen. If all 
the experiments be performed in a proper manner on one and 
the same specimen of ferromagnetic metals, we may feel assured 
of being able to discern the true merits of the theory, or to 
detect its various defects, not only from qualitative [)oints of view, 
but also in various quantitative details. 



Jour Sc Coll Vol XIII PI »VI 































~ 


~" 


~ 


















































































































































' 


r~ 




1 


































































































































































































1 




















































































































































- 




















































































































































































































































Y 


K 


1 
















































































































































































' 






































































*• 


H 


> 








































iXn 


r,k 










































































/A\ 










1 


TOI 


H 


in 




o» 






W 


im 






































<iu 






)S 


rm 


, 








-> 


•^ 


^ 
















" 








djfm 




JOJ^m 


1 








































































\^4 


















i 


am 




/(W. 




















































, 


<> 












/ 




?" 






\ 


















































v\ 


X 


6 


































M 






















































































\ 


\ 




/ 








\ 


, 


























\ 






























































\<,\' 






































































s*= 
















\ 


\ 


1 


/ 












~~- 






















































































Sax 






























































r" 


























\ 


1 
























5; 


:^ 










' 


A 
























W»l>w 


































Av<^^ , 
















































































\ ' 


f 




























-r 










!N 




























































V •.^^:;ii4i 


^ 




































































a 










V 


1 






B- 


17 






























1 -^ 


^ 


, 
























































•0 —^ 


IF- 




=^= 


- 


= 




=^ 






J 


Zi 


«- 




-» 


r 


E 


- 






- 


- 




_ 






5 




E 


n 


— 


- 


- 




- 


I 


J[ 


{= 


E 


E 




— 


E 


E 


- 


E 


E 




E 


Eï 


- 


- 


- 




I 






El 


L 


_ 




-■ 


P" 




^^ 


^ 


__ 


r= 




^ 


~__ 


=; 


(5- 




- 


— 


- 


- 


- 


i 


1 1 . ^ 


S 


^ 


„J 


y- 


E^ 




E 


3 


^ 


'l_ 




- 


- 




- 


-U 




- 


- 


- 






- 


- 






J. 






^5 






^ 




; 


- 


- 


- 






= 


-. 


f 


— 




^ 


- 




ü 


t- 


£L 


^ 


-1 




3?E 


- 














L 

V 


g^-^=3 






" 








l~ 


=^ 


dT 


n^ 


L 


E 







E 


= 




- 


- 


- 


- 


- 










^ 




^ 


' 




































































^ 






{ 


/ 
















if 


























, 


/ 














an' 


1 


1 








rf d»/ 


fnumrfni rur^Tii* I, 


, 












i / 




/ 










JVÙ 


M 


w 


re 


It 


im 




0, 


Jl 


un 




— 












































^ 


_; 


< 


' 




' 










f^ 










































/3"cki. 




A,U«l„l,iJ; 


Î tJ-/..U ,1 
















/ 


T 


/ 


















di 


m 






»71 


171 




























































^ 






































'J 




















■' 1 ■ 





















H—/ 


/ 




- 


- 


- 


- 


- 


- 


- 


- 


- 




- 










- 


- 










- 


jp 


Tn 


- 


In 


IL 


Vo 














_ 


- 


_ 


^ 




- 


_^ 




1 












s- 




-. 




- 








_ 




— 






*> 


-Is 


_ 


_ 


— - 


__ 














_ 


_ 


_ 


_ 




T7V 




~ 


" 


~ 


"■ 


"" 




~ 




~ 


~ 




~ 










__^ 


z:- 












"~ 


~ 






~ 


~ 










~ 






~ 


— 


- 




~ 


r- 




1 — 


— ' 




- 


k 




- 




- 




- 




— 




-J 










<-■ 


;'- 


-- 


- 


— - 


-~ 
























^ 






- 


- 






- 




- 


- 


- 


- 


- 




k 
















































-J 


ff 












































































































































■< i// 






























-^ 
























































































































































, 


-J/-L 




















, 




^ 








' " r 




























































F 




5 
























" 




■ 


'. "■ 


~ 




~ 






~ 












~ 






~ 








— 




~ 


— 


— 


— 


— 






11- 


















^ 


& 


i 






_ 






■* 


" 








m 


pi 


rj 


» 


m) 










m 


" 


c. 




























h 


















^ 


" 




" n\ 






"~ 






~ 






"~ 




Tic 


1 




~ 


^ 






~ 




~ 


" 


" 


~ 


~ 




/ 


















^ 




■ 
































































iv> T^« 


























' 


A- 


f,a 


















































1/ 


/ 





















^ 


^ 












\ 


• 






















































n 


»« 




,, 












' 


~ 


r~ 


:Kv 




"~ 






~ 








l-m 


"»r7 




— 


^ 






■~ 




— 


-1 


1 — 








Jlj 




















^ 






\ 














<' 


P 







«3 








































, 


^ 




















ilTlffA 


2)7 




, 












■~ 


~ 


- 


-\n 




~ 


~ 






~ 


~ 










/ 


M 


ïœ 


■~ 






~ 






~i 


~ 


~ 


~ 


fl 


~Y 


L 




















\^ 








^ 
















































































































' 


V 


•*;■ 
















^7.9/n 


.»4< 


m 






















i ! 


1 






















^-s 


\ 


^ 






^3 




J 




































_ 






















v^o 










































ill 


\ 






\ 
































. 






~ 


~ 


"■ 


~ 




























\ 


• 


^ 
















' 


' 






P 








_, 


— 




































-^ 














































\ 




\ 














■ 








■ 






"" 




— 


~ 




~ 


— 




rf^-i 
























A 


^ 






ts 








^ 


=-^ 


_ 


_ 






[= 


—' 






































H- 


^ 










— 




__ 


























~ 


~ 






A 




^ 




\ 


~ 






~ 






~ 








~ 






\~ 






~ 




~ 


~ 




1 


_ 










IJ 






s^ 












\ 




.^ 








_ 








^ 






- 


































^ 


■--f 




-, 






■^ 














































■"=:r 










h 




















-1 


~ 


H 


— 




-A 


- 








I 


z 




I 


L 




J 




_, 






N 


"^ 


N: 


:^ 


>=; 


- 


- 












r 


- 


'-c 




- 


- 






- 




- 






- 


- 


~-i 




- 


- 




— r 


" 




= 






= 




= 


-r- 






- 










— 


— 


' 


- 


- 


— 










- 


J 


^^^ 


F= 


S! 


=: 


:3 


U^ 


3 




3 


E 


= 


= 


--U- 




= n 


- 








P 


— 




P 






- 




- 






p 




S 


5 






■^ 










- 


= 


-, 


T-i 




= 


S 


§ 




^ 


P 


- 






- 


= 


- 


\ 


- 


~ 







— 




- 


- 




- 




- 


I-ns* 


9il 
2?9 


— 








— 


— - 


-- 


— 


~ 


~ 




— 


— 


_ 


- 


_ 


- 


- 


Î- 


- 




~ 


EE 


— 


:5 


r^ 


= 




—I 


■ 


=H=H 


^ 







n 








_ 


_ 




_ 


_ 




_ 




P 




1* 


_ 










V 


_ 




_ 








^ 
































\ 












































" 






— 


~ 




\- 






~ 


~ 


~ 


=^ 




-^ 


— 




^ 


== 









~ 






~ 






' — — 


- 


— 




— ' 


_ 










_ 




_ 


_ 




_ 




_ 






_i 


_ 




«_ 


_ 




i. 


^ 








































- 














s 


_ 




— 


— 




' 














" 




~ 


^ 


— 


— 


-~ 


~ 


^\ 


"■^ 






" 


— 


— 


— 






■^ 




— 


— 






" 


— 


— 




~ 


~ 


:= 


= 


= 


- 


— 




- 








- 


- 


- 


-] 


- 




- 




- 






- 


- 


- 


- 


- 


- 


- 


- 


=i 








- 


- 






- 


- 






- 




- 






-^ 


_ 




V 


ï= 




rr: 


-_z 


_ 


_ 


































"~ 


" 




-] 


t^ 


_ 


_ 


'^ 


= 




' 


-H 




~ 






~ 






H 


" 










" 


~ 


~ 


~ 


~ 




-- 


- 


- 


- 


- 


~ 


- 




- 


- 


- 


- 


- 


- 






- 


- 




- 




- 


- 


- 


- 


- 


- 




- 


- 




r 


»7 

— 


- 






- 


- 


- 


- 




- 


- 


» 


\ 


~-ss 


= 




Zt 


-- 


s 


z 


e: 


- 


E: 


- 


- 


- 


- 


- 


- 


- 


- 


- 




- 


- 


-- 


-- 


-L. 


- 


-M*t 


X 


- 


- 


z 


- 


_ 


z 


^ 


\ 


- 


- 


- 


- 


- 


- 


- 


_ 


.üeiä 


1rs 


Î, 


- 


_ 


_ 



The Interaction between Sulphites and Nitrites. 



Bv 



Edward Divers, M. D., D. Sc, F. R. S., Emeritus Prof., 

and 
Tamemasa Haga, D. Sc, F. C. S., 

Professor, Tokyo Imperial University. 



The present paper gives an account of a series of experiments, 
the results of which seem to leave no room for doubt as to the 
truth of the following propositions respecting the sulphonation 
of nitrites : — (1) normal sulphites are inactive upon nitrites ; 
(2) pyrosulphites are not active to their whole extent upon ni- 
trites ; (3) pyrosulphites are active in their entirety upon nitrous 
acid or its equivalent of nitric oxide and nitric peroxide (nitrons 
fumes) ; and (4) sulphurous acid and nitrous acid or the oxides 
and water equivalent to them interact of themselves and in such 
a way that the base of the sulphite, that may be used in place 
of the sulphurous acid, is needed only to preserve from hydro- 
lysis the products of their interaction. Concerning these asser- 
tions we would point out that the first directly contradicts the 
conclusions drawn by other workers from their experiments ; the 



282 E. DIVERS AND T. HAGA : 

second is novel like the first, the facts on which it is based having 
been misunderstood ; tlie third has indeed been already enunciated, 
but not with intention to limit it to a strict interpretation, and 
only theoretically, without any experimental treatment of it ; and, 
lastly, the fourth has been also made before and upon the basis 
of experiment, but experiment quite inadequate to justify it. 

The establishment of these propositions, taken along with 
what we have already published as to the constitution of Fremy's 
salts will then allow of the further assertion being made, that 
the interaction of nitrous acid with a pyrosulphite results in their 
conversion into a two-thirds normal hydroximidosulphate (and 
water), and that all the other sulphazotised salts are secondary 
products simply derived. It is thus established that the only 
interaction between sulphites and nitrites is one of the greatest 
simplicity, instead of being full of complications, as hitherto 
believed. 

I. — a. A Normal Sulphite inactive upon a Nitrite. 

Dipotassium or disodium sulphite is quite inactive upon a 
nitrite. In establishing this fact wo have mixed solutions of 
normal sulphite and nitrite in proportions varying in different 
experiments, and left them in closely corked flasks, almost full, 
for days and for weeks. No change has ever happened. When 
coloured with rosolic acid, a drop of dilute acid would at any 
time, as at first, discharge the colour. Had any action occurred, 
alkali hydroxide must have been generated (as to the possibility 
of which see sect. II. /;.). A portion of the solution to which 
had been added a drop of dilute sulphuric acid reacquired the 
pink colour of the rosolic acid when left to stand for some time 



INTERACTION BETWEEN SULPHITES AND NITRITES. 283 

— the minute quautity of pyrosulphite which the acid had pro- 
duced having slowly interacted with the nitrite, but when that 
was used up, no more action occurred and, at any time added 
one drop of dilute acid would again remove the colour of the 
solution. 

No further proof of the activity upon a nitrite of a normal 
alkali sulphite is wanted, but additional evidence of the fact is 
readily obtainable. Thus, in the case of the potassium salts, 
while the action of pyrosulphite upon nitrite shows itself (excej^t 
in cases of high dilution) by the formation of the insoluble 
nitrilosulphate, no separation of this salt occurred in the above 
experiments. Again, barium chloride, added, at any time, to the 
mixed and to litmus very alkaline solution of normal sulphite and 
nitrite, precipitated all the sulphur as sulphite (with also a very 
little sulphate) and left the nitrite in solution, neutral to litmus. 
Had sulphazotised salts been present, precipitation of the sulphur 
would have been incomplete, and the mother-liquor would have 
preserved a strong alkalinity to litmus and when acidified soon 
have deposited barium sulphate in the cold and at once if boiled. 

Fremy, Claus, and Raschig all believed in the activity of di- 
potassium sulphite upon potassium nitrite, although the last 
named chemist recognised the value also of the pyrosulphite, as 
Berglund had done before him. Fremy apparently used sulphite 
neutral to litmus and took it to be the normal salt, and Claus 
certainly did so. Since, therefore, they used sulphite which was, 
for the most part, pyrosulphite, no evidence upon the point in 
question can be gathered from their work. We would account 
for Fremy's finding sodium sulphite inactive upon sodium nitrite, 
otherwise inexplicable, by assuming that the solution of sodium 
sulphite which he tried happened to contain no pyrosulphite. 



284 E. DIVEES AND T. HAGA : 

Claus's statement that some potassium sulphite neutral to litmus 
was as active upon nitrite, after lie had added potasli in * excess ' 
to it, as it was before only requires us to assume that the excess 
spoken of was large enough to give the solution a markedly 
alkaline action upon litmus and yet small enough to leave much 
pyrosulphite unchanged. 

I. — h. Poiassluni Hydroxide not a Factor in the Formation of 
the Sidphazotued Salts. 

That a normal sulphite, potassium or sodium, remains still 
inactive upon nitrite, when alkali hydroxide is added, was ascer- 
tained by leaving the three substances together in solution in a 
closed flask for some time, as in the experiments where no 
hydroxide was present, and then precipitating with barium chlo- 
ride after addition of ammonium chloride, and finding no sulphur 
compound left in the filtrate. (Ammonium chloride prevents 
precification of hydroximidosulphate, this Journal 7, 48, 69). 

On the many occasions we have had to prepare sulphazo- 
tised potassium salts by submitting solutions of nitrite and hydrox- 
ide to the action of sulphur dioxide, taking care to keep the 
solution briskly agitated, we found that, even in ice-cold solu- 
tions, precipitation of these very sparingly soluble salts only 
began from the point at which there was no more hydroxide 
left, and then went on freely until the solution had become 
neutral to lacmoid paper. In proportion as the hydroxide dis- 
appeared, sulphite became abundant, whilst from the time that 
the replacement of hydroxide by sulphite was complete, the 
quantity of sulphite steadily decreased as the sulphazotised salts 
formed, sul})hazotised sodium salts being very soluble no preci- 



INTERACTION BETWEEN SULPHITES AND NITRITES. 285 

pitation occurs during their preparation, and with these, therefore 
we made an experiment to determine quantitatively wliat happens 
up to the point when the last portion of hydroxide disappears, a 
point indicated by rosolic acid losing its pink colour. 

Washed sulphur dioxide was sent in a steady stream into a 
solution of 11.21 grams real sodium hydroxide and 19.34 grams 
sodium nitrite in about 198 grams water, kept in active motion 
and immersed in ice. The sodium compounds were in molecular 
proportions; more nitrite would not have mattered. In a short 
time a 10 cc. pipetteful was removed ; soon after a second ; and 
not very long after a third one, just when the pink colour of the 
rosolic acid present had disappeared. The three portions were 
treated alike. Each was mixed with excess of solutions of am- 
monium and barium chlorides and the precipitate filtered off, 
oxidised to sulphate, washed with dilute hydrochloric acid, and 
weighed as barium sulphate. The ammonium-chloride filtrate 
was evaporated to dryness, during which operation all the sul- 
phur of the sulphazotised salts was converted to sulphate by the 
nitrite and ammonium chloride. The soluble salts were washed 
out with dilute hydrochloric acid, and the barium sulphate 
collected and weiohed. The solution of salts removed from the 
barium sulphate was concentrated and then heated under pressure 
for hours, after which it was found to be still clear and there- 
fore free from sulphate, which must have formed by hydrolysis 
had any sulphazotised salt escaped decomposition during the 
evaporation with nitrite and ammonium salts. 

We give the quantities of sulphur dioxide found in each 
pipetteful as sulphite and as sulphazotised salts, and also state 
these quantities as parts per iiundred of the total sulphur dioxide 
which had entered it. 



286 E. DIVERS AND T. HAGA : 

Sulphur dioxide 1st 10 cc. 2ncl 10 cc. 3rd 10 cc. 

as 

Sulphite .0662 grms. =96.6 o/o .r204grms. = 96..59^ .3996 grms. = 01.59^ 
Snlphazot. .0023 „ = 3.4^ .0047 „ = 3..59^ .0365 „ = 8.5«^ 

It will be seen that all but 3.5 per cent, of the sulphur 
dioxide entering the solution in the early stages of the experi- 
ment remained in the form of sulphite, and that even up to the 
time when the last of the hydroxide had been consumed, all but 
8.e5 per cent, of the total sulphur dioxide was in the slate of sulphite. 
That it must be impossible to prevent all temporary local excess of 
sulphur dioxide wdll be at once admitted, as also that it must 
be difficult in the later stages to keep down this local excess to 
very narrow limits. Therefore it will seem in the highest degree 
probable, if not certain, from this experiment that sulphur diox- 
ide, equally with normal sulphite, does not act upon nitrite in 
presence of alkali hydroxide. 

Now Freiuy believed that potassium hydroxide helps the 
formation of sulphazotised salts and endeavoured, accordingly to 
keep some of it always present when passing sulphur dioxide 
into a solution of potassium nitrite. This he did by adding it 
occasionally during the process, and to such an extent that at 
the end of the operation the mother-liquor of his suits was al- 
ways not merely alkaline but caustic and destructive to filter 
paper. Clans did not go so far as to believe that the potash 
exercised any specific influence upon the action between the sulphur 
dioxide and the nitrite, but, agreeing wdth Fremy as to its 
value in precipitating and preserving the sulphazotised salts, he 
adopted the precaution to stop passing in sulphur dioxide when 
the alumina contained in the potassium hydroxide began to pre- 
cipitate, since this occurs while the solution is still strongly 
alkaline to litmus. Kaschig, in attempting to prepare Fremy 's 



INTERACTION BETWEEN SULPHITES AND NITRITES. 287 

sul^jkazate, also used precipitation of alumina as tlie indication to 
stop the process while yet alkali remained. 

Thus, then, it would seem, Fremy, Claus, and Raschig, the 
last in less degree, have all prepared sulphazoiised salts without 
difficulty, under conditions which we pronounce to be in- 
compatible with their production. To remove this apparent 
contradiction in results it is sufficient to assume, for one thing, 
that, in Fremy's way of working, success followed only because, 
temporarily and locally, the point of saturation of the alkali 
was reached and exceeded again and again where the gas entered 
the solution, — a state of things, never avoidable altogether, above 
all at the time when the potassium hydroxide is nearly exhausted- 
There is nothing to show that, to check this, he kept his solu- 
tion well agitated. Secondly, we can assume, with great proba- 
bility that his solution often lost its alkalinity between the additions 
of the hydroxide which he made. Working as we believe he 
actually did, we have found it easy to get results such as his. 
So far as Claus and Raschig followed Fremy's method, their 
results are equally open to objection, while it is to be remarked 
of tlieir alumina indicator that, not only is normal sulphite 
alkaline to litmus but, as we have found, any aluminium present 
is precipitated as hydroxide just when the sulphur dioxide has 
converted all alkali into normal sulphite. They will, therefore 
in their experiments have preserved none of the alkali unchanged 
and most probably have generated also some pyrosulphite. There 
is besides indirect evidence in Claus's work that normal sulphite 
is either inactive or only very slowly active upon nitrite, for when, 
having taken no excess of this salt, he stopped the process just 
after precipitation of alumina, much of this nitrite remained in 
the solution, while, as we have just pointed out, much normal 



288 E. DIVEKS AND T. HAGA : 

snlpliite must also have been present. The two salts were, there- 
fore, together in solution unchanged. Raschig, too, found that 
sulphite and nitrite are inactive upon each other when in presence 
of potassium hydroxide dissolved in only its own weight of water. 

II. — a. Even Pyrosulphite only active upon a Nitrite till 
it has become Normal Sulphite. 

Pyrosulphite, neutral to lacmoid paper and containing there- 
fore, neither sulphurous acid nor normal sulphite, freely sulpho- 
nates nitrite, but is far from being all consumed in the process, 
as it has been represented to be by Claus, Berglund, and Kaschig. 
Quantitative experiments have shown us that, when pyrosulphite 
is left in solution with excess of nitrite in a closed vessel for a 
considerable time, about one-third of the sulphite remains inactive 
by becoming converted into the normal salt, separable, as in 
other cases, from the sulphazotised salts by precipitation with 
barium chloride in presence of ammonium chloride. From this 
it follows that 3 mois, pyrosulphite are needed to convert 2 
mois, nitrite into hydroximidosulphate (this Journal, 7, 19) 
and not 2 mois, only, as had been supposed. The third mol. 
sulphite remains unavoidably in the solution but all the nitrite 
sulphonated :~2NaNOo + SNa^SA + OHo = 2NaoHNS20, + 2Na2S03 
using less pyrosulphite, some nitrite remains at the end along 
with normal sulphite. That sodium pyrosulphite is not easily 
all used up in sulphonating sodium nitrite was observed by 
Raschig. 

Not only hydroximidosulphate but a little nitrilosulphate is 
formed when a pyrosulphite acts upon a nitrite, but this need 



INTERACTION BETWEEN SULPHITES AND NITRITES. 289 

never be enough, with ordinary care, to cause much less than 
one-third of the sulphite to remain inactive. If excess of pyro- 
sulphite is used the interaction appears to be — NaN02 + 2Na2S205= 
NagNSgOc) + NaaSOa, but we have not made any quantitative deter- 
mination of the sulphite remaining, the qualitative evidence being- 
sufficient. 

The interaction between pyrosulphite and nitrite proceeds at 
first very rapidly and with great elevation of temperature, but, 
when the temperature is kept down by cooling, soon slows down, 
so as to require many hours for its completion. The normal 
sulphite seems here to inhibit the action of the pyrosulphite, just 
as does the salt of a weak acid inhibit the action of that acid, 
an effect now well recognised. This consideration points to the 
propriety of looking upon the passage of pyrosulphite to normal 
sulphite as its action as an acid upon the nitrite, and not as the 
yielding up of half of its sulphurous acid for the sul phonation 
of the nitrite, the interactions being 2 NaNOi + Na 38205 + 0112 = 
2HN02 + 2Na2SO„ and then 2HN02 + 2jS'a2S205=2Na2HXS20; 
(see section III ci). 

II. — b. Alkali not produced in the Sulphonation of 
a Nitrite. 

One of the most remarkable things, according to Clans, is 
the production of potassium hydroxide by the formation of Freray's 
salts through the agency of a sulphite. He explained this pro- 
duction by the equation~KN02+2[v,SO,-r 20H2=K2HNS20,+ 
3KII0. Such an equation was also published by Berglund 

o 

(Lunds Univ. Arskr. 1875, 13, 14). Raschig gave the same 
equation for results obtained by himself and, in order to express 



290 E. DIVERS AND T. HAGA : 

Other results, gave also the equation — NaN02+2NaHSOo= 
Na2HNS207 + NaHO, Finding also, and again in agreement with 
Oaus, that dipotassium liydroximidosulphate does not combine 
at once or even at all with potassium hydroxide, he argued that 
this salt cannot have a similar constitution to that of Fremy's 

* basic ' sulphazotate because potassium hydroxide is produced 
along with it instead of being combined with it as Fremy's 

* basic ' sulphazotate. 

Now, all this is wrong in fact on the part both of Claus 
and Kaschig, as we have already shown (this Journal 7, 30) 
or here show in other sections of the present paper, except as to 
the generation of alkali hydroxide, which we now proceed to 
deal with. Claus's emphatic statement, supported as it is by 
Berglund and by Kaschig, that potassium hydroxide is formed 
when a sulphite meets a nitrite in solution, rests upon no other 
evidence than what we now set down in full, recalling the fact 
(section I, a.) that between the normal sulphite and nitrite there 
is really no activity of any kind. A solution of sulphite made 
neutral to litmus and a solution of nitrite of either potassium or 
sodium become hot and strongly alkaline to litmus when mixed 
together, and then contain much liydroximidosulphate and nitrilo- 
sulphate, both neutral to litmus, which soon crystallise out if 
they are the potassium salts. That is all these chemists had as 
evidence for the production of the hydroxide ; let us add to these 
facts that addition of excess of barium chloride removes all the 
alkalinity. It follows, since pyrosulphites are a little acid to 
litmus and normal sulphites are very alkaline to it, that the 
phenomena depended upon offer no grounds whatever for the 
l)elief that alkali hydroxide is produced. Except by the use of 
lime, baryta, or other base, there is, we believe, only one way by 



INTERACTION BETWEEN öULPHITES AND NITRITES. 291 

which potassium hydroxide can be generated from potassium 
sulphite and that is one made known by us, namely, treatment 
of the sulphite first with nitric oxide and then with alcohol and 
water (this Journal, 9, 106). 

III. — a. A Pyrosulphite all active upon Nitrous Acid. 

As remarked at the end of section II. a., a pyrosulphite 
appears to act as an acid upon the nitrite and then sulphonates 
the nitrous acid itself, only indirectly, therefore, sulphonating 
the nitrite of a metal or of ammonium. One third of the pyro- 
sulj^hite should, accordingly, be replaceable by some other acid, 
and so it proves to be (section III. d). It is not new to formuhite 
the sulphonation of HNO2, and to speak of ' nitrous acid' as the 
reacting substance, for (passing over Fremy) Raschig has already 
done so. But, whereas we would be understood to confine the 
activity to nitrous acid itself, or its acidic equivalents (sect. III. d), 
such was not the thought of Raschig, who only wrote H as a 
general symbol, and ' acid ' as a general term, while representing 
metal nitrites as active by generating alkali hydroxide. 

Nor is it new to learn that nitrous acid can be sulphonated. 
By treating a dissolved sulphite with nitrous acid (nitrous fumes) 
Fremy did succeed in obtaining sulphazotised salts, but the dif- 
ficulty of moderating the flow of the gas, and the presence in it 
of nitrogen peroxide and nitric acid made the operation so in- 
convenient, he said, that he did not use it in preparing any of 
the salts he examined, and gave no further attention to it. We 
have taken up the matter, since untouched and unmentioned, 
where Fremy left it half a century ago. Our work has been 
very simple but very eflective, and has consisted in f^^ubjecting a 



292 E. DIVERS AND T. HAGA : 

solution of pyrosLilpliite (and of ijormal sulphite, but of that we 
treat in sect. III. b.) to nitrous fumes which act as nitrous an- 
hydride when of the riglit composition. The gases were found 
to be fully absorbed by a concentrated solution of potassium 
pyrosulphite kept cold in a liask immersed in ice and brine and 
well agitated. Soon an abundant precipitation began of hydrox- 
imidosulphate mixed with a little nitrilosulphate. While still 
much pyrosulphite remained, the process was stopped and the 
mother-liquor at once drained olf. In this way we had great 
success in getting much hydroximidosulphate and only a little 
nitrilosulphate, notwithstanding the presence all along of so much 
pyrosulphite ; for, as was [)ointed out by us long ago, in suffi- 
ciently cold solutions sulphonation to nitrilosulphate hardly 
occurs. 

The next five sections (III. b, c, d, e, f) treat of various 
mixtures which, iioni the acid constitution of one of the com- 
})onents, behave like that of nitrite and pyrosulphite, that is, as 
if each contained pyrosulphite and nitrous acid. 

III. — b. JVormal Sulphite also all active upon Nitrous Acid. 

Keplacing the pyrosulphite used in the last experiment by 
the normal sulphite, it was found that again but in this case 
gradually, hydroximidosulphate precipitated, as well as very little 
nitrilosulphate. But here potassium nitrite proved to be another 
product, which by gradually replacing the potassium sulphite 
in the solution allowed the process to be carried very far to- 
wards completion. The reaction is expressed by the equation — 
:3HNO, + 2K,S03 = 2KNO, + OH, + K,HNSA, ûom which it is 
seen that only one-third of the nitrous acid becomes sulphonated, 
the rest being used up simply as an acid. 



INTEKACTION BETWEEN ÖULPHITEÖ AND NITllITEÖ. 293 

This interaction is what, we believe, Raschig must iiuidver- 
tently have got, when seeking to prepare Peloiize's !?alt (hyponi- 
trososulphate) by the use of nitric oxide. The conditions are 
favourable to the production of the nitrito-hydroxiinidosulphate 
(this vol., p. 222). 

III. — c. Action of Sulphur Dioxide upon Normal Sulphite 

and Nitrite. 

It has been shown in this paper (sect. I. b) that the hydrox- 
iniidosulphate which, from the first, accompanies the normal 
sulphite as joint product of the action of sulphur dioxide upon 
iilkali nitrite and hydroxide, keeps steadily to small proportions 
to the sulphite until nearly all the hydroxide has been saturated. 
After that point is passed and when, therefore, sulphur dioxide 
is meeting a mixture of nitrite and normal sulphite, examination 
of the solution, by the method already described, shows that, 
along with a greater production of hydroximidosulphate than 
before, there is pyrosulphite produced in no insignificant quantity. 
This remarkable growth in the quantity of pyrosulphite, considered 
along with the fact (sect. II. a) that it is itself active upon 
nitrite proves that much of the sulphur dioxide goes altogether 
to the normal sulphite. Only after the greater part of this salt 
has been acidified to pyrosulphite is the sulphur dioxide active 
in sulphonating the nitrite, whicli it tlien does by combining 
with it in conjunction with the pyrosulphite, thus : — 

2KN02+KoSA + 2S02 + OH,=2K2HNS20;, the hydroximi- 
dosulphate being produced in this way with much greater facility 
than by the pyrosulphite alone because of its ^jroduction not 
being accompanied here by the regeneration of normal sulphite 



294 E. DIVEES AND T. HAGA : 

with its inhibitory effect upon sulphonation (sect. II. a). In 
this change it still holds true that it is nitrous acid itself which 
is sulphonated, the potassium leaving the nitrite to enter the 
sulphonate radical, and being replaced by hydrogen. 

Claus held that there could be no difference between the 
effect of submitting a nitrite to the action of a sulphite and that of 
mixing it with a solution of hydroxide and then treating it with 
sulphur dioxide. The contents of this section and section II, a 
show that essential difference exists between the courses and 
results of the two procedures. 

III. — d. Action of Carbon Dioxide and of an Acid Carbonate 
upon Normal Sidphite and Nitrite. 

As would be expected, the gradual addition of one of the 
stronger acids to a solution of normal sulphite and nitrite leads 
to the formation of sulphazotised salts. But even carbon dioxide 
and the acid carbonates of the alkalis are effective in exciting ac- 
tion in a solution of these salts. Concei-ning the activity of 
carbon dioxide there is nothing to add to what was published in 
our first paper (J. Cli. Soc, 1887, 51, 061), that the gas is very 
slowly absorbed by the mixed salts in solution though not by 
either salt alone and at the mean temperature, and that sulph- 
azotised salts are then produced. Normal carbonates of the alkalis 
are inactive. 

It is known that nitrites are not decomposed by carbon 
dioxide, and also that alkali carbonates are decomposed by 
pyrosulphites as freely at the mean temperature as by sulphur 
dioxide itself. Accordingly, we have found that [)0tassium or 
sodium acid-carbonate dissolved along with excess of normal 



INTERACTION BETWEEN SULPHITES AND NITRITES. 295 

potassium or sodium sulphite gives off carhoii dioxide to n cur- 
rent of decarbonated air much to the same extent as when dis- 
solved alone in water. But sodium acid-carbonate may be added 
to an ice-cold solution of sodium pyrosulphite, containing also 
much normal sulphite, and be only very gradually decomposed 
with effervescence. Indeed, an ice-cold concentrated solution 
of normal sodium sulphite will deposit some acid-carbonate when 
charged with carbon dioxide. 

It is, therefore, not surprising that sodium or potassium 
acid-carbonate has a very marked action upon mixed normal 
sulphite and nitrite. When the three salts are left together in 
solution in a closed vessel for a d;iy or two, much sulphazotised salt 
is formed, so that after carbonate and excess of sulphite have 
been precipitated by baryta and barium chloride in presence of 
ammonium chloride, the filtrate from the precipitate when boiled 
with acid gives much barium sulphate and reduces cupric hydr- 
oxide freely. The interaction of the salts may be expressed 
by tlie equation-KNO., + 2K,S03 + 3KHC0, =KoHNS A + -SKoCO, 
+ OH2, but since the two-thirds normal hydroximidosulphate is 
to a small extent converted by normal carbonate into a more 
nearly normal salt and acid -carbonate (this Journal, 7, 32), 
the change expressed by the above equation cannot proceed to 
completion. 

III. — e. Action of Sulphur Dioxide upoji Normal Carbonate 

and Nitrite. 

When sulphur dioxide is added to two mois, nitrite and one 
mol. normal carbonate until the solution becomes acid to lacmoid 
paper, the only products are hydroximidosulphate and carbon 
dioxide. This was long ago pointed out by us, and also that 



296 E, DIVERS AND T. HAGA : 

siilpliite and acid cnibonate are intermediate products, the latter 
of wliieli separates for a, time from concentrated solutions. We 
liavc made fuitlier experiments to ascertain the effect of the first 
portions of the sulphur dioxide in producing hydroximidosul- 
])liato, which, where alkali hydroxide is used, we have shown to 
be insignificant. 

These experiments were carried out in the same way as those 
for testing the effect when sodium hych'oxide is employed (I. b) 
but with the modification of making two pipettings each time 
instead of one, and of weighing both instead of merely measur- 
ing them, then in tlie one we determined tlie sodium, as sulphate 
and used tlie result for calculating what fraction of the original 
solution the other quantity was in which we determined sulphite 
and sulphonates. We thus made ourselves independent of the 
chancre of volume durins; the reaction caused bv loss of carbon 
dioxide and gain of sulphur dioxide. We found in this way, 
admitting of no refined accuracy, that at a later sampling the 
solution contained at most, as much as o^/o per cent, less sodium 
than at an earlier sampling, a diffei'ence however hardly large 
enough to need attention. 

The flask for receiving the portion for the sodium determin- 
ation was previously weighed empty but that for the other 
portion was weighed containing some concentrated solution of 
sodium hydroxide, placed there to arrest all action in the pipette- 
ful (1 topped into it. In the first portion could be seen, by its 
changes on standing, how necessary the sodium hydroxide was 
for fixing the composition of the solution at the time it was 
sampled ; sometimes acid carbonate was deposited by it, some- 
times hardly at all ; sometime« the precipitated acid carbonate 
slowly disappeared sometimes not. The solution used contained 



INTERACTION BETWEEN SULPHITES AND NITRITES. 297 

1 part of sodium nitrite in 4.64 parts of water, besides the cal- 
culated quantity of anhydrous sodium carbonate. 

The results of the experiments showed that hydroximidosul- 
phate was largely produced from the beginning, in proportion 
to the sulphite also formed. Thus, in one experiment, when 2ö 
per cent, of the sulphur dioxide required for complete sulpho- 
nation had been passed in, 55.3 per cent, of it had become 
sulphonate, the rest (44.7 per cent.) sulphite. When 53.6 
per cent, of the sulphur dioxide required had been used, 
74.9 per cent, of it had become sulphonate and 25.1 per cent, 
sulphite. In another closely comparable experiment, when 33.7 
per cent, sulphur dioxide of that required had been absorbed, 
62.7 per cent, of it had become sulphonate and the rest sulphite; 
when 44.4 per cent, of the whole had been used, 72.75 per cent, 
of it had become sulphonate ; and when 62.2 per cent, of the 
whole had been used, 81.5 per cent, of it had become sulphonate. 
That is to say, as for the last statement, when 20.2 grams of 
sodium nitrite (with carbonate) had received 37.5 grams sulphur 
dioxide, 23.3 grams of this had become sulphonate and 14.2 
grams had become sulphite. 

Uniform results are here, however, as when hydroxide is 
started with, only obtained by uniform working, of which the 
following experiment is a good example. A solution of sodium 
nitrite and carbonate was divided approximately into one-fifths and 
four-fifths, and both portions were treated, as nearly as could be, 
alike, their unequal quantities making the onty difference. The 
smaller portion when it had received 20 per cent, of the full amount 
of sulphur dioxide was found to contain 61.8 per cent, of it in form 
of sulphonate, 38.2 per cent, of it as sulphite. The larger por- 
tion, having received 25 per cent, of the amount necessary for 



298 E. DIVERS AND T. HAGA : 

its full snlplionation, was found to have only 53.3 per cent, of 
it as sulphonate and 44.7 per cent, of it as sulphite, as already 
given ; had we stopped here at 20 per cent, sulphur dioxide, as 
we did with the smaller portion, the difference would have been 
more striking still. The difference observed was due to the smaller 
portion having, in relation to its quantity, received sulphur 
dioxide four times more rapidly than the larger portion had, the 
stream of sulphur dioxide having been steady and closely alike 
in the two cases. The result was that local saturation was less 
checked by the agitation of the flask in this case than when 
the much larger portion of solution was under treatment. 

The lack of uniformity in the results here described, does 
not affect in the least the evidence they afford that the sulphon- 
ation of nitrite in presence of carbonate differs greatly in its 
course from that it runs in presence of alkali hydroxide. 

Respecting the formation and destruction of sulphite in the 
process, this salt was observed to be produced rapidly until in 
quantity it had become equivalent to about one-eighth of the 
sulphur dioxide needed for sulphonation of all the nitrite. Then, 
for a time, its quantity remains nearly steady, all sulphur dioxide 
entering the solution during that time becoming sulphonate. 
Finally, it steadily lessens in quantity as more sulphur dioxide 
is added, and disappears just at the end of the sulphonation. 
The more rapidly the sulphur dioxide is blown in at first, the 
less of it becomes sulphite, and the more snlphonates, as already 
stated above. 

One other striking thing observed in these experiments was 
the great variability of the point at which acid carbonate first 
precipitated, as well as the variability of its quantity. With 
quick working acid carbonate precipitated much earlier and in 



INÏEEACTION BETWEEN SULPHITES AND NITRITES. 299 

much larger quantity than in slow working, proportionately, that 
is, to the fraction the solution had received of the quantity of 
sulphur dioxide needed for complete sulphonation of the nitrite. 
Thus, while, with quick working, acid carbonate separated in 
abundance when 20 per cent, of all sulphur dioxide had been 
absorbed, it only precipitated, and then much less copiously, 
when 44 per cent, of all sulphur dioxide had been supplied 
relatively more slowly to the solution. In another experiment 
it showed itself only when 53 per cent, of the sulphur dioxide 
had been added. The main condition, therefore, for early pre- 
cipitation of acid carbonate is rapid addition of the sulphur 
dioxide at first, — the same condition as favours growth of sul- 
phonates at the expense of sulphite. 

Now for the discussion of the results. It becomes highly 
probable from a consideration of these results, together with what 
we know of the several substances concerned, that the first action 
or tendency to act of sulphur dioxide when it enters the solu- 
tion is to convert carbonate into normal sulphite and acid 
carbonate, and to leave the nitrite untouched, and that this ac- 
tion remains prominent so long as much normal carbonate 
is undecomposed. Though this cannot be shown experimentally, 
it is certain that this action does take place, for its products 
present themselves freely, products which could not be derived 
from the sulphonation of the nitrite. Both normal sulphite and 
acid carbonate are active along with sulphur dioxide in sulphon- 
ating nitrite. 

In accordance with what is stated in III. d. the normal 
sulphite and acid carbonate together slowly disappear of them- 
selves from the solution when addition of more sulphur dioxide 



300 E. DIVERS AND ï. HAGA : 

is stopped, owing to sulplionatiou of the nitrite and reconversion 
of acid carbonate to normal carbonate — - 

]S[aNO, + 2Na2S03 + 3NaHC03=Na2HNS20, + 3Na2C03+H20. 
such a mode of sulphonation will therefore be also in operation 
when the entrance of more sulphur dioxide has not been arrested, 
but it is very slow in presence of normal carbonate and may- 
be disregarded as a factor in the process of sulphonating when 
sulphur dioxide is also at work. Here we would insert that only 
to simplify discussion do we speak of normal sulphite and carbon 
dioxide, or even acid carbonate, being together unchanged ; these 
substances, as previously stated, act on each other to a large 
extent in ice-cold solutions, and in uur work we met with pre- 
cipitated acid-carbonate at times when it could only be there in con- 
sequence of carbonic acid withholding sodium from pyrosulphite. 
That in the earlier stages of the process, when much 
carbonate is present, the normal sulphite plays a very small part 
ill the sulphonation not only follows from the observation of its 
rapid increase in quantity at first but is also shown by its then 
nearly constant quantity for a long time though sulphur dioxide 
is still entering the solution and forming sulphonates. Only 
later, as the carbonate gets consumed, does the sulphite become 
an important factor in the sulphonation by freely becoming pyro- 
sulphite, for then its quantity rapidly falls. 

The part played by sulphite in the early stages being thus 
insignificant, we have to seek in the carbonates the source of the 
early considerable sulphonation of the nitrite. It would be un- 
reasonable to assume, with acid carbonate present, that the normal 
carbonate takes part in sulpnonation ; equally so to assume that 
it remains inactive to sulphur dioxide. We are therefore com- 
pelled to recognise that sulphonation goes on only after con ver- 



INTERACTION BETWEEN SULPHITES AND NITRITES. 301 

tioii of all carbonate locally present to acid carbonate and 
sulphite has been effected. Then the reaction that ensues is — 

NaNOs + NaHCOs + 2S0, = Na.H NS,0, + CO, 
When all normal carbonate in the solution has been acidified by 
the carbon dioxide, the sulphite becomes as active as the acid 
carbonate and neither salt gets consumed before the other. 

While it seems certain that first the sulphur dioxide converts 
the normal carbonate into normal sulphite and acid carbonate, 
and only then produces hydroximidosulphate by acting on the 
nitrite along with acid carbonate in the earlier stages and on 
both this and normal sulphite collaterally in the later stages, the 
experimental results show that local saturation must take place 
largely where the sulphur dioxide enters the solution, since so 
much sulphonate is produced along with the sulphite. In con- 
sequence of the activity of acid carbonate, local saturation be- 
comes twice as difficult to prevent as when hydroxide is used 
in place of carbonate. 

If in order to impede local saturation we slacken the rate 
of passage of the sulphur dioxide into the solution, we meet with 
a good amount of success. Thus, it was shown by the results of 
experiments already given, that the slower rate gave proportion- 
ately less sulphonate and more sulphite. But the effect of 
slowness in passing in the gas has its limit, in consequence of 
the continuous though slow interaction which takes place between 
nitrite, normal sulphite, and acid carbonate whereby sulphite 
disappears to give place to sulphonate. It follows that too slow 
as well as too rapid an addition of sulphur dioxide is unfavour- 
able to the accumulation of sulphite, rather than of sulphonate, 
in the solution, and that a medium rate of supply is best for 
raising the proportion of sul[)hite. 



302 E. DIVERS AND T. HAGA : 

There remains to be explained the great variability in the 
commencement of precipitation of the acid carbonate. This takes 
j)lace the sooner the faster tlie sulphur dioxide is blown into the 
solution. When it occurs in the earlier stages of the process, it 
is, therefore, accompanied by greater predominance than usual of 
production of hydroximidosulphate over production of sulphite. 
It does not however depend upon this, ibr while sulphur dioxide 
liberates a molecule of carbon dioxide in changing carbonate into 
sulphite, four mois, of it are needed to liberate one mol. of 
carbon dioxide in changing carbonate and nitrite into hydrox- 
imidosulphate. 

An explanation is suggested by a consideration of the fact 
that wdieu working the process at a moderate rate, the first 
crystallisation of acid carbonate takes place long after the point 
at which the solution must contain the maximum of the salt, at 
least potentially, the point, that is, when half the carbonate has 
become either sulphonate or sulphite. When it does occur the 
quantity of it in solution has become much less. Only where 
crystallisation is started early by a very rapid addition of sulphur 
dioxide, does the acid carbonate continue to separate out in much 
such quantity as it could do at the stage of the process reached. 
The cause in one word is supersaturation. The acid carbonate, 
it would seem, is slow to begin to precipitate from the solution 
while that is not charged with carbon dioxide. At a medium 
rate of working this only happens in the later stages, any normal 
carbonate and even much normal sulphite present keeping down 
tlie quantity of carbon dioxide, but by a rapid rate of working 
local saturation occurs and the acidified portion of the solution 
then crystallises. Once crystallisation has been started, it proceeds 
unchecked. In slower working when crystallisation only begins 



INTERACTTOX BETWEEN SULPHITES AND NITRITES. oOo 

late in the process, the amount of salt separating is small, and 
generally depends then for its existence npon its ])Ower to resist 
the action of acid sulphite in ice-cold solutions. The solution 
when, potentially at least, it is richest in acid carbonate, was 
found by us to crystallise soon, if left to stand in closed vessel, 
although sulphonation which is destructive of acid carbonate was 
slowly going on in it. 

III. — ;/. Primary Action of Sulphur Dioxide upon a Nitrite. 

Solution of sulphur dioxide added to that of potassium or 
sodium nitrite produces a sulphate and either nitric or nitrous 
oxide, according as one or other of the interacting substances is 
in excess. That is the ordinary well-known result, but there are 
two ways of limiting the extent of the action so as to get either 
hydroximidosulphate and nitrous acid or the undoubted products 
of their transformation. By these ways, the interaction of sul- 
phur dioxide and a nitrite is shown to be — 

2KNO2 + 2S0, + OH, = K.HNS A + HNO,. 

The more important way to thus limit the action is by an 
experiment first tried by Glaus {Ber., 1871, 4, 508 ; see preceding- 
paper) which consists in adding an alcoholic solution of sulphur 
dioxide to excess of potassium nitrite in strong aqueous solution. 
For this experiment gives, as we have ascertained, potassium 
nitrito-hydroximidosulphate which precipitates and ethyl nitrite 
which boils off' by the heat of the reaction : — 

3KNO2 + 2SOo + CoHoO = KN02,K,HNS,0, + CM, NO,. 
By becoming ethyl nitrite the nitrous acid is rendered inactive 
on the hydroximidosulphate, which is thus saved from oxidation. 

The other way of tracing the earlier action of sulphur dioxide 



304 E. DIVERS AND T. HAGA : 

lipon a nitrite was found out by Raschig, when trying to prove 
anotlier point (sect. IV. a.). He added the nitrite to excess of 
sulphur dioxide, both being in very dilute and well cooled 
solution, evaporated down find neutralised the solution with 
chalk, and again evaporated the filtered solution. After much 
potassium sulphate had crystallised out, potassium amidosulphate 
was finally obtained, as proof that hydroximidosulphate had 
been formed at an earlier stage. Our own experiments have 
yielded us an earlier product of the degradation of this com- 
pound. 

At the time when Kaschig published his observation, we 
published {J. Oh. Soc, 1887, 51, 659) one of ours, that silver 
nitrite and mercurous nitrite, when decomposed by sulphur 
dioxide solution, yield a substance answering to the copper test 
for hydroxylamine. This we now know to be hydroxyamido- 
sulphuric acid, but at the time we took it to be hydroxylamine 
itself. We have also found that, after adding a dilute solution 
of sodium nitrite to excess of a cooled solution of sulphur dioxide 
and then blowing out of the solution the residual sulphur diox- 
ide by a current of air, enough hydroxyamidosulphate (hydrolysed 
hydroximidosulphate) is present to be easily identified by the 
copper test for it. A hydroxyamidosulphate is distinguishable 
from hydroxylamine in applying this test by finding that the 
mother-liquor of the cuprous oxide (which need not be filtered 
off) gives sulphurous acid when acidified (this Journal III, 225). 

Though less successful than Claus's experiment, Raschig's 
method is serviceable for showing that the alcohol used in that, 
plays only a secondary part. While excess of nitrite is success- 
fully used in that experiment, the sulphur dioxide must be in 
excess in Raschig's method. To understand this, it has only to 



INTERACTION BETWEEN SULPHITES AND NITRITES. 305 

be remembered, firstly, tliat nitrous acid would oxidise liydrox- 
imidosulpliate at once, and secondly that sulphurous acid sulplion- 
ates the hydroximidosulphate slowly enough to allow a little of 
it being secured in a hydrolysed state. 

IV. — a. Sulphonation of Nitrous Acid by Sulphurous Acid. 

Fremy believed that certain of his sulphazotised salts are 
formed in the first action of sulphurous acid upon nitrous acid. 
From this belief Clans strongly dissented, holding that the presence 
of a base (as salt) was essential to the production of these acids. 
Raschig considered that his experiment of treating potassium nitrite 
with sulphur dioxide in excess (sect. III./.) proved the correct- 
ness of Frerny's belief ; but that cannot be admitted since potas- 
sium is present in this experiment playing the part of base. It 
is, however, quite piacticable to establish Frerny's belief and that 
no base wdiatever is necessarv to brino' about the formation of 
sulphazotised acids. 

When a solution of sulphur dioxide, better ice-cold, is treated 
with a relatively small quantity of nitrous fumes passed on to 
its surface while it is being well agitated in a flask, and is 
then deprived of remaining sulphur dioxide by a rapid cur- 
rent of air, or even by quick boiling, it will give a good reaction 
for hydroxyamidosulphuric acid with the copper test. A 
little deviation in the composition of the nitrous gas from that 
of nitrous anhydride is not of importance. If the object is only to 
get amidosulphuric acid, the solution of sulphur dioxide is left 
to stand for a day after it has received the nitrous acid without 
expelling what is left of the sulphur dioxide. If it is then 
evaporated on the water-bath and further concentrated in the 



306 E. DIAŒRS AND T. H AG A : 

vacuum-desiccator, the amidosulphnric acid will crystallise out 
from the sulphuric acid with which it is accompanied (this 
Journal, g, 230). We have purified tlie acid by recrystallisa- 
tion, and have hydrolysed it ;it 150°, by means of hydrochloric 
acid, into acid ammonium sulphate ; we have also completely 
volatilised the acid by heat thus proving the absence of base 
accidentally derived. 

Nitrosyl sulphate dropped into much excess of cooled solu- 
tion of sulphur dioxide also yields the hydroxyamidosulphate re- 
action with copper sulphate and potassium hydroxide. 

IV. — b. Influence of the Base of the Nitrite or Sulphite. 

Although Fremy held that sulphurous and nitrous acids 
combine together, he did not believe that the resulting sulph- 
azotised acids could be obtained in this way, because of their in- 
ability to exist in absence of a base. Moreover, he considered 
that a strong base is influential in bringing about the formation 
of these acids, even though he had had no success with such a 
base as sodium. The only hydroximidosulphates he could prepare, 
indeed, were those of potassium, but from ammonium nitrite he 
got the nitrilosulphate, and also obtained evidence that calcuim, 
strontium, and barium nitrites are convertible into amidated 
sulphates. 

We have just shown (sect. IV. a.) that the interaction of 
sulphurous and nitrous acids does not require the presence of 
any base at all for the actual production of sulphazotised acids, 
although such presence is essential to preserve unchanged the 
first product of the interaction. To serve this purpose some 
bases will doubtless be inferior to others, and those w^hicli do 



INTERACTION BETWEEN SULPHITES AND NITRITES. 307 

not freely form soluble pyrosulphites are difficult to work with. 
Otherwise, the luitiire of the base seems to be a matter of in- 
difference. Since the time of our early publications on the sub- 
ject, we have extended our experiments to several other nitrites 
than those of sodium, mercurosum, and silver, with the results 
we now record. 

Ammonium salts. — Ammonium nitrite solution was prepared 
by triturating silver nitrite with its equivalent of ammonium 
chloride dissolved in about five times its weight of water, and 
filtering off silver chloride over the pump. To this solution, 
after it had been cooled in ice, w^as added a little less than its 
equivalent of ammonia-water which had just before been con- 
verted to sulphite by passing sulphur dioxide into it. jMore 
sulphur dioxide was then passed into the mixture until it red- 
dened lacmoid-paper. In this way the ammonium nitrite was 
almost all sulphonated, without any evolution of gas having 
occurred till just at the last, when slight nitrous fumes appeared. 
Some of the solution was hydrolysed and tested then with copper 
sulphate and potassium hydroxide ; it was thus shown to have 
contained abundance of ammonium hydroximidosulphate. 
Another portion of the solution not hydrolysed gave a large 
precipitation of dipotassium hydroximidosulphate on addition of 
potassium chloride. 

Barium salts. — Some barium hydroxide was converted into 
sulphite by putting it in water and passing in sulphur dioxide ; 
the barium sulphite was then, for the most part, brought into 
solution by passing in more sulphur dioxide. The product was 
added gradually to a solution of a little more than its equivalent 
of barium nitrite, which had been purchased of excellent quality. 



308 E. DIVEES AND T. HAGA ! 

Having neglected to cool our solutions we had reason to fear 
tliat our experiment was a failure; for along with very much 
precipitation there was a somewhat large evolution of nitrous 
gases. But for our purpose we had been amply snccessful. The 
solution was only faintly acid to litmus and remained so for 
hours. Both it and the precipitate contained lai'ge quantities of 
barium hydroximidosulphate. The precipitate also contained 
sulphite and sulphate, the latter being the complement to the 
nitrous fumes produced. The hydroximidosulphate was extract- 
ed from the precipitate by a solution of ammonium chloride. 

Calcium salts. — A solution of calcium nitrate, free from 
magnesium, sodium, potassium, and other ordiary impurities, was 
heated with well-washed spongy lead until nitrogen oxides and 
ammonia began to form. The filtered, very alkaline, solution 
was freed from lead by hydrogen sulphide not used in excess. 
Calcium hydroxide was then removed by carbon dioxide, (it was 
interesting to find that, contrary to assertion, carbon dioxide 
cannot be used to precipitate lead in presence of calcium salt, 
since calcium precipitates before lead.) A solution of calcium 
sulphite in sulphurous acid was prepared just before use, in the 
same way as the barium salt had been, except that carefully pre- 
pared calcium carbonate took the place of barium hydroxide. 
With the calcium nitrite somewhat in excess of the calcium 
sulphite, the solution of the latter was gradually poured into 
the former, both solutions having ice floating in them at the time. 
No gas was given off and only a moderate quantity of precipi- 
tate was formed, which consisted of sulphite. The filtrate was 
neutral and contained the full quantity of hydroximidosulphate 
expected. 

Zinc salts. — Zinc nitrite solution was prepared by précipita- 



INTERACTION BETWEEN SULPHITES AND NITRITES. 309 

ting zinc sulphate witli barium nitrite and filtering. Zinc sul- 
phite in solution in sulphurous acid was made from zinc oxide 
in water and sulphur dioxide. The two solutions, suitably pro- 
portioned and with ice floating in them were mixed. iSTo gas 
came off, zinc sulphite precipitated, and the solution proved to" 
contain zinc hydroximidosulphate present in it in large quantity. 
Mercurous salts and silver salts. — Experiments, already re- 
ferred to in sect. III. /. of this paper, sufficiently establish that 
mercurous and silver nitrites are readily sulphonated. It is nOAV 
evident that the sulphouation of nitrites is a general reaction, 
essentially independent of the nature of the base, which only 
effects the preservation of the products. It is not the salts which 
are sulphonated but nitrous acid itself. 

V. — What Nitrous Acid becomes when Sulphonated. 

In the paper preceding this it has been established that 
neither the abundant experimental work of other chemists and 
ourselves nor theoretical considerations afford any support to the 
view that the double sulphouation of nitrous acid into a hydrox- 
imidosulphate occurs in two stages, or that a monosulphonated 
nitrous acid, ON'SOgH or (HO)2N'S03lI, must be the first product 
of its change. In the present communication it is shown that 
the acidity necessary for the sulphouation of a nitrite points 
clearly to the fact that it is in every case the acid itself, and 
not its salts, which is directly sulphonated. We are, therefore, in 
the position to affirm that the fundamental action in the form- 
ation of all Fremy's sulphazotised salts is the interaction between 
actual nitrous acid and a pyrosulphite, in which they unite al- 
ways to form the one substance, the two-thirds normal hydrox- 



310 DIVERS & HAGA : INTERACTION BETW. SULPHITES & NITRITES. 

imidosiilpliate corresponding to the pyrosulphite acting — 
HONO + (S02K)-S03K=HON(S03K)2. The origin of all the other 
salts out of this salt has been traced, partly by others and part- 
ly by ourselves, and need not be gone over again here. 



Contributions to the Morphology of Cyclostomata. 

II.— The Development of Pronephros and 
Segmental Duct in Petromyzon.^^ 



By 



S. Hatta, 

Professor in the College of Peei's, Tokyo. 



With Plates XVII- XXL 



The following pages contain the second of a series of 
studies on the later stages in the development of Petromyzon, the 
first having already been published some time since in this 
journal ('97, vol. x, pp. 225-237). 

Our knowledge of the earliest development of the excretory 
organs in the lampreys is still somewhat incomplete. This 
circumstance is, I believe, mainly due to the want of recent 
investigations upon the subject. Since the appearance of the 
works by Müller (75), Scott ('82), Shipley ('87), Goette 
('88), KuPFFER ('90), and others, ten years or more have 

1) It was my intention to publish this paper shortly after the appearance of my preliminary 
notice in 1897, (Annot. zool. Jap., vol. I. pp. 137140) but various unavoidable circumstances 
have combined to cause the delay. Meanwhile I have had opportunities of renewing my study 
on various points and the results here given are different from tliose of tlie preliminary 
puper in several important respects. 



312 s. H ATTA : 

elapsed, and, so far as I am aware, no important fact has been 
added during the interval by any renewed researches. I need 
not, therefore, apologise for the ^publication of the present paper 
which embodies the results of my study on the subject during 
the last few years. 

The investigatio]! of a longitudinally stretched, or a meta- 
merically arranged, organ-system such as the neural canal, the 
chorda, the pronephros, &c., is rendered peculiarly difficult 
in Petromyzon by the fact that the longitudinal axis of the 
embryo in early stages describes a semi-circle. Some sections 
in a series of cross-sections of such an embryo are therefore 
unavoidably cut in planes which meet the longitudinal axis 
of the embryo in variable degrees of inclination ; consequently a 
structure stretching in the direction of this axis is cut through 
obliquely, as, for instance, the neural cord shown in figs. 2 and 
3, PL xvii. The vertical dimension of the cord is not in reality as 
long as is represented in these figures. To gather accurate notions 
of the form, the position, &c., of a given structure, therefore, 
it is necessary to compare series of sections of two or more 
embryos of as nearly the same age as possible. Further, the 
difficulty of observation is greatly increased by the crowd of 
yolk -granules in cells, especially by their reaction against stain- 
ing fluids. Certain fluids such as haematoxylin, borax- carmine, 
&c., either stain diffusely all the parts, or act on the granules 
more intensely than on the other contents of cells, so that we 
can not discriminate different kinds of tissues. This difficulty 
was, however, obviated b}' employing picro-carmine. The 
embryos were stained in toto in this fluid, decolorized to the 
proper degree in acid-alcohol, and then washed in 909^ alcohol. 
In the sections of specimens thus prepared, the histological struc- 



MOEPHOLOGY OF CYCLOSTOMATA. 313 

tures are distinguishable very clearly, being almost entirely dis- 
colored in all parts except nuclei which are stained intensely. 

I wish here to express my warmest thanks to my former 
teachers, Pkof. Miïsukuki and Prof. Ijima, for much in- 
valuable advice and for their kindness in looking through the 
manuscripts and the proof-sheets of this paper. 

To avoid confusion the present paper will be divided into tw^o 
sections, the first of which will contain mere descriptions, while 
in the second will be given a historical review^ and conclu- 
sions. 



I. Descriptive. 

A. — The Pronephros. 

The youngest stage of the embryo which I have to deal 
with in the present investigation, is only a little advanced beyond 
an ellipsoidal gastrula ; it is intermediate between Stage I and 
II of the list given in my first contribution above referred to 
(fig. 1, A. and B, of that article). The head-fold forms a pointed 
protuberance at one pole of the ellipsoid, while the blastopore 
is plainly visible at the opposite pole. The prominent neural 
ridge extends longitudinally from the anterior end of the head- 
protuberance to the dorsal lip of the blastopore. 

For the general relation of the germinal layers at this stage, 
I refer to fig. 1 which is drawn from the same series as the 
section represented in fig. 18 of my former work.^^ It represents 
a transverse section though the dorsal region of an embryo in 
the stage above described. As seen in this figure, the solid 

1)S. Hatta, On tlie form, of the germ. lay. in Petr.: tliis Journal, vol. Y, 1S91. 



314 s. H ATTA : 

neural cord is interposed in the median line between the right and 
left halves of the mesoblast, underneath the epiblast composed by 
of a single row of columnar cells. The epiblast is always limited 
with a sharp contour against the mesoblast. The latter consists, on 
each side, of a dorsal id.), a ventral {y.) and a median {m.) row 
of cell.«. The dorsal row represents the parietal, and the median 
and ventral rows together form the visceral, layer. Both the 
visceral and parietal layers of the mesoblast show, in the prox- 
imal portion, a regular arrangement of a high columnar epithel- 
ium, while distally this arrangement is more or less disturbed. 

The above structures are localized in a small portion on the 
dorsal aspect of the ovoid embryo, the remaining larger part 
being taken up by yolk-cells compactly loaded with yolk-granules. 

In this early stage, we can detect, therefore, neither in the 
epiblast nor in the mesoblast any structure whatever which is to 
be regarded as the rudiment of the pronephros. 

Period 1. 

In the next following stage, which corresponds approximate- 
ly to the early part of Stage ii {loc. cit., fig. 1, B), certain alter- 
ations are met with in the mesoblast. The most important of 
these is its metameric segmentation. This process first begins at 
the neck^^ and proceeds both forwards and backwards. At the 
present stage, there are found 16 or more mesoblastic somites.^^ 
At about the time when this process has extended to the 
anteriormost section of the mesoblast a second change arises in 
the mesoblast, viz. the first appearance of the pronephros. 

l)The term neck is used for tlie sake of convenience to designate the slender region where 
the liead-fold ])asses over into tlie hind ghjbiüar part. 

2) The exact number of tlie somites can not be reckoned, for the metameres become 
indistinct posteriorly. 



MOKPHOLOGY OF CYCLOSTOMATA. 315 

Fig. 2 represents a section through the middle of the fifth 
somite^^ of the embryo mentioned above. The general features of 
tlie germinal layers and of other primitive organs are essentially 
the same as before. The epiblast {ep.) is a single row of columnar 
cells and is sharply bounded from the structures beneath it ; the 
neural cord (71.) remains still solid.^^ In the mesoblast, however, 
two portions are distinguishable : the proximal portion composed 
of high columnar cells {mt.V and a.pn.2) which undergoes 
metameric segmentation, and the distal portion consisting of 
a loose group of somewhat irregularly shaped cells (Im.) which 
remains unsegmented and constitutes the lateral plate. It 
is noteworthy that the former takes up the largest portion of 
the mesoblast, while the latter is represented by a small portion ; 
these two portions represent respectively the parts of the same 
name in the mesoblast of AmpMoxus. However, between them 
there exists no distinct limit in the lamprey ; the one passes 
gradually over into the other. Although the visceral layer shows 
no sign of constriction, the parietal layer is notched at about the 
middle of the proximal segmented portion (.i). The parietal layer 
distal to this notch is composed of a regular cylindrical epithelium 
{a.pn.2), which is slightly arched against the epiblast, so as 
to cause an indentation in the latter, wdiile the visceral layer 
of the corresponding portion consists of a more or less disturbed 
row of high columnar cells. As the subsequent history teaches, 
the proximal half of this extent {mt. V) represents the myotunie"^ 



1) The somites are reckoned from the anterior end. Tlie first, i.e. the foremost lies 
immediately behind the auditory vescicle when the vescicle comes into view. 

2) The vertical diameter of the neural cord in tigs. 2 and 3 is shown greater than it 
really is, the sections passing obliquely owing to the bending of the longitudinal axis of 
the embryo, as noted in the introduction (p. 312). 

3) This term liere means the Sclero-rayotom of German authors. 



316 s. H ATTA : 

and the distal portion (a.2)n.2) constitutes the Anlage of the 
pi'onephros — the name I assigned to the same in my preliminary 
paper ('97). To avoid tiresome reiteration, 1 shall often speak 
of them in the following pages simply as the " xinhige " and when 
it is necessary to refer to special ones, as ihe Anlage lirst, the 
Anlage second, etc. in tlie order of their position in t1ie series of 
mesoblastic somites, beginning from the anterior end. 

In the somite next following, i.e. the sixth (fig. 5), the 
mesoblast shows almost the same condition as that already des- 
cribed ; but in the somite preceding the fifth, i.e. in the fourth, 
the Anlage of the pronephros is a little more advanced in 
development (fig, 3). On the left side of fig. 3, the fourth 
myotome is sliced only at its hind wall {mt.IV), while, on the 
light, it is cut through in the middle {mt.IV). On the right 
half of the section, no marked progress is visible in the meso- 
blast except the separation of the myotome, which shows a pen- 
tagonal outline {mt.IV) and consists of high cylindrical cells 
from the lateral plate formed of a loose mass of cells {hi.). In 
the left half, however, the state of things is quite otherwise : 
the Anlage of the pronephros {a.pn. 1) together with the corres- 
ponding visceral layer is entirely constricted off from the myotome 
{ml. IV), although it. is still connected with the lateral plate {Im.). 
The cells composing the Anlage are compactly set together and 
arranged more or less in a radial manner ; the Anlage itself is 
rounded off at the proximal end. The lateral plate, on the other 
hand, still consists of loosely grouped cells of variable shape. 

The Anlage is thus always (before and after its separation 
from the myotome) histologically very distinct from the lateral 
[)late ; one might therefore often be misled to suppose that there 
is no organic connection between these two structures. 



MOEPHOLOGY OF CYCLOSTOMATA. 317 

Fig. 4 represents a section passing between the two somites 
above mentioned (the fourth and fifth) and is ranch magnified 
(Zeiss, E^ 2) to iUnstrate the finer structure of this portion. The 
structural cells are all loaded with an enormous quantity of 
ovoid corpuscles or yolk-granules. The epiblasi {ep.) consists 
of a single row of cubical cells and shows a sharp limit against 
the structures inside it. The irregularly polygonal mass of 
cells [mi.V) is the anterior wall of the fifth myotome. Two 
rows of variously shaped cells (Im.) constitute the lateral plate 
which is histologically quite like that in the soraitic portion, being 
composed of irregularly quadratic cells and tapering towards 
the distal (ventral) extremity (compare with the lateral plate, 
Im., in figs. 2 and 3). However, in the proximal portion, 
where the Anlage of the pronephros consisting of a regular row 
of tall columnar cells w^ould be found in the somitic portion, w^e 
see here a group {x) of a few cells of faint appearance, forming 
the proximal edge of the lateral plate. By a comparative study 
of two or more series of sections, it is easily demonstrated that 
these cells are a piece of the somite lying in front and have 
nothing to do with the Anlagen. To elucidate this point still 
further, I have drawn fig. 7 which represents a section through 
the intersomitic plane between the sixth (fig. 5) and the seventh 
somite (fig. 6). In this part the Anlage of the pronephros 
is developed still more weakly, and the mesoblast remains in a 
more primitive state. In the proximal edge of the lateral plate 
{x), no special structure is detected, but the edge fades away without 
a distinct limit. By comparison with Fig. 4, w^e can not find any 
marked diiïerence ; thus, here likewise, there is no cellular con- 
nection between the Anlagen in the two succeeding somites. 

From the fifth somite backwards for 9 or 10 somites, the 



318 s. HATTA : 

mesoblast presents almost the same feature of the Anlage as in 
the fifth somite mentioned above. Fig. 5 represents a section 
through the sixth somite, next behind the fifth ; when compared 
witli fig. 2 no marked difference is detected in regard to the 
structure of the mesoblast. But in some segments the development 
of the Anlage is somewhat weaker than in others, as seen in 
fig. 6, which shows a section throngh the seventh somite; while 
in a segment posterior to this somite, we find the Anlage as much 
pronounced as in the sixth somite. However, generally speaking, 
the Anlage of the pronephros in an anterior somite develops 
further than that in a posterior. It must be remembered that 
the somite in which the Anlage has already become expressed 
does not pass over suddenly into the somite in which no trace 
of it is to be seen ; but its development gradually grows less and 
less distinct from the anterior to the posterior part, until finally 
no trace of it is perceived. 

In the present stage, therefore, the Anlage of ilie pronephros 
is detected in more than 4 somites but is completely separated from 
the 7nyotome only in one segment, viz. the fourth somite^\ and it 
has no genetic connection either with the Anlage in the next 
following somite or icith the epiblast ; and it must he noticed that 
tve find the foremost Anlage not exactly beneath the fourth myotome, 
but always underneath its hind border. 

Figs. 8-17 represent sections throngh a still older em- 
bryo of this stage, having about 20 somites. The epiblast {ep.), 
the neural cord (n.), and the chorda (ch.) are essentially the same 
as before. Being cut through somew^hat obliquely, the myotomes 

l)Such a case is very rare. In most specimens examined, the Anlage separated from 
the myotome is fcinnd in many segments, so that we can hardly decide in whicli segment 
the separation takes place first. 



MORPHOLOGY OF CYCLOSTOMATA. 319 

on the two sides do not exactly correspond. On the right side 
of fig. 8, the hind border of the fonrth myotome (mt.IV) is 
cut through; the Anlage of the pronephros [a.jjn.l) presents in 
section an oval shape, consisting of columnar cells radially ar- 
ranged and containing a cavity of an irregular form. The 
histological structure of the Anlage is as compared with that 
in fig. 3, more or less loose^', and the Anlage itself is there- 
by also distended. The lateral plate (Im.) shows, however, no 
marked progress. The left side of this figure and the right 
of figs. 11 and 12 represent the section through the fifth myotome 
(mt.V) and the Anlage of the pronephros (a.pn.2) for that somite. 
The Anlage presents almost the same development as that just 
described. The left half of fig. 12 and the right half of fig. 
13 shows the sixth somite {mi.VI) and the Anlage belonging to 
it (a.pn.3). It can be inferred from the arrangement of its 
component cells that the Anlage has been just constricted off from 
the myotome, as is shown by the fact that the cells at the point 
marked with x of the visceral and parietal layers are not yet 
rearranged to form a continuous layer, — a condition which is 
observed not infrequently in younger embryos. Fig. 14 shows 
on the right side a section through the hind wall of the sixth 
myotome ; the Anlage beneath it [a.pn.3) is, therefore, the hind 
part of that represented on the right side of fig. 13 : it is 
entirely cut off from the myotome {mt.VI), and the two layers 
at this point have completely fused together, enclosing a com- 
paratively wide cavity. The same condition is observed in the 



1) When tiie pronephric Anlagen are cut ofl' from the myotome, their structure is at 
first loosened, that is, their component cells become loosely s-et together. Later the cells 
multiply tliemselves, and are again compressed by mutual pressure; giving a compact structure to 
the Anlage — })rohably the same condition observed by van Wyhe in Selachian embryos 
('89, p. 470). 



320 s. HATTA : 

section throngli tlie anterior border of the somite. This phase 
of constriction is doubtless earlier than that shown in fis;. 3. The 
right half of fig. IG and the left of fig. 14 is from the section through 
the mid-plane of the next following somite, i.e. the seventh. 
The Anlage of the pronephros {a.pn.4) is not yet cut off from the 
myotome {mi.VII), but the process is beginning as shown by a slight 
constriction and an inclination to arch out, while the suddenly 
weakened lateral plate {Im.) forms the distal (ventral) continuation 
of it. This feature of the mesoblast reminds us of the youngest 
stage of the Anlage described above (compare with figs. 2 and 5). 
In the section passing through the anterior or the posterior border 
of the somite too, the same condition of the Anlage, as on the 
rio'ht side of fio". 14, is observed. 

From the facts mentioned above, it is easily understood that 
the separation of an Anlage from a myotome begins with the 
constriction which takes place at the anterior and the postej'ior 
border of the somite, and that the middle portion is the last to 
be cut off, so that the cavity of the myotome {myocoelome) com- 
municates with the peritoneal cavity, during some time, through a 
narrow passage at the middle point. 

Myotomes when cut off from the Anlage of the pronephros 
assume a pentagonal form (see fig. 3) constructed of a dorsal, a 
lateral, a ventral, and two median sides, each of which is composed 
of a regular row of tall cylindrical cells. The dorsal and lateral 
rows of cells constitute the parietal layer of the myotome, while 
the three other sides represent the visceral layer (compare with 
the description on p. 314). 

For about ten segments behind the seventh somite, the 
Anlage of the pronephros shows the same condition as that seen 
on the right side of fig. 16. Fig. 17 represents a section 



MORPHOLOGY OF CYCLOSTOMAïA. 321 

through the twelfth somite ; we can find no marked difference 
between tlie Anlage in the seventh somite and in this. The 
segments Ivins; still farther backwards are not cut throuo;h exact- 
ly transversely in this same series of sections, owing to the cause 
stated above (pp. ol2 and 315), so that we can not trace the dif- 
ferentiation of the mesoblast from the anterior to the posterior 
part in this one series. But 1 could demonstrate from several 
other series of sections that the Anlage of the pronephros is, in 
the present stage, found in no less than 15 somites. 

Figs. 9-11 represent the contiguous sections through the 
intersomitic portion, on the right side, between the first and 
second Anlagen, i.e. between that of the fourth, and that of the 
fifth, somite. Fig. 9 is from the section next behind that shown in 
lig. 8 ; the portion [cd.) lying j^roximal to the lateral plate {Im.) 
presents no longer a weak appearance as in younger embryos 
(see the statement on p. 317 and figs. 4 and 7), but is occupied 
by a compact cellular structure (cd.) which suddenly passes over 
into the loosely composed lateral plate {Ivi.). Fig. 10 is from the 
section next posterior to fig. 9 and next anterior to the second 
Anlage represented in Jig. 11 and shows almost the same con- 
dition as in lig. 9, with respect to the structure in the proximal 
portion of the lateral plate. In other words, in the intersomitic 
portion between the first and second Anlagen, a cellular coj'd 
has become established, which connects these two Anlagen. It 
is this cord which gives rise to the collecting duct or Sammelrohr 
of RÜCKERT ('88), putting all the pronephric tubules in communi- 
cation. 

On the left side of figs. 9, 10, and 11, the contiguous sec- 
tions through the intersomitic portion l)etween the Anlagen second 
and third, are represented. In figs. 9 and 11, the condition of 



322 s. HATTA : 

the structure (cd.) at the proximal portion of the Literal plate is 
almost the same as that on the right side just described, although 
it is here somewhat weaker in development than there. The section 
represented in fig. 10 intervenes between the two mentioned above ; 
in this section, the structure in question {cd.) is weakest in develop- 
ment, consisting of four or five cells only. In the left half of fig. 15 
which represents the section through the intersomitic portion be- 
tween the sixth and seventh somites, there is found no structure 
to be compared with the cord mentioned above, the proximal 
edge of the lateral plate [x) being of the same condition as that 
in figs. 4 and 7. In fact, the cord appears after the complete 
separation of the Anlage from the myotome, and ivhen it is first 
established, the nearer the plane of a section to the Anlage either 
anterior or posterior, the thicker the cord. For instance, of the 
above three sections (tlie left side of Figs. 9-11), the middle 
(fig. 10) is the weakest. But this unequal development of the 
cord is soon made even by its growth as seen in the case of the 
cord between the Anlagen first and second (the right side of 
figs. 9 and 10). 

The history of this cord as given above shows that it has doubt- 
less the genetic relation with the Anlage of the pronephros. In 
early stages, no such structure is found in the intersomitic portion, 
but it becomes established one after the other with the develop- 
ment of the Anlagen. The cord is in section, thickest near 
the Anlage and weakest in the midway between two consecutive 
Anlagen, when it is first established. These facts give naturally 
an impression that it is growing out of the two consecutive An- 
lagen backwards and forwards and these two growing ends meet 
at some point in the midway between these two Anlagen, finally 
to fuse together. This ])oint of meeting is, I think, indicated by 



MORrHOLOGY OF CYCLOSTOMATA. 323 

the part where the duct has been described above as weakest. 
It is also the fact that repeated cell-multiplication takes place 
at the outer rim of each Anlage of the pronephros. One might 
suppose that the product of the cell-division would contribute 
only to the growth of the Anlage itself aud has nothing to do 
with the cord ; but this is not the case : the Anlage does not 
grow at the outer (lateral) end, as it might seem, but by cell- 
division within its own structure. I have never observed any 
case of cell-proliferation along the dorsal edge of the lateral 
plate in an intersomitic portion, although the cords appear, in 
later stages, to have some connection with, that edge, when they 
are fully establisiied (see the right side of figs. 9 and 10) ; this 
connection thus is not primary, but secondary. The epiblast has, 
from the first, no share in the formation of the cord, always 
showing a sharp contour against the mesoblast below. 

There is thus no difficuliy in acceptim/ the view that the 
connecting cord is formed of the intersomitic cell- outgrowths which 
are budded out of the anterior and posterior rims of each Anlage 
of the pronephros and are subsequently fused together. The cord 
isj therefore, originally brought about by the conßuencc of the free 
extremities of the Anlagen. 

Further development of the Anlage of the pronephros may 
be intelligible by refering to fig. 18 which represents a sec- 
tion through a little older embryo of Stage ii. The epiblast {ep.) 
consists of a single layer of cubical cells as before ; the neural cord 
(n.) is still solid. On the left side of the figure, the hind border 
of the fourth myotome {pit. IV) is cut, while on the right, the 
mid-plane of the fifth myotome is met witli. A comparison with 
the corresponding parts in the younger stages (figs, o and 8) will 



324 s. HATTA : 

plainly dernonstrate a progressive change undergone by the 
pronephric Anlage. The Anlage on the right side [a.pn.'Z) 
presents a feature much like that seen in fig. 3, notwithstanding 
some points of progress. The Anlage on the left side (a.pn.l), 
however, shows a considerable progress ; it has become much more 
compact by the active multiplication of its component cells. 
Owing to mutual pressure, the cells are compressed and their 
nuclei are regularly arranged, describing together an ellipsoidal 
figure. The inside of the ellipse encloses a comparatively large 
lumen, which is standing in connection with the body- cavity 
represented, at the present stage, only by the boundary line of 
the parietal and the visceral layer of the lateral plate (Im.). 

In a little more advanced embryo, the cross-sections of which 
are rej)resented in figs. 20-31, the neural cord {n.), the myotomes 
(mt.), and the Anlage of the pronephros show some progress as 
compared with those described in the preceding pages. The epiblast 
(ep.) consists, as in the embryo just described, of a single layer 
of cubical cells and is limited by a sharp line against the struc- 
tures l^elow. The component cells of the neural cord^^ become 
arrariged in two layers, leaving, in the anterior secti(jn of the cord, 
a vertical fissure-like lumen in the median line of the cord, which 
represents the beginning of the central canal (figs. 20-21, &c.). 
The posterior part of the cord is still solid, although the position of 
the central canal is marked by a vertical line produced by cell- 
boundaries (tig. 26) just as described in the foregoing pages. The 
myotomes are, in the anterior region, likewise enlarged, probably 
owing partly to multiplication of component cells and j^artly to 

]) Owing to liie same cause as the sections rejuesented in Figs. 2 and o, the vertical 
(liaiueter of tlie neural cord in Figs. 20-23 is shown somewhat longer than it is in reality. 



:\rORPHOLOGY OF CYCLOSTOMATA. oZO 

the loosening of the composition of the tissue'^ and assnme the 
lijhape of a scalene triangle (figs. 20-23) ; tlie median side of the 
triangle (mus.) represents the visceral, and the two otlier sides 
[ruf.) the parietal, layer of the myotome. In the posterior region, 
they are yet of a compact structure of a pentagonal form, en- 
closing a cavity (figs. 25-31, mt.VII-X). 

The anteriormost Anlage of the pronephros is found as before 
under the hind part of the fourth myotome, the section of which 
is represented in fig. 20 {apii.l). It shows a considerable 
development : the component cells, which are of high columnar 
character are no longer compressed, but the tissue is more or less 
loosened. Thus the Anlage itself is distended, and its upper 
(dorsal) angle becomes acute and grows in between the epiblast 
and the myotome. The internal cavity of the Anlage also be- 
comes conspicuous. The Anlage of the pronephros under the 
next posterior myotome (the fifth) is not so advanced as in 
the last somite (the fourth). In fig. 22 is shown the section 
through the hind part of this somite and of the pronephric An- 
lage belonging to it {mt.V and a.pn.2), a section through the 
mid-plane being unfortunately wanting in this series of sections. 
The next posterior Anlage is found just under the sixth myotome 
and represented in fig. 24 {a. pa. 3) together with the hind border 
of the myotome {mt.VI). The Anlage shows a compact structure 
wdiich is probably due to a rapid multiplication of the constitu- 
ent cells. The next following Anlage of the pronephros is found 
beneath the seventh myotome (fig. 26, a.j^n.é). It shows no 
further development than the separation of it from the myotome 
and the fusion at the retrenched ends of the two lavers of 
mesoblast : it is in the same stage of constriction as that in the 

l)See the foot note in p. 319. 



326 s. HATTA : 

right of fig. 13 wliicli represents the Anlage third in a younger 
embryo. 

The Anlagen above referred to are connected with each other 
by the solid connecting cords (figs. 21, 23, and 2;j, ed.), which 
are found between these Anlagen just as described in the younger 
stages. Of these connecting cords, that between the Anlagen 
second and third (fig. 23, cd.) is the thickest, while that be- 
tween the Anlagen third and fourth (fig. 25, cd.) is the weakest 
in development, owning probably to its having been just established. 
The cord between the Anlagen first and second (fig. 21, ed.), 
however, is rather weaker as compared with that in the next 
posterior intersomitic plane (fig. 23, ed.). Such a case is occa- 
sionally met with ; but this is doubtless not normal ; in most 
cases examined, the cord is thickest in anterior segment and de- 
creases in thickness gradually posteriorly as seen in the last 
example (see pp. 321-322 and figs. 9-11). 

The Anlage of the pronej)liros belonging to the eighth so- 
mite and that of tlie ninth somite are not completely cut olf 
from the myotome to which these respectively belong (figs. 28 
and 30). They are, however, constricted already at the anterior 
and posterior borders of the segment : fig. 27 shows the posterior 
part of the Anlage fifth, and fig. 29 rej^resents tlie anterior 
part of the Anlage sixth respectively. Such a case is observed in 
the younger stage described in the foregoing pages (pp. 319-320). 
Compare these two figures with the right half of Fig. 14 and the 
description on p. 319 : here the phase of constriction is a little 
more advanced than there. The anterior part of the Anlage fifth 
and the posterior part of the Anlage sixth, the figures of which 
are omitted, have features much like those seen in figs. 27 and 29. 
In these two segments, the central portion of the Anlage is 



MORPHOLOGY OF CYCLOSTOMATA. 327 

just in the process of being- constricted ofif from the myotome, and we 
can not decide by this case alone which segment (whether the an- 
terior or the posterior) is the further developed ; a comparative 
study of other examples shows that the separation in the posterior 
segment follows that in the anterior. The state of the mesoblast in 
the next posterior segment, i.e., the tenth segment (fig. 31), is quite 
different from that just described ; it is in a more primitive condition 
of development. The Anlage of the pronephros {a.scl.) presents only 
an indication of constriction, — a feature which we have observed 
repeatedly in embryos of younger stages (compare with figs. 2, 5, 6, 
14, and 16). From this segment backwards, a few segments show 
almost the same condition. Still further posteriorly, the structure 
of the mesoblast can not be readily observed, since the planes of 
sections incline by degrees in the cranio-caudal direction, owing, as 
above stated, to the bending of the longitudinal axis of the embryo. 

In all the segments mentioned above, the lateral plate {Im.) 
consists of a loose tissue of cells of variable shape, and the Anlage 
passes over suddenly into the lateral plate just as in the embryos 
described in the foregoing pages. 

Ill this stage, therefore, the Anlage of the pronei^hros is com- 
pletely separated from the myotome in 4 somites, i.e., from, the 
fourth to the seventh inclusive ; and these are connected icith one 
another by the intersomitic solid cord. In the following 4 or 5 
somites, the constriction is just going on, while in a few of still 
more posterior somites it is indicated merely by a slight dejyressioii 
in the parietal layer of the mesoblaH. 

Period 2. 

Tn the embryos which belong to Stage in, we observe a 
decided advance in several respects. Figs. 32-50 represent a series 



328 s. HATTA : 

of cross-sections through one of these eniliryos whicli lias abont 25 
somites. The neural cord {n.) is reduced in size and in the anterior 
part has a conspicuous canal. The myotomes which showed 
before a pentagonal outline, in the anterior part of the body have 
now assumed an elongated lozenge-shape^^ and is composed of an 
outer, and an inner, layer of long cells whicli have begun to dif- 
fentiate themselves. The inner layer (mus.) represents the median 
and ventral rows of the pentagonal myotome mentioned on p. 320 
and, therefore, corresponds to the visceral layer of the myotome ; 
the outer layer {cuL) is the product of the dorsal and lateral layers 
and constitutes the parietal layer. The pronephric Anlage is com- 
posed of high columnar cells which are plainly distinguishable 
from the much shorter elements of the lateral plate (Im.). The 
component cells of the Anlage of the pronephros which we generally 
found to be compressed in the foregoing stage (pp. 324 and 32Ö), 
are now more or less loose, and the internal cavity of the Anlage 
is somewhat widened, being distended by the loosening of the cells.'' 

The peritoneal cavity is, at the present stage, still repre- 
sented merely by the boundary-line of the parietal and visceral 
layers of the lateral plate. 

In the present stage, the foremost Anlage of the pronephros 
is, as before, found under the hind part of the fourth myotome 
(figs. 32-35, a.pn.l). The Anlage shows, in section, a circular 
outline and is composed of high columnar cells arranged in a 
radial manner. The internal cavity of it is confined no longer 
to one section, but it is observed in three or more sections ; it 
is most spacious in the hind part of the fourth somite (fig. 33) 
or in that part where the cavity is visible from an early period. 

])Afew myotomes in the anterior somite'^ tend to assume this shape alrcaily in the last 
stage (see figs. 20, 21, and 22). 

2) See the foot-note on p. ?>ld. 



MOKPHOLOGY OF CYCLOSTOMATA. 329 

From this part backwards it gradually decreases in width 
until no space is perceptible. Anteriorly the cavity is also some- 
what narrowed, Init not as much as in its posterior contin- 
uation, aud ends blindly rather suddenly at its anterior end 
(hg. 32). The anterior portion of the Anlage forms a blunt conical 
tube (Fig. 32, a.pn.l) projecting anteriorly and lying between the 
dorsal edge of the lateral plate {Im.) and the lower surface of 
the fourth myotome {mt.IV). The existence of this conical 
tube^' gives us a strong impression that originally there must have 
been present an Anlage of the pronephros in the anterior segment 
which was connected by a connecting cord with the Anlage 
belonging to the fourth somite, but had disappeared during the 
phytogeny and that this conical tube is the remnant of this 
connecting cord.-' 

The next posterior Anlage, which is found under the liftli 
myotome (%s. 37-30, a.pn.2) and shows an outline much re- 
sembling that represented in tigs. 32-35, has an internal cavity 
of irregularly triangular form, extending through three sec- 
tions, of which the foremost section contains the most spacious 
cavit}^, while in tbe others the lumen grows smaller and smaller. 
The pronephric Anlage in the next following somite (tigs. 41-43, 
a.pn.3) has an outline much like that shown on the left side of 
Figs. 18 and 24, being in the same phase of development, that is, 
it is of the form of an isosceles triangle whose two basal angles 
touch the myotomes. This Anlage is found under the sixth 

1) Tlie internal cavity of this conical tube is not entirely closed, but there is clearly seen 
a small canal {%) directed towards tlie median side and opening below the myotome. I can 
not decide, at present, -whether this canal is normal or abnormal ; for I cau not make out the 
corresponding structure on the opposite side and have no other embryo of exactly the same 
stage, in which the structure in question would probably be found, if it be of some definite 
meaning ; I also can not detect any trace of such a canal in eml)ryo.s of advanced or younger 
stages. 

2) See the descrijition under I'eriod. 1. 



330 s. HATTA : 

myotome and contains the internal cavity extending likewise 
for three sections, of Avliich however the hindmost contains the 
widest cavity, while it is diminished in width anteriorly : in other 
words, the width of the cavity enlarges in inverse direction 
as compared with that in the preceding two somites. The 
Anlage fourth under the seventh myotome (figs. 45 and 46, a.pn. 
4) has an oval outline like that shown in fig. 26 and encloses 
an internal cavity, which covers two sections and is anteriorly 
wide and posteriorly narrow. The two following Anlagen which 
are detected under the eighth and ninth myotomes respectively 
(tig. 48, a.pn.ô and tig. 50, a.pn. 6) show almost the same con- 
dition of development as in the somite just described ; the 
internal cavity which they contain is likewise extended into two 
sections ; the width of the cavity is about the same in these two 
sections, being of a fissure- like form. 

The solid cord which is observed in the embryos of the last 
stage connecting the Anlagen with one another, is also found 
here. The cord in the intersomitic plane between the Anlagen 
first and second (fig. 36, ed.), that between the Anlagen second 
and third (fig. 40, ed.), and that between the Anlagen third and 
fourth (fig. 44, ed.) are all comparatively short, so that they 
are in each stretch confined to only one section, while that in 
the two posterior intersomitic planes, i.e. between the Anlagen 
fourth and fifth, and between the Anlagen fifth and sixth, the 
cord is extended in each case into four sections. In this latter 
part, the cord is in a primitive condition ; the component cells 
are actively multiplying. Hence these four sections all show 
similar features. I have endeavoured to show in fig. 47 one of 
these sections which is taken from one of the four sections between 



MOEPIIOLOGY OF CYCLOSTOMATA. 331 

the Anlagen fourth and fifth, and in fig. 49, one hetween the 
Anlagen fifth and sixth. 

This inequality in the length of the intersomitic solid cord 
is, I believe, due to differences in the degree to which the canali- 
zation within the Anlage has extended into the connecting cord. 
In the anterior section of the pronephros, this process has already 
proceeded to some extent into the interior of this cord, while in 
the posterior, the cavity is still confined entirely within the 
Anlage itself. The whole system of the pronephros at the present 
condition may be compared to a bamboo-cane with nodes and 
internodes ; in the anterior section of the system, the nodal septum 
has Ijecome very thin, while it has a considerable thickness 
in the the posterior. As will be shown further on, all these 
septa entirely disappear later when the collecting duct is fully 
established. 

From the fact mentioned above, it Avill be easily seen that the 
})rocess of canalization in the pronephric system of Peiroviyzoii 
begins in the internal cavity of the pronephric Anlage in each 
segment and is extended into the intersomitic connecting cord. 
The direction in which this process proceeds seems, generally 
speaking, to be from the anterior section to the posterior ; for 
in most cases, not only the internal cavity in each Anlage is 
spacious anteriorly and narrowed posteriorly, but the cavity in 
anterior somites is extended more, or canalization goes on further, 
than in the posterior section of the system ; although the pro- 
gress in the opposite direction is occasionally met with. 

From the tenth somite backwards, five or six segments show 
the same condition of the mesoblast as in the eighth and ninth 
somites, after which the series can not be studied, owing to the 
inclination of the planes of sections, referred to above. 



332 s. HATTA : 

In ail tlie segments above referred to, tlie lateral plate of the 
niesoblast shows the same condition as in the foregoing stages, but 
has become more distinct from the Anlage of the pronephros. 

In the present staye of development, then, the Anlage of the 
pronephros, is cat oß' from the myotome in more than 10 segments, 
and the canalization has advanced in the anterior section of the 
system, to a state just ready to put the Anlayen in the succeeding 
somites in communication ivith one another, although the inter- 
somitic connecting dtict in the posterior part remavns still solid. 

Figs. 51-Ô8 were drawn fron] a series of sections through 
one of the older embryos in this stage. The intei'nal structures are 
developed much more than in the embryo just described. The 
cells forming the visceral layer of the myotome have been differen- 
tiated into the muscle-plates, while the parietal layer is composed 
of cubical cells. The Anlagen of the pronephros have acquired, 
in most cases, a tubular structure and have grown dorsally, being- 
folded out from the body-cavity ; I will accordingly call them the 
pronephric tubules. 

On the right side of iig. 51, the foremost tubule ipt.l) is 
visible, which no longer contains the internal cavity but is 
converted into a solid mass of cells occupying the space beneath 
the fourth myotome. This consolidation is not due to retrogres- 
sive changes, but is effected by very active cell-multiplication 
which takes place within the tissue. The cross-section of the 
collecting duct seen on the left side of tig. 02 [cd.) which re- 
presents the third section behind the last, is likewise solid. The 
tubule on the right side of this figure {])t.2) and that on the 
left side of tlie third section ])Osterior to it (Iig. 03, pl.'^) are 
respectively the second tubule of the right and left side found 



jrOEPHOLOGY OF CYCLOSTOMATA. oôo 

under the fiftli myotome ; both are of a triangular form nnd 
contain a very spacious internal cavity of tlie same shape. On 
the right side of fig. 53, the sixth myotome and the third tubule 
are shown. The section next posterior to fig. Ô3 (fig. ô4) shows 
the cross-section of the collecting duct (cd.) on the right side and 
a slice of the hind wall of the second tubule on the left (^9^.2). 
The cells composing the duct are closely set together, although 
arranged more or less radially, acquiring a tubular form. As 
has been repeatedly mentioned above, the epiblast is, as in 
the foregoino' staoe, marked off from the mesoblast as well as from 
the Anlage ; but at the ^^rcsent stage, the second tubule (figs. 54, 
pt.2) pushes against the epiblast, probably in consequence of an 
enormous multiplication of its component cells, so as to cause 
the latter to be a little elevated externally. It must be remarked 
here that the Anlagen, especially the first and the second, when 
they first assume the tubular form, are brought into an intimate 
relation with the epiblast, striking against it. In some of 
my sections, a mitotic figure is seen at that point of the epiblast^' 
(fig. 54, :c). This might lead some to assume a genetic connection 
between the epiblast and the pronephros in Petronyzon ; but there 
is, I believe, in reality no such relation. If the epiblast gives some 
cells destined to build the pronephros or a part of it, cell-prolifera- 
tion or some other mode of cell-production would necessarily be 
observed in the epiblast in the preceding stages or at least, in 
the stage here spoken of. In the foregoing stages, the epiblast 
had, as has been repeatedly mentioned above, a sharp limit against 
the structures inside it. At the present stages also, it is marked off 
by the boundary-line of cells from the tissue .of the tubule, 



l)Tn the series of sections, from which fig. 54 is drawn, I ohserve mitotic fignres at 
tliat point in several sections. 



334 s. HATTA : 

sliowiDg no structural alteration. Mitotic figures are met witli 
not infrequently in that part of the epiblast (fig. 54, a-) ; their 
axis lies, however, in all the cases examined parallel to the 
plane of the epiblast, giving us an impression of the resulting 
cells contributing to the formation of no other part than the epi- 
blast itself; on the contrary, within the structure of the tubule 
the cells are rapidly multiplying (figs, ol, pt.l and fig. »54, 
j)t.2), showing that the growth of the tubule is actively going 
on. In fact, the connection, or rather the intimate contact, of 
the pronephric tubule with the epiblast is a temporary con- 
dition ; the separation follows immediately afterwards, and the 
tubule returns soon into a state similar to that seen in fig. 52 
{pL2). 

According to Rickert ('88), a similar case is observed 
in Selachian embryos : the tubules become connected secondarily 
with the epiblast — what caused him to believe that the latter might 
give some constituent elements to the tubules. 

The third section behind that represented in fig, 54 (fig. bb) 
shows, on the right side, the fourth iptA) and, on the left, the third 
tubule (p^.5) respectively. The latter is not so far developed 
as its counterpart on the opposite side (fig, 53, pt.3), while the 
former presents a great progress: it consists of a definite epithel- 
ium and contains a distinct cavity of triangular shape, although 
the corresponding tubule on the opposite side (fig, 6ß, pt.4) 
which is found in the third section behind the last, is much less 
advanced in development. The fifth tubule, the tubule on the 
right side of fig. 56 (pt.ô), is somewhat more developed than that 
which belongs to the anterior somite (the fourth tubule on the 
opposite side) ; but it has a feature much resembling the fourth 
tubule on the same side (fig. 55, pL4) and the second on the 



MORPHOLOGY OF CYCLOSTOMAÏA, 335 

opposite side (fig. 53, pt.2). In short, in this series of sections, 
tlie tubnles on the right side, are all more advanced than those 
on the opposite side. Tlie sixth is very primitive in development ; 
fig. 57 represents the section, on the left side, through the anterior 
part of the ninth somite and, on the right, the posterior part of 
it. The left tubule is sliced at its anterior wall, but the right 
tubule is cut through in its mid-plane. Tt is composed of two 
layers of columnar cells, but no cavity has yet appeared in 
the interior. 

From the tenth somite backwards, the Anlagen are cut off 
from both the myotomes and the lateral plate, and constitute the 
segmental duct or the posterior continuation of the collecting 
duct, which is distinctly traceable for 7-8 somites. Not infre- 
quently, however, a somite is met with, in wdiich the segmental 
duct is not yet cut off from the lateral plate at the ti.iie wdien 
the separation is finished in a majority of somites, as seen in 
fig. 58 which represents a section through the twelfth somite. 
The left half of the figure show^s the duct entirely cut off from 
the lateral plate, while the right exhibits the state not yet 
separated. The same structure is made out in two contiguous 
sections, so that one might mistake it for a pronephric tubule. 
This point will be described further on. 

The relation of the pronephric tubule and the peritoneal 
cavity is not so simple as in the last specimen ; besides the 
pronephic tubule, there is seen another structure which projects 
out of the inner angle of the peritoneal cavity (figs. 52, 53, 6by 
and bQ, c.p.). This projection is originally a fold of the peritoneal 
\vall and gives rise, as subsequent history shows, to the radix of the 
mesentery, whence the gonads and the mesonephric tubules are 
derived. It will here be called briefly the " coehmic projection.^^ 



336 s. HATTA : 

At that point of the visceral layer of the mesoblast, where the 
Anlage of the pronephric tubule passes over to the lateral plate, 
it is alwa3^s many cells deep (figs. 55 and 50, c.p.), and the pro- 
jection in question is brought aljout by repeated division of these 
cells. The projection formed is consequently seen in each somite 
and thus shows a segmentai arrangement. Its component cells 
are soon re-arranged into an epithelium, and the pouches thus 
formed push their way between the myotome and the hypoblast. 
The coelomic projection appears, at first sight, to be homo- 
logous with the coelomic pocket described by Price ('97) in 
Bdellostoma. The coelomic pocket is, however, according to Price, 
the product of both the parietal and visceral layers of the lateral 
plate and is afterwards converted into the cavity between the 
glomerulus and Bow3ian's capsule of the Malpighian cor- 
puscle ; the fioor of the pocket forms Bowman's caj)sule, 
and its roof together with a part of the pronephric tubule is 
transformed into the cover of the glomerulus ('97, p. 213). The 
coelomic projection in Petromyzon is, on the contrary, formed 
out of the visceral layer of the distal half of the somite and gives 
rise, as just stated, to the radix of the mesentery, from which 
partly the mesonephrie tubule and partly the gonads are formed.^' 

Fiîïs. 59-63 are from a series of sections throuo-h an older 
embryo of the same stage. In this series of sections, a further 
development of the coelomic projections is clearly seen ; the first 
figure (fig. 59) shows the section through the second tubule, fig. 
60 through the third, and so forth. In the first 3 figures and 
on the right of fig. 62, the coelomic projection {cp.) presents an 

1)1 will not here fiirtlier discuss this structure, as I iutend to do so in a future paper 
iu whicli tlie developuieut of the iiiesonephros in Pdromij-ym. will l)e delt with. 



MORPHOLOGY OF CYCLOSTOMATA. dû/ 

epithelial structure, forming the continuation of the peritoneum 
and folding out from the peritoneal cavity. Beneath the first 
tubule, there is found no rudiment of the projection ; under the 
second (fig. 59) it is very weak, while beneath the third (fig. 60), 
fourth (fig. 61), and fifth (fig. 62), tubule, respectively it is most 
vigorously developed. But on the left side of figs. 61 and 62 it 
is again in a primitive condition, just as in the last series of 
sections (figs. 52, 53, 55, and 56). 

The coelomic projections are not confined to the anterior region 
where the pronephric tubules are found, but it is found likewise 
in the posterior part where only the segmental duct develops. 
Fig, 6o show^s the section through the thirteentli somite ; on this 
section, the duct is cut off from the m3"otome and a well developed 
coelomic projection {cp.) is observed ; I will return once more 
to this subject further on. 

Leaving the coelomic projection in this stage of development, 
I will return to the origin of the Anlage of the pronephros and 
give somewhat more exact details on the subject. Since the piece 
of the mesoblast called above the Anlage of the pronephros forms 
for a time the |)roximal portion of the lateral plate, one might 
presume that its whole mass wall be transformed into the pro- 
nephric tulnile and will not partake in the formation of the perito- 
neal membrane. I was at first of this opinion, l)ut a careful 
observation of sections through the embryos in each stage showed 
my error. 

To illustrate this point satisfactorily, I have given, in the 
annexed wood-cut (Wood-cut 1), a series of semi-diagramatic 
figures, which show the successive phases of the changes going 
on in the structure. A shows the first indication of the Anlaue 
of the pronephros before the separation of it from the myotome ; 



338 



s. HATTA 



a-b indicates the extent of the 
coelomic projection. When tlie 




Wnud-L'ut 1. — Seiuidiagniniatic tigiues to 
illustrate the siK;ces.sive jjliases of the 

evolution of the nephrotome. 
I'roiu the right side of fig. 16. 
from fig. •'>. 

from the right side fig. 18. 
from the left side of the same, 
from the left side of fig. 53. 
from the right side of fig. 55. 
from the left side of fig. 61. 



called the coelomic projection 
plate of epithelial cells {G, c.p. 



Anlage ; c-d shows that of the 
myotome is cut oft', the point 
of the parietal layer indicated 
by a becomes fused with the 
point r of the visceral layer 
(B, ac). This piece of the meso- 
blast assumes an ellipsoidal sha2)e 
(C). The component cells of 
this ellipsoid are multiplied by 
active cell-divisions, and the 
piece almost loses its lumen and 
gets a compact consistence {D). 
Meanwhile the cells in the space 
d-c remain inactive. Conse- 
quently the piece acquires a tri- 
angular form (E), whose upper 
sharp angle, together with the two 
sides enclosing this angle, gives 
rise to the pronephric tubule. 
The lower (median) obtuse angle 
now begins to grow by cell- 
mulplication {F) and produces 
a small knob {F, c.p.), which 
grows further and pushes in be- 
tween the myotome and the 
hypoblast (the upper wall of the 
enteric canal). This cellular 
projection is that which has been 
It is reduced into a thin 
) and assumes then the form of 



MORPHOLOGY OF CYCLOSTOMATA. 339 

;i true fold of the visceral layer of the lateral plate. At the 
same time, the upper angle or the pronephric angle develops 
further and assumes a tubular form composed of a single layer 
of columnar cells {G). 

The peritoneal cavity begins, therefore, at the point, from 
which the coelomic projection starts, and the part of the layer 
dorsal to this point is all appropriated to the formation of the 
pronephric tubule. The nephrostome will be found, therefore, by 
the point where the tubule passes over to the projection. 

I will add a few words on the diftereutiation of the myo- 
tome, so far as concerns the topographical relation of it to the 
Anlage of the pronephros. The myotome consists, at the present 
stage (Stage ui), of the inner and outer layers which constitute 
respectively the Mu.^helblatt and the Cutuhlatt of German 
authors (lig. 59 and 60, nms. and cut.). The cells composing the 
Mu^helblait {mua^ are, simply differentiated into a transverse row 
of the muscle-plates. The outer layer [cut^ undergoes, however, 
subsequently a series of interesting changes : it folds in, just as 
the Sklcrablatl or sclerotome described by Hatciiek in Amphioxua 
('88) between the Blushclblatt and the chorda and the neural tube.'' 
As is well known, Rabl ('88) has homologised Hatcher's Sklerab- 
latl with his Sdcroionidivcrtikcl of Selachian embryos, which is 
the evagination of the ventral part of the visceral layer of the meso- 
blastic somite. This part of the somite (the sclerotome) corres- 
ponds, I believe, exactly to the ventral row of the pentagonal 
myotome in my embryo (see pp. 314 and 320), which comes after- 
wards to form the ventral part of the cutis-layer (see figs. 21, 22, 
23, 36, 37, 43, 49, ö9, 60, &c.). When the myotome is not yet 
separated from the rest of the mesoblast (fig. 2), this part of the 

1) This biihject will he trealeJ fif in an independent article. 



340 s. HATTA : 

selerotomic layer forms, as has been seen above, a direct continua- 
tion of the visceral layer giving rise to the coelomic projection. 
(The successive changes of the myotome are seen in figs. 1, 2, 3, 
18, 21, 22, 43, 49, 60, &c.) 

From the above account, it can be inferred that the ventral half 
of the mesoblastic somite in Petroinyzon, which gives rise to the 
pronephic Anlage and the coelomic projection, is doubtless homo- 
logous with the " inicrmcdiaie cell-mass " of Balfouk described by 
liim in Selachia and, therefore exactly coincides with the ** Nephro- 
tom " of Ki'CKERT.^* Ho far as concerns its future destination, how- 
ever, the results arrived at by me slightly deviate from their views. 

A series of sections through the oldest embryo of this stage 
is represented in figs. 64-76, The epiblast has undergone no 
histological change, but remains, as before, one cell deep. Many 
structures, however, exhibit a remarkable progress. The muscle- 
layer {iiiua.) of the myotome is, for instance, further difierentiated, 
now consisting of a transverse row of long nniscle-cells, although 
the cutis-layer {rut.) is still composed of short cubical cells. In the 
anterior region, the true coelome {pp-c), l)ecomes conspicuous en- 
closed l)y the parietal {m./>.) and the visceral {i/).r.) layers of the 
lateral [)late, both of which consist of a single row of cubical cells. 
The ventral edges of the lateral plates on both sides do not, however, 
yet meet in the ventral median line. The walls of the enteric 
canal too are, in the anterior region, reduced into a single cell layer. 

A great alteration is met with in the prouephric tubules. 
They have assumed a cylindrical form composed of tall columnar 
epithelium and have grown dorsally, pushing in between the 
myotome and the epiblast, causing the latter to be elevated a 

1) In si)ite of the discussion liy Kl'CKEkt ('89, jip. l'J-20) on the inexactness of the expres- 
sion "inlerniediule cell-muss," I honiologise, witli many aulhois, ihese l\\o tcnns with each othei'. 



MORPHOLOGY OF CYCLOSTOMATA. 341 

little. The internal lumen of the tubules are ]iut not only in 
wide communication with the peritoneal cavity, but also in direct 
continuation with one another through the collecting duct, which 
consists of a regular columnar epithelium-cells arranged radially 
and now encloses a conspicuous lumen. 

In the foremost of these twelve sections (fig. 64), we notice that 
a structure {pt.l) consisting of a few cells projects at the outer 
corner of the proximal edge of the lateral plates and lies in contact 
with the outer wall of the myotome on either side. This structure is 
found under the anterior border of the fifth myotome and T infer 
that it is a remnant of the first pair of the pronephric tubules which 
begins to decline in the present stage. The reason wh}- it is found 
not under the fourth myotome as in all the stages hitherto des- 
cribed but beneath the anterior border of the fifth myotome, 
consists probably in its shifting backwards ; for we find, in this 
series of sections, another pair of the tubules under this same fifth 
myotome. A comparison of this figure with fig. Q>ö representing 
the next posterior section will make the matter clear. On the 
left side of fig. 65, the same remnant structure {pLl) together with 
the collecting duct [ed.), which connects the first and the second 
tubules can be observed, while the cross-section of the collecting- 
duct in the corresponding intersomitic plane is seen on the op- 
posite side {ed.). The next following section is shown in fig. Gß ; 
the tubules (pt.2) on both sides communicate freely with the 
peritoneal cavity ; these are found beneath the hind part of the 
fifth myotome and are the second pair of the tubules ; the open- 
ings (list. 2) to the peritoneal cavity ^are, therefore, the second 
nephrostome of the pronephros. The tubule on the left side is 
weaker than that on the right, since a larger part of the left tubule 
is visible on the section next posterior which is represented in fig. 



342 s. H ATTA : 

67 (p/.2). The shape, which tlie tnhules of the second pair (fig. 06, 
pt.2) assume at about this stage, is a characteristic triangle, 
whose two angles, the one directed dorsally and the other directed 
medially, are acute and whose outer (lateral) angle is obtuse 
(see the left side of fig. 59 and the right side of fig. (^ß,pl;.2) ; so 
much so that we can easily determine by this feature the fact 
of their being the second pair. This peculiar shape of the second 
tubule is retained for a considerable time as will be seen further 
on. On the right side of fig. 67, the collecting duct (cd.) is cut 
through transversely ; on the left, the same duct {rd.) and the 
hind part of the second tubule {pt. 2) are seen. 

At the point where the nephrostome oj^ens to the peritoneal 
cavity, the visceral peritoneum at the median corner of the latter 
projects out between the myotome and the hypoblast (figs. 66 and 
67, c.p.) ; beneath the collecting duct, however, no such structure 
is detected (see the right side of figs. 6f5 and 67). 8uch a pouch 
is repeated in each nephrotome (see figs. 66-75, c.p.) and is what 
has been called above the coelomic projection.^' 

The next following sections shown in figs. 68 and 69 show 
the third pair of the tubules {pt.3) to be of the same structure. 
In these two sections the tubules are cut through lengthwise, and 
the nephrostomes {nst.3) on the two sides come into view in sym- 
metrical manner. The tubules are so simple as to need no further 
explanation. Fig. 70 represents a section through the intersomitic 
plane between the sixth and the seventh somites, and next posterior 
to fio;. 69. It shows on either side the cross-section of only the 
collecting duct (ed.), consisting of radially arranged cells. Fig. 
71 is from a section through the seventh somite and is the third 



1) In this series of sections^ we often see the coelomic projection on sections passing 
tlirongh the intersomitic plane; Init this is the piece of it belonging to eitlier the anteriorer 
the posterior somite. 



MORPHOLOGY OF CYCLOSTOMATA. 343 

behind the section sliown in fig. 70 ; the tubules of the fourth 
pair {pt.4) show themselves symmetrically on both sides; they 
are somewhat less developed as compared with those of the last pair. 
Fig. 72 is the section next behind fig. 71 ; it shows on both sides 
tlie collecting duct {cd.) together with the coelomic projection {c.j}.) 
which is a part of that of the anterior segment. Fig. 73 is the third 
section posterior to that just described ; the blastoderm becomes 
more flattened than in the foregoing sections ; it shows on both sides 
the tubules of the fifth pair (pt.ô) ; the condition of the tubules 
and nephrostomes {nd.ô) is much like that in fig. 71. Fig. 74 
is from the fourth section posterior to fig. 73 ; the right tubule 
{pt-6) of the sixth pair and its nephrostome {nst.6) are visible on 
the right side, while the collecting duct {cd.) is cut through on the 
left. The sixth tubule and nephrostome on the opposite side are 
observed in the next anterior section which is not figured. The 
segments back of the ninth somite have no trace of the tubule, but 
the cross-sections of the posterior continuation of the collecting- 
duct, the segmental duct, are repeated in each section. Fig. 75 
represents a section through the sixteenth somite ; the cross- 
section of the segmental duct {sd.) on either side is seen ; it 
always occupies the space where, in the anterior region, the 
tubules or the collecting duct is found. 

This condition, however, is not continued to the dorsal lip 
of the blastopore. As I have stated in my previous paper ('91), 
many processes of development are much delayed in the hind region, 
so that we are here reminded of wdiat were seen in the anterior 
region of the younger stages. Fig. 76 represents the fifteenth sec- 
tion from the dorsal lip of the blastopore and passes through about 
the twenty-third somite. In this comparatively late stage, in 
which many mesoblastic organs have developed in the anterior 



344 s. HATTA : 

region, the neural cord {71.) is still solid ; the mesoblast {vhs.) is 
many-cell-layered and its metameric segmentation is still going on. 
On the right side of the figure, the section passes through the 
mid-plane of the myotome showing no sign of its separation 
from the rest of the mesoblast, while on the left, which shows the 
intersomitic portion, the process of separation {/nt. and Im.) is going 
on. On both sides, however, there is no structure that can be 
recognised as the Anlage of the segmental duct. 

T/ir six pairs of prone-phiHc tubules observed in this stage 
are the maximum number for Petromyzon ; this stage ought, there- 
fore, to be regarded ecs the highest p)oint of development toith reference 
to the pronephros. Even in the present stage, the foremost tubules 
show a tendency to degenerate. 

Period 3. 

The embryos of Stage iv, which have about 35 mesoblastic 
somites, present a remarkable progress. The head-fold is much 
prohjuged ; in older embryos of this stage, it begins to twist 
('1)7, fig. 1, D). Figs. 77-91 i-epresent sections through one of 
these embryos. In some myotomes, the sclerotomic fold goes 
deeper between the muscle-layer and the chorda. The parietal 
layer of the lateral plate i;m-p.) is much lessened in thickness, 
so that it is reduced, in the dorsal region (the posterior two thirds 
of the pronephric extent), into a thin epithelial lining of the 
body wall (see figs. 79-85). The coelomic projection is likewise 
reduced into a thin plate (figs. 82-86, c.j)) except in the anterior 
two segments of the pronephic region, in wdiich it still keeps the 
characters of the younger stages (figs. 77-81, cj).), only folding in 
deeper than in the foregoing stages. The visceral layer {m.v.) of the 



MORPHOLOGY OF CYCLOSTOMATA. 345 

peritoneum still consists of a cubical or rather cylindrical epithe- 
lium. The pronephric tubules are, in general, much prolonged 
and begin to coil in the dorso-lateral direction, so as to cause an 
elevation in the epiblast. The walls of the tubules consist of a 
regular row of cylindrical cells, which passes over suddenly into 
the thin peritoneum (figs. 77-86), except in the region of the second 
pair of the tubules, where the parietal layer {iti.p.) of the lateral 
plate still retains the character of the younger stages, being 
composed of cylindrical epithelium like the tubules themselves 
(figs. 77-79). At some regions, even a few mesenchyma-cells {inch.) 
appear, — for instance, beneath the chorda (see figs. 80, 82, and 84), 
in the median ventral space (see figs. 81 and 82), and also inside 
the lateral epiblast (see figs. 77 and 80). 

Fig. 77 shows a section through the fifth somite and 
therefore corresponds to fig. Qß, which represents the section 
through the same plane of an embryo at a younger stage. The 
longitudinal section of the second tubule {pt.2), together with the 
corresponding nephrostome {nd.2), is seen on the left side of the 
figure, greatly resembling the tubules of the same pair in the 
younger stage (compare with fig. Q)Ç>). On the right side, the 
nephrostome {nst.2) alone is observed ; the tubule proper is to 
be seen in the two following sections which are represented in 
figs. 78 and 79. 

Beneath the myotome anterior to the one just described, 
there is found neither a tubule nor any structure that may be 
regarded as the remnant of it. In the space between the epiblast, 
the myotome and the lateral plate, however, a few scattered 
cells (fig. 77, vich.) are found. I at first supposed that these 
might be disconnected component cells of the first pair of 
tubules ; ])ut, as free cells of quite the same character are found 



346 s. HATTA : 

in other places, for instance, in the space between the lateral 
plate and the epiblast (figs. 79-81, mch.), I have been compelled 
to conclude that they have no genetic relation with the pronephros, 
but are niesenchymatous cells which are destined to form the 
blood-vessels and corpuscles. 

As the embryo was somewhat twisted, the sections did not 
pass through the lateral walls of the body in an exactly trans- 
verse plane, but unavoidably obliquely, on either side, as the 
continuous serial sections represented in figs. 70-86 show. 

While on the left side of fig. 78 the posterior portion of the 
second tubule {pt.2) is seen, the second nephrostome {nst.2) is ob- 
served on the right together with a cross-section of a tubular struct- 
ure (ed.). This latter might be taken as a slice of the anterior 
border of the second tubule, but is, in my opinion, the remnant of 
the collecting duct which once connected the second tubule with tlie 
first and forms, at present, a tubercle in front of the second tubule, 
the first tubule having disappeared ; for the second tubule on that 
side is observable in the next following section represented in 
fig. 79 {pt.2) showing its characteristic features stated above 
(p. 342). 

On the left side of fig. 79 and on the right side of fig. 81, 
we see the anterior half of the third tubule {pi.S) and nephro- 
stome {nsL3) of each side, their respective posterior half being 
found on the left side of fio-. 80 and on the risjht side of fis;. 
82 {pt.3 and 7isL3) ; the tubules are bent laterally and dorsally, 
probably caused by the prolongation of their tubular portion ; for 
their nephrostomal part and dorsal blind end retain their original 
position. This is the first step in the convolution of the pro- 
nephric tubule. 

As seen on the left side of figs. 79 and 80, the tubule {pt.3) 



MORniOLOGY OF CYCLOSTOMATA. 347 

of the third pair shows a new character : the dorsal blind end 
and the nephrostomal portion of the tubule are more or less expand- 
ed, while these two portions are united by a slender middle trunk. 
When compared with the tubule of the same pair on the opposite 
side represented in figs. 81 and 82 [pLS), this character of the 
tubule will be understood more clearly : the dorsal expansion is seen 
in fig. 81, while the nephrostomal widening is observed in fig. 82. 
The tubules of the following two pairs show the same feature. 
On the right side of fig. 80, only the collecting duct between the 
second and third tubules is found. The left tubule of the fourth 
pair is shown on the left side of figs. 81 and 82 (pi. 4) ; fig. 83 
shows a cross-section through the collecting duct (cd.) between the 
third and fouith tubules and a slice of the anterior wall of the 
right tubule {pt.4) of the fourth pair which is prolonged and bent 
like the tubules of the last pair. The fifth pair of tubules is seen 
on the left half of fig. 84 on one side (j^^.ô) and on the right 
of fig. 85 on the other {pt.o). It is not developed as much as 
the more anterior pairs, but shows considerable progress as com- 
pared with the tubules in fig. 73 which represents the younger 
stage of the same pair. 

These four pairs of the tubules (from the second to the fifth) 
contain a spacious lumen and stand in wide communication 
with the peritoneal cavity, which becomes, at the present stage, 
conspicuous from this region forwards. 

In fio;. 86 which shows the fifth section behind the section 
shown in fig. 85, the space on the left side which is occupied, 
in the more anterior region, by the tubule or the collecting duct, 
is replaced by the cross-section of a duct [scl.) with an oval out- 
line and an ovoid lumen. This is the segmental duct under the 
tenth myotome {mt.X). On the right side ol" the figure, however, 



348 s. HAïTA : 

besicles tlie cross-section of a duct (ed.), there is seen a pronephric 
tubule {pt.6), tlie long axis of which is directed vertically to 
the inner surface of the epiblast. It is found just beneath the 
ninth myotome where the sixth tubule should be found. Al- 
though the parts of it are also to be seen in two consecutive sections 
(the one represented in fig, 86 and another preceding it), 
the communication of its lumen with the collecting duct is not 
to be found anywhere. In some embryos, the tubule loses the 
connection with both the body-cavity and the duct. The structure 
{pt.6) in question is, I believe, nothing else than the remnant of the 
sixth tubule which is in a stage of degeneration, and the duct (cd.) is 
doubtless the collecting duct between the tubules of the fifth and 
sixth pairs. Compare the segmental duct (sd.) on the left side with 
the collecting duct (cd.) on the right side just described ; the latter 
has a wide circular lumen, whilst in the former it is slender and 
compressed. This difference of character between these two 
ducts is noticeable for some time in the younger stages. 

To sum up ilie results obtained in this stage the tubules of the 
third to the ßj'th pairs are vigorously developed, ivhile the second 
is very weak, the sixth retrograding, and the first has entirely 
disappeared. 

In the present stage, a peculiar structure is observed inside 
the walls of the body-cavity (figs. 77-85, pp.1-3). At some points 
of the peritoneum, a thin plate which consists, in cross-section, of 
one, two, or three cells, projects from the peritoneal wall into the 
body-cavity; it will be called here shortly the '"''peritoneal parti- 
tion.'''' A peritoneal outgrow^th is found at the level wehere the 
coelomic projection passes over to the visceral layer of the lateral 
plate ; at the same level, another peritoneal outgrowth starts up out 
of the parietal layer. These two outgrowths meet at midway 



MORPHOLOGY OF CYCLOSTOMATA. 349 

and cut off a long chamber from the body-cavity along the 
openings of the nephrostomes (figs. 77-85, pp. l). This longitudi- 
nal chamber communicates anteriorly as well as posteriorly with 
the body-cavity below, which is represented, in those parts, ])y the 
boundary of the parietal and visceral layers of the lateral plate. 
This is the uppermost partition. The second partition is weaker 
in development and is detected a little more ventrally, projecting 
likewise from the parietal and the visceral layers of the lateral 
plate. It is most obvious in the region beneath the third and 
fourth nephrostomes (81-84, pp.2). We find the third partition 
still more ventrally, which is weakest in development ; its extent 
is almost the same as the second (figs. 81-84, pp.3). 

These partitions disappear after a short existence ; in a little 
older embryo, none of them is detected, as will further be seen. 
This is probably the same structure as the *^ pcritmieale Scheide- 
wände " or *' Pentonealbrüche " described by Goette in Petromyzon 
fluviatUis ('90). As to the meaning of the structure I have 
nothing to sa}^^^ 

It is important here to illustrate the topographical position 
of the pronephros and the relation of it to other parts ; for these 
become definite for the first time in the present stage. For this 
purpose, a series of sagittal sections is instructive (figs. 112-114)."' 
A few anterior myotomes {/nt.lI-V) are seen in fig. 112, which 
represents the section nearest the median line. In the posterior 
part of these myotomes, four cell-layers are distinguishable ; the 
outmost layer (ep.) is the epiblast ; the cell cord (cd.) inside the 
epiblast is the longitudinal section of the collecting duct, and 



l)See the liistorical review under Peironyzon. 

2)Tlie embryo, from which these figures are drawn, is a little younger than tliat just 
spoken of. 



350 s. HATTA : 

its caudal continuation {sd.) is the segmental duct ; while the inner 
two layers {m.v. and 7ii.p.) present respectively tlie parietal and 
visceral layers of tlie lateral plate. Below these structures, the 
roof of tlie enteric canal covers the fore-gut (fff.) and the com- 
mencement of the mid-gut which forms the passage of the enteric 
cavity from the anterior slender portion to the posterior wide 
cavity. In fig. 113 which represents the section next outside the 
last, the lateral walls of the five myotomes from the first to the 
fifth {int.I-V) are noticed ; the cell-mass {au.) seen next anteriorly 
to the first myotome {mt.l) is a slice of the wall of the left auditory 
pit. The cross-sections of the pronephric tubules from the second 
to the fourth {pt.2-4) follow immediately behind the fifth 
myotome ; an oblique section of the fifth tubule (pt.ô) and the 
nephrostomal part of the sixth tubule (j^t.O) are also obvious 
behind the fourth tubule. The nephrostomes of the second 
{nst.2) and fifth {nst.ô) tubules are seen only in part, while a 
larger part of the third, fifth and sixth nephrostomes is seen in 
the next figure (fig. 114, 7ist.3, o, and 6') which represents a 
section still further outside. The sixth tubule, which is of weak 
development and has a wide nephrostome, is visible in these two 
sections (figs. 113 and 114, pLG and nst.G). Except the first tubule 
which has already disappeared without leaving any trace, the five 
tubules are all thus seen in the dorso-lateral aspect of the hind 
section of the fore-gut and the commencement of the mid-gut. 
The pronephros is thus situated in the neck which connects 
the head-protuberance with the globular abdominal portion. 
Below the pronephros, the narrow passage of the fore-gut passes 
through to unite the fore-gut with the wide space of the mid-gut, 
where after \vards the liver [1.) is found. Underneath the passage 
of the fore-gut, a group of mesenchymatous cells {mch.), which 



MOEPHOLOGY OF CYCLOSTO:\LVTA. 351 

constitutes the earliest fundament of the heart, is detected. The 
present position of the pronephros, — dorsal to the heart, anterior 
and dorsal to the liver, and along either side of the chorda, — 
is retained by it for a coinparatively long period (see fig. 97) ; 
in later stages, the liver is somewhat shifted backwards, so 
that the pronephros now comes entirely in front of it (see 
fig. 115). 

In an older embryo of this stage (figs. 92-96) the median 
folds of the coelomic projection, the component cells of which are 
very much flattened out, go in deeper towards the median line 
to meet with its counterpart on the opposite side. The second 
tubule (figs. 92 and 93, pt.-J) has become weaker, as a com- 
parison of these figures with figs. 77 and 79 will show. On the 
contrary, the tubule of the next pair (fig. 94, pt.S) has much 
elongated and is bent considerably in dorso-lateral direction, 
so that we can no longer observe the nephrostome together with 
the tubule itself on the same section. The following tubule, 
the fourth (fig. 95, pt.4), is likewise well developed ; the fifth 
(fig. 96, pt.5) is more or less weak in development as compared 
with the tubules of the two foregoing pairs. In short, these 
three pairs (from the third to the fifth) make parallel pro- 
gress with the development of other structures, for instance, the 
mesenterial fold or the muscle-segments. This is a fact that 
is to be observed too in the younger embryo of this stage, as 
above described. At the present stage, we can find no trace of 
the tubules beneath the ninth myotome, where the tubules of the 
sixth pair ought to be found, but only the cross-sections of the 
collecting duct or the anteriormost part of the segmental duct 
are seen. 



352 s. HATTA : 

Tliu>i, the tubules of the third, fourth, and fifth, pairs con- 
tinue to grow, lühile the fird pair has disappeared^ in the early 
part of the present stage (or at the end. of the foregoing stage) ; 
the sixth has already coinmenced retrogression and the second, is 
also growing iveaher and weaker. 

Tn the oldest embryo of this stage, there is to be seen no 
marked change in tlie ])ronephros, but tlio peritoneal lining is 
reduced into a very thin plate of :i deiinite epithelium every- 
where except at the pericardial portion, where the cells still have 
a columnar shape. The mesenchy matous cells accumulated on 
the median ventral line of the body are arranged in a certam 
order to be transformed into the cardiac tube. The third, fourth, 
and fifth tubules are also markedly prolonged and project into 
the body-cavity so as to cause the parietal layer of the perito- 
neum to fold between the epiblast and the body of the tubule 
(see this Journal : vol. x, Pt. xviii, figs. 8, \\ and 10). Tn some 
of the embryos, the tubules of the second pair undergo degenera- 
tion. I have met with, in this series of sections, the same con- 
dition of the sixth tubule as on the riglit of fig. 86, tlie right 
tubule having entirely disajipeared. 

Period 4. 

The embryos in the next advanced stage (Stage v) are much 
diminished in size, assuming a form of a retort or of a pistol 
('97, fig. 1, F). Figs. 98-106 represent sections through an 
embryo of this stage. The posterior larger section of the fore- 
gut comprising the pronej^hric region, has been reduced into a 
slender tube [fg.) which is bounded by almost a single layer of 
high cylindrical cells. The parietal layer of the peritoneum as 



MORPHOLOGY OF CYCLOSTOMATA. 353 

well as the coelomic projection ir.///.) is very much decreased in 
thickness and encloses the peritoneal cavity [pp.c.) that has now 
become spacious, while the visceral peritoneum is still thicker than 
other parts. The mesenchymatous cells found in the foregoing stage 
on the median ventral line are transformed into a thin layer of the 
endocardium of the heart and its anterior continuation {h. and tr.a.) 
which are suspended by the dorsal and the ventral raensenteries, 
and enclosed in the thick pericardial coat still forming a con- 
tinuation of the peritoneum. I can detect, however, none of the 
traces of the peritoneal partition which was developed so markedly 
in the hist stage that one could not possibly overlook them. I 
have endeavoured to trace tlie mode of disappearance of this 
structure, but have only found that in one lot of embryos the 
whole set of the structure was present while in the other no trace 
of it was perceptible. Unfortunately I have found no embryo in 
an intermediate condition. 

The few cells observed from the last stage beneath the chorda, 
and also in the space outside the pronephric tubules on both 
sides are more or less multiplied. As the former group is 
transformed tinally into the dorsal aorta, and the latter into the 
anterior cardinal vein of either side, I shall call them the tract of 
the dorsal aorta and of the anterior cardinal veins respectivel}'. 

Fig. 1)(S represents the section through the hind border of 
the branchial region. On either side of the enteric tube a small 
space {pp.c.) of the body-cavity is surrounded by the peritoneal 
epithelium, still consisting, in this part, of somewhat cubical 
cells. The ventral edges of the peritoneal membrane of both 
sides are just meeting at the median ventral line. A few mesen- 
chymatous cells .{tr.a.) found in the space between this meeting 
point and the ventral wall of the enteric canal, are destined to 



354 s. HATTA : 

form the anterior continuation of the cardiac tube or the truncus 
arteriosus. An irregular cell-structure {x) is seen on either side 
above the dorsal corner of the body-cavity and inside the tract 
of the anterior cardinal vein. It is this structure about which I 
could not at first decide with certainty whether it was a slice of 
the hind wall of the branchial chamber or a part of the pro- 
nephric tubule. All the cases examined, however, point towards its 
being a part of the tubule ; the structure is detected in the anterior- 
most part of the ])ody-cavity which wedges in, at about this 
stage, to the branchial region with a sharp angle (see fig. 97). 
The narrow space (fig. 08, pp.c.) found intervening between the 
structure and the peritoneal walls is a part of this cavity. One 
might suppose that the space may be the coelomic cavity of the 
branchial region ; but, the sjmce between the parietal and the 
visceral peritoneum of the branchial region is consolidated already 
in the preceding stage, being filled up with variously shaped cells 
of mesenchymatous nature (see fig. 07). 

In the next following section shown in fig. 00, a tubular 
structui'e {'pt.2) with an oval outline is seen on either side at 
the place where the pronephric tubule ought to be found. Its long 
axis is directed just like a tubule (compare with figs. 101, 102, ifcc. ). 
This is doubtless a part of a pronephric tubule ; but the corres- 
ponding nephrostome which ought to be found either in the 
section in front (fig. 98) or behind (fig. 100), can not be detected 
in either of them. The nephrostome must, therefore, be looked 
upon as having degenerated ; and since this pair of the tubules 
is, in fact, detected underneath the fifth myotome, it must be 
identified as the second pair of the tubules. The section represented 
in fig. 100 shows on both sides the cross-sections of the collecting 
duct {ccL). On the left side, a cellular structure connects the 



MORPHOLOGY OF CYCLOSTOMATA. 355 

peritoneum and the collecting duct ; it is the posterior wall of the 
tubule in figs. 99. Fig. 101 represents the section through the axial 
plane of the third tubule, the nephrostomes of which are recog- 
nized more clearly in the section behind it (fig. 102, nst.3). The 
tubules of this pair are comparatively not long. The fourth pair 
of the tubules and their nephrostomes are obvious in fig. 104 
{pt.4 and nst.4) which represents the third section behind that of 
fig. 102 ; the tubules much resemble those of the pair in front, 
showing the same convolutions as these. It is a peculiarity of the 
present stage that the aperture of the nephrostomes of the third 
and the fourth pair is not so wide open as in the last stage or 
as in more advanced stages Î It is always nearly closed and slit- 
like, so that we can hardly trace the communication between 
the lumen of the tubule and the body-cavity. 

Fig. 103 represents the section intervening between the sections 
shown in figs. 102 and 104. On the right side, the collecting 
duct alone, and on the left side, the duct together with a small 
part of the fourth tublile, is shown. The peritoneal membrane 
on the dorsal end of the body-cavity is folded far into that 
cavity (fig. 103, bs,). This fold is traceable from the anterior part 
of the third tubule to the hind part of the fourth (figs. 100-104, fe.). 
The space enclosed in this fold communicates freely with both the 
tract of the dorsal aorta under the chorda and the tract of the 
anterior cardinal vein outside of the })j()nephros and contains a 
number of mesenchymatous cells which probably wander in from 
the tract of the aorta and the anterior cardinal vein. As sub- 
sequent history shows, this structure constitutes the beginning of 
the glomerulus of the pronephroi^. 

Figs. 105 and 106 represent two contiguous sections immedi- 
ately posterior to the section shown in fig. 104. In fig. 105 we 



356 s. HATTA : 

observe ou either side the cross-section of the collecting; duct 
{cd.) together with a part of the fourth tubule [ptA) ; the longi- 
tudinal section of the fifth tubule (pLo) is seen on the right 
side of fig. IOC), standing in wide communication [nst.ô) with the 
body cavity. This is the hindmost tubule. In the sections lying 
behind this, the cross-section of only the segmental duct is 
repeated. 

Thus the tubules of the second pair undergo, at the present 
stage, complete degeneration. This process begins, in this case, 
as above seen, at the nephrostome and proceeds upwards to the 
collecting duct, — a process which is just the reverse of what 
is observed in the reduction of the tubules of the sixth pair and 
probably also of the first pair, in Ijoth which cases the tubules are 
first cut oft from the collecting duct and the scîparation from the 
peritoneal cavity follows afterwards. 

Period 6. 

In the Stage vi, embryos have developed so far that all 
the organs have received their definite forms and proper position 
with the exception of the middle and the hind portion of the gut, 
whose development is much delayed on account of the yolk-mass. 
Having absorbed the yolk-granules, the component cells of most 
organs are much diminished in size. 

Figs. 107-110 have been drawn from a series of sections 
through an embryo in this stage. The enteric canal {fg.) is much 
diminished in diameter, presenting, in section, an elongated heart 
shape. The peritoneum becomes very thin in all its parts with the 
exception of the pericardium and the coat of the inuicm arteriosm, 
in which iio component cells are of cylindiical or cubical bbape. 



MORPHOLOGY OF CYCLOSTOMATA. 357 

The peritoneal membrane lining tlie enteric canal immediately 
behind the branchial region is also thicker as compared witli 
other parts (fig. 107), being composed of a single layer of cubical 
cells, — a ^peculiarity observed since the last stage (compare figs. 
98-90 with fig. 107). 

The pronephric tubules as well as the collecting duct are 
composed of a regular epithelium of cylindrical cells ; the former, 
moreover, are much prolonged and, in some parts (fig. 108), much 
coiled, so that the peritoneal cavity which was almost a liollow 
space in the last stage, is filled up with the tubules and the 
cardiac tube. 

Fig. 107 represents the section through the hind part of the 
sixth myotome; a pair of the tubules {pt.3) is hanging down in 
the body-cavity immediately behind the hind wall of the branchial 
chamber. On the right side, the axial plane of the tubule is 
cut through, while, on the left, the anterior wall of it is sliced ; 
these are the tubules of the third pair. They show no bending 
in the antero-posterior direction, but are curved laterally and ven- 
trally. The component cells are, in the nephrostomal portion, 
taller in comparison with those in other parts of the tubule or the 
collecting duct. The fourth tubule and nephrostome are seen on 
the right side of fig. 108, while on its left side, the communication of 
the corresponding tulnile on the opposite side with the collecting 
duct is recognizable. The left ne2:)hrostome is found in the third 
section behind this, which is not figured. This pair of the tubules 
exhibits, in section, constrictions at two or three points owing 
to their curving somewhat in the antero-posterior direction (see the 
tubule on the right side of fig. 108). Fig. 109 is the section im- 
mediately behind the last and shows the cross-sections of the 
collecting duct {cd.) and a piece of the left fourth tubule {j)t.4). 



358 s. HATTA : 

A pair of the glomeruli (figs. 108 and 109,^/.). is seen adher- 
ing on the median side of the tubule on each side and lined with 
the visceral peritoneum. The glomerulus represented in the 
last stage by a folding of the peritoneum which covers the tubules 
from the third pair to the fourth^*, is reduced, at present, into a pair 
of sacs of this membrane projecting on each side between the 
fourth and the fifth tubules ; the other part of the folded membrane 
becomes adhered firmly to the walls of either the tubules or the 
body-cavity leaving no space of sacculation, — in short, a pair 
of long fokls, extending from the anterior part of the third 
tubule to the fiftli in the last stage, is reduced into a pair of sacs 
found in the position just mentioned. The inside of the sacs 
is compactly filled up with free-cells and communicates with 
the aorta tract and with the space outside the pronephros, 
where free-cells to be afterwards transformed into the anterior 
cardinal vein have been observed already from the foregoing 
stage. 

The section represented in fig. 110 fortunately passes sym- 
metrically through a pair of the nephrostomes {nst.a) and of the 
tubules ipt.a) hanging down in the peritoneal cavity. This is 
the fifth or the hindmost pair of the i^ronephric tubules in the 
present stage. The communication of the tubules with the collect- 
ing duct is seen in the section behind this. The tubules present 
also some antero-posterior bendings. Posterior to this, no tubule 
is found. 

The pro7iephric tubules m the j^resent stage are, therefore, re- 
duced into the miniinum number, i.e., three pa ir.s~\ all of 2ühieh are 
retained so long as the organ functions as the excretory apparatus 

1) See p. 355. 

2) We occasionally find the fonr tubules to persist, and the additional tubule is the sixth. 



MOEPnOLOGY OF CYCLOSTOMATA. 350 

during the larval life of Petronyzon. E>^pecially it mud he noticed 
that the foreiiiod i^air of the pershteiit tubules {the third pair) is- in 
close contact with the hind border of the hind uudl of the hi-anchial 
chamber ivhc re, in the foregoing stage, the second pair of the tubules 
was found, this latter Jiaring d isapjwarcd in the course of the 
last stage. It follows that the two somites, to which the first and 
the second, pair of the tubules have belonged, hare now entered into 
the formation of the branchial region. 

The development of the pronephros after this consists only 
in tlie prolongation and the convolution of the tubules, no further 
change taking place with reference to the number of the tubules 
or to their histological structure, until the system undergoes 
degeneration to be replaced by the mesonephros, which functions 
as the excretory organ for the whole subsequent life of Petro- 
myzon. 

The convolution of the tubules is hard to make out. I 
have reconstructed them from a number of sections ; some of 
these are diagramatically given in the annexed woodcut 
(Woodcut 2). 

With the growth of the muscle-segments the collecting duct 
is prolonged, so that the points of connection of the tubules 
with that duct become farther apart from one another, while the 
nephrostomal portions of the tubules retain more or less their 
original positions ; in this Avise, the tubules are laid in oblique 
positions directed anteriorly and posteriorly [A) and have no other 
curvature than the ventro-lateral bending (the frontal projection 
of the curvature is shown in F). Then the antero-posterior 
bending begins to take place. The foremost tubule is curved 
forwards in its whole length, while a small curvature in the distal 



360 



s. H ATTA : 



(nephrostomal) portion of the two following pairs is directed 
backwards. The nephrostomes retain their first position {B). 

Now the secondary curvatnres take 
place {0). The nephrostomal part 
of the foremost pair is crooked just 
like that of the two hind pairs in 
B ; the middle tubule is bent for- 
wards like the foremost tubule. 
The hindmost tubule makes a 
small forward curvature and a 
large backward bending. In the 
next stage, D, the foremost and 
the middle undergo no marked 
change, but the secondary curva- 
tures of the hindmost tubule are 
much more strongly expressed. In 
E the foremost receives a second- 
ary curvature directed backwards 
at the middle part ; the middle 
acquires a curvature in opposite 
direction; the hind tubule undergoes 
no marked change except in the 
increased de2;ree of the ori2;inal 
curvatures. It seems that the 
subsequent bondings always take 
place in the curved portion until there arises a system of com- 
plexly convoluted tubules filling up the chest cavity. 

As has been said, throughout these phases the positions of 
the nephrostomes are not markedly changed, retaining the same 
condition for a considerable period. The bendings of the tubules 




Wood-cut 2. — Di ashrams 
sliowing the convolutions 

of the tubules: in later stages, 
t. pronephric tulMiles. 

Sil. segmental dnct. 



MORPHOLOGY OF CYCLOSTOMAïA. 361 

are caused, therefore, by the growth of the tubule at tlie point 
of bending. 

The curvature in tlie vcntro-hiteral direction is ver}^ simple 
and undergoes no remarkable change ; its projection is showw in /'. 



B. — The Segmental Duet (cml the Genital Cells. 

For the sake of simplicity, the development of the segmental 
duct and of the vascular system in the pronephros haa been 
entirely put aside in the description given above. 

As already alluded to, the origin of the segmental duct in 
Petroiny^on is extremely difficult to make out, because its formation 
goes on rapidly at a comparatively young stage. The early 
process of its formation is essentially the same as in the prone- 
pliric tubules. In the anterior region, the infer //lediate cell-mass 
or the nejjJiTotome (see p. o40) behaves itself in precisely the same 
manner as in the Anlage of the pronephric tubules ; the difference 
is that it is cut off from the lateral plate and is transformed into 
the duct, while in the case of the tubule it retains the continuity 
with the lateral plate. If fig, 31, which represents the section 
through the tenth somite {i.e. the somite, from which backwards 
the Anlagen are converted to the segmental duct) be compared 
with the left half of figs. 2, 5, 6, 14, and the right half of 
fig. 16, in which the Anlagen all develop to the pronephric 
tubules, it will be found that there is no difference between them ; 
in fact, they are morphologically equivalent to one another. 
Such an Anlage is, posterior to the pronephros, not confined to 
the tenth somite, but, as has been already repeatedly said (pp. 317, 
320, and 327), is observed for some segments further backwards 
(see fig. 17). 



362 s. HATTA : 

The Anlage thus pronounced in each somite soon assumes a 
characteristic oval form, being completely cut off from the myo- 
tome to which it belongs (compare the right side of fig. 3 with the 
right side of fig. 58 and see the description on p. 335). The mode 
of constriction is also the same as in the case of the pronephric 
tubules ; the indentation begins at the anterior and posterior 
borders of the somite, and the middle portion is cut off last 
(compare with the explanation on p. 320). 

Here also, the coelomic projection is formed in the same mode 
and at the same point as in the case of the pronephric tubules (see 
left side of fig. 63, r.y;.). 

Up to about this time, the Anlage shows a feature much 
resembling that of the tubule, so that one who has not followed its 
further history might mistake it for a pronephric tubule (compare 
the left side of fig. 63 with figs. 67-74). But cell-mutiplication 
which occurs almost invariably in the case of the pnmephric 
tubules, is not observed in the Anlage of the segmental duct 
which is soon cut off from the lateral plate (including the coelomic 
projection) and assumes a characteristic tubular structure com- 
posed, in cross-section, of radially arranged cells of columnar shape. 
Its position is always on the parietal aspect of the dorsal (proxi- 
mal) angle of the peritoneum where the coelomic projection passes 
over into the lateral plate (see fig. 75). This separation of the 
Anlage of the duct from the lateral plate goes on, it seems to me, 
on tlie whole from the anterior part to the posterior, but often ir- 
regularly ; for not infrequently, the duct in some anterior somite is 
connected with the lateral plate, while it is already cut off com- 
pletely in posterior somites. In fact, there are some somites in 
which the separation is very much delayed and I have often been 
sur})rised to find what appeared like a prone})hric tubule in a 



MORPHOLOGY OF CYCLOSTOMATA. 363 

somite (see fig. 63) far backward of the posteriormost tubule 
which is found in the ninth somite. 

The segmentally arranged Auhigen of the segmental duct 
are secondarily united with one another just as in the case of the 
collecting duct in the pronephric region. This union seems to 
take place during the separation of the Anlage from the myotome 
and is finished before it is separated from the lateral plate ; for, 
when the Anlage first comes int(j view, there is no intersomitic 
cord as in the case of the pronephric tubules and the duct is seen 
already consisting of radially arranged cells (fig. 58) when it is 
cut off from the lateral plate. When established, the duct is the 
same in structure in both the somitic and intersomitic spaces ; a 
cross-section of such a duct in the intersomitic portion is shown 
ou the right side of iig. 75, while that in the somitic portion 
is seen on the left of the same figure. 

This condition of the duct is already traceable, in Stage iir, 
for no fewer than 10 somites from the hindmost pronephric tubule 
backwards, and it forms a direct posterior continuation of the 
collecting duct. The duct remains awhile as a solid cord of cells 
arranged radially in cross-section, but it soon acquires a lumen 
(figs. 75, 86, and 87, .sd). The further development of the duct 
goes on more promptly than that of the tubules in the hind part, 
and therefore, the embryos at such a stage (Stage iii) have a well 
developed duct and more or less primitive tubules (compare fig. 
74 with fig. 75). 

In the hind region, where yolk-cells are crowded, the pro- 
cess is much delayed and more or less modified. Instead of the 
diilerentiation of the cells in dtu, it seems to me, a few cells 
are detached from tlie nephrotome; a number of cells is produced 
by repeated division of these cells (fig. lU, aM.) and becomes ar- 



364 s. H ATT A : 

ranged a« in the Anlagen in the anterior region. Fig. 89 represents 
the section through the twenty-eighth somite in the series of sections 
shown in iigs. 77-86 ; it is the hindmost section in this series of 
sections, in which the cells just spoken of are detected ; there are 
found a few cells {((..sd.) of this kind which show no definite struc- 
ture, hut are scattered. In the next anterior section (fig. 88) the 
cells are arranged more or less radially. In the sections lying 
further anteriorly to this a perfect tuhe is formed as seen in fig. 87 
(•sï/.) which shows the frontal section through the seventeenth to 
twenty-third somites in the same series as the above two figures.^^ 
In what somite this nKxlified mode of the formation of the duct 
begins I can not tell with exactness, but it is certain that the duct 
arises by the differentiation of the nephrotomic cells in situ more 
than 10 segments back of the hindmost pronepliric tubule. I have 
considered it possible that tliese cells {a.-sd.) might be epiblastic in 
origin, Ijut I can not find that the cells composing the epiblast over 
this cell-group show any sign of mutiplication ; while on the 
other hand, the cells on the dorsal edge of the lateral plate 
(which corresponds to the nephrotome in the anterior part) are 
very active. I see, therefore, no escape from the conclusion that 
these cells are mesoblastic in origin. 

Also in the anterior part (jf the body, the epiblast consists 
throughout these phases of development always of a single 
layer of cubical cells and shows a sharp contour against the 
structure inside it, being, in most cases, intervened by a space. 
Naturally, mitotic figures are observed at several points, but the 
products of these cell-divisions contribute only to the extension 
of the epiblast itself, as may be inferred from the direction of 



1) By the l)ending ul" tlic body-axis, some sections in a series of cross-sections arc iin- 
avoidablv cut throutrh frontallv. 



MOEPHOLOGY OF CYCLORTOMATA. 06/) 

the spindles, tlio long axes of which are directed always parallel 
to the surface of the layer. / have nowhere observed any traee of 
either the proliferation or of the casting off of eells from the einhlaü 
to give rise to the segmental duet. 

In the cloacal region, the formation of the segmental duct 
o-oes on a little earlier than in the reojion next anterior to it. 
In sj^ite of much effort, I fiiiled to observe the very beginning 
of the formation at the cloacal opening, and I have nothing to 
tell of its earliest stage. In the series of sections from which 
figs. 77-89, are drawn, I can not yet find in the adjacent part 
of the cloaca any trace of the duct ; but in the section repre- 
sented in fig. 00 wdiich passes through the dorsal lip of the 
blastopore of an embryo with about the same number of the 
mesoblastic somites (34 or 35) as the one just referred to, the 
duct already breaks through into the cloacal cavity {eo.sd.Y\ Fig. 
91 represents the next ventral section which passes through the 
dorsal part of the blastopore {hp.). As seen in these two sections, 
immediately inside of the blatlopore [hp.], where the hypoblast 
passes over into the epiblast, the walls of the cloacal cavity send out, 
right and left, a symmetrical pair of diverticula-^ {c.dv.), forming 
an acute angle, the inner side of which is a part of the enteric wall, 
Avhile its outer side is the direct continuation of the epiblast. 
The walls of this diverticulum pass over into the segmental duct 
{sd.). The eoiiimunieation of the segmental duet with the eloaral 
cavity is found, therefore, at the point where the epiblastie layer 
of the lip is reflected inside and passes over into the hypoblast. 
This point of communication is, however, shifted far inside and 



l)This opening is found in the same vertical plane as the n4th or 35th somite. 
2)Tlie right diverticulum only is seen in figs. 90 and 91, the left one being ol)served in 
another section which is nniigurcd. 



366 s. HATTA : 

clorsally when the development proceeds further (fig. Ill, co.sd. 
and c.dv.). 

I have also met with two cases (figs. 00 and 111), in which 
I have observed some epiblastic cells of the external lateral 
walls of the blastopore multiplying actively and having mitotic 
spindles [x) with axes directed perpendicularly to the plane of the 
epiblast, while the duct comes in firm connection with that 
point of the epiblast, — the connection is so firm that the duct 
and the epiblast appear to form one and the same tissue. At 
this point, thus, there is every appearance of epiblastic cells 
partaking in the construction of the segmental duct. 

The collecting duct pertaining to the ninth somite forms 
the segmental duct in that segment, having lost the connection 
with the tubule. 

Up to Stage II, the duct is represented by the segmental 
Anlagen in about 8 segments back of the ninth somite ; in Stage 
III, these Anlagen are converted into the duct in about 10 an- 
terior segments ; while in the course of Stage iv it opens out 
into the cloacal cavity. 

From the above account, it is easily conceivable that the An- 
lage of the segmental duet and that of the jjronephrie tubule are 
perfectly homologous, and that the duet is a continuation of a series 
of abortive pronephric tidmlex in the hind region. 

Underneath ten and more myotomes lying posterior to about 
the fifteenth somite the proximal portion of the lateral plate, 
which corresponds to the nephrotome, contains peculiar large 
cells (figs. 87, 88, and 89, //r-.) loaded with an enormous quantity of 
yolk-granules ; the other mesoblastic cells in this part, being 
much flattened out, form a thin layer o^or these cells. These 



]\rorvPTTOLOc;Y of cyclostoimata. HOV 

peculiar cells are, I think, the equivalent of the primitive 
genital cells found in the corresponding part of the Amphibian 
and Selachian body. 

Up to Stage Tir, these cells can not be distinguished from 
other mesoblastic cells which are equally rich iu yolk-granules. 
In Stage iv, they become conspicuous ; and in Stage v, again 
indistinguishable from other constituent cells of this part. 



C. — The Vascular System in the Fro?iephros. 

In early stages, no trace of the vascular system is perceived 
in the pronephros. What is recognisable as a fore-runner of the 
vessel is represented by mesenchymatous cells scattered in the 
space between the primary germinal layers (figs. 77 and 82, 
7?ich.). These free cells are detected, during Stage iv,^' in three 
tracts, viz., beneath the chorda, beneath the ventral wall of the 
enteric canal and outside the jn-onephric tubules on either side (figs. 
77, 79, 80, 81, and 82, mch.). In Stage y, or at the end of Stage iy, 
the cells below the hind section of the fore-gut are converted into 
the endothelium of the heart and of the vessels which are its 
direct continuations. The cells beneath the chorda are destined 
to be transformed into the dorsal aorta, and the cells on either 
side of the pronephros constitute the first indication of the cardinal 
veins. It is these three vessels — the aorta and the two cardinal 
veins — which come iu relation wüth the pronephros. 

In the embryos in which the degeneration of the tubules is 
still going on, there is no special vessel supplying the pronephros; 
but when the process is over, a pair of long blood-spaces (figs. 
100-104, bs.) is found in communication with the aorta-tract. 

1) A few of them are observed here and there already in Stage III. 



368 s. HATTA : 

They are the spaces formed by the slackening and fokling of the 
median peritoneum which coats the pronephric tubules, as above 
stated (p. 355). The fold, i.e. the space, extends throughout 
almost the whole length of the pronephros (figs. 100-104) and 
contains numerous free-cells. But, as the peritoneum finally 
adheres to the median walls of the tubules in each nephric 
segment, the space becomes divided into three pairs : the anterior 
pair, which soon disappears, is found between the tubules of 
the second and third, the middle is detected between the third 
and fourth, and the posterior between the fourth and fifth tubules. 
These spaces communicate directly, — medially with the aorta tract 
and externally with the tracts of the anterior cardinal veins, 
which emerge, in later stages, in the pronephros. They are the 
blood-spaces which, I believe, correspond with the intersomitic 
arteries demonstrated by Paul Mayer and others in Selachia. 
AVhen the tubules develop further, the arterial portion of 
these blood-spaces disappear except the middle portion where it 
is sacculated and filled up with the mesenchymatous cells 
(figs. 108-109, gl.). This portion is the structure which is called 
the glomerulus of the pronephros ; it is found one on each side 
(see pp. 355 and 358). 



Having followed, in the foregoing pages, the successive pro- 
cesses which take place in the development of the pronephros 
in Petronyzon, step by step, I will give a short resume of the facts. 

1. In the earliest part of Stage ii, the ni esoblast consists 
simply of A\\q parietal (dorsal) and the visceral (median and 
ventral) layers. The proximal portion of the mesoblast is dis- 



MOEPHOLOGY OF CYCLOSTOMATA. 369 

tiiiguislied, first of all, in the histological structure from the 
distal portion : the former is composed of columnar cells, and 
the latter of irregularly shaped cells. Only the proximal portion 
Avhicli occupies the largest part of the mesoblast undergoes the 
metameric segmentation and gives rise to the scleromyotome and 
the nephrotome (in the sense of Kuckert) ; the distal smaller 
portion remains unsegmented and is later converted into the 
flattened epithelium of the peritoneum. 

2. The earliest traces of the pronephros are noticeable in 
exceedingly young stages, that is, in the early part of Stage ir, 
in which the embryo has about 16 somites. 

3. They are expressed in the form of a diverticulum of the 
parietal layer of that section in each mesoblastic somite, which 
forms the ventral half of the segmental part of the mesoblast 
and is called the nephrotome. This is the Anlage of the pro- 
nephric tubule and not of the segmental duct. 

4. The pronephric diverticulum or the Anlage is brought 
about by the evagi nation of the parietal layer in each nephrotome, 
enclosing a part of the primary coelomic cavity. 

5. The nephrotome is separated from the proximal portion 
of the segmented mesoblast and forms awhile the proximal portion 
of the unsegmented mesoblast or the lateral plate. The separation 
begins with an indentation in the anterior and posterior borders 
of the mesoblastic somite ; the myocoelome communicates for some 
time by a narrow passage with the general coelomic cavity. 

6. The Anlao-e has no histological connection either with 
the preceding or the following Anlage or Avitli the other 
germinal layers ; it is, tlierefore, segmental in origin and myo- 
meric in positiou. 

7. The Anlagen are developed, in Stage ii, in about 12 



370 s. tïATTA : 

segments and are eut off from the scleromyotome in 4 segments. 
In Stage iii, the separation of the Anlage from the myotome 
goes as far backwards as the sixteenth or seventeenth segment. 

8. The anteriormost Anlage is found in the hind part of the 
fourth somite and is the first to arise ; the second follows it, and 
so forth. 

9. The Anlagen in each somite are secondarily united with 
one another by the solid cellular cord which is budded out of 
the anterior and posterior rims of the Anlagen themselves ; thus 
the collecting duct {Saiiiinelrolir in the sense of KÜckert) is estab- 
lished. This j^rocess is originally to be looked upon as the 
com in a; to2;ether of the ends of the tubules. 

10. The canalization of the collecting duct begins within each 
Anlage and proceeds, generally speaking, posteriorly, until the 
Anlagen in front and back are put in free communication. 

11. Each Anlage grows dorso-laterally and acquires a tubular 
form. The collecting duct is shifted gradually in a dorso-median 
direction ; finally it comes to lie between the myotome, the 
mesentery, and the chorda dorsalis. 

12. The tubules open in the coelomic cavity at the lateral 
angle of the dorsal corner of that cavity. 

13. In the somites posterior to the ninth, the tubules are, 
during Stage iii, cut off also from the lateral plate and establish 
a long duct running, on each side, along the dorsal aspect of 
the lateral plate where originally the tubules opened. Tliis is 
the segmental duct ; the tubules and the coUectino; duct in the 
somites anterior to this constitute the glandular part of the 
pronephros. 

14. The glandular part, or the pronephros proper comprises 
six somites, from the fourth to the ninth. The maxinuim number 



MORPHOLOGY OF CYCLOSTOMATA. 371 

of the prouepliic tubules which in attained by the embryo in 
Stage III, is, therefore, six pairs. 

lo. The tubules of the first and second pairs come, in Stage 
III, temporarily in close contact with the epiblast, but do not 
receive cells from it ; they soon return to their original condition. 

16. The anterior extremity of the system shows, from the 
first, degenerating features. The first, second, and sixth of the 
tubules begin, during Stage iii, to decline ; and at the end of 
Stage IV, or the beginning of Stage v, the tubules are reduced 
into the minimum number, which consists of three pairs from 
the third to the fifth. These three pairs function as the actual 
excretory organ for a considerable length of time. 

17. Ketrogression is first met \vith in the first pair of tbe 
tubules, which decline probably without further development, 
soon after their separation from the myotome is completed ; they 
seem to atrophy from the free end. The next pair degenerating 
is the sixth, which is at first cut off from the collecting duct 
and remains for a short time, but soon disappears without leaving 
a trace. The second pair persists for some time seemingly to 
function as the excretory organ, but it atrophies already in 
the early part of Stage v, the communication wnth the coelomic 
cavity being first obliterated ; and in Stage vi, none of the struc- 
ture remains to be recognized. 

18. The foremost pair of the persistent tubules comes to 
lie in close contact with the hind wall of the branchial chamber. 
The two mesoblastic somites wdiich correspond to the first and 
second nephromeres should therefore be looked upon as having 
entered into the formation of the branchial region. 

The stages in which the tubules appear and abort in different 
somites are shown in the annexed table. 



372 



ö. HAÏTA 





l-H 
1 


S 
ce 


1—1 

h-l 
O 

m 


(—1 

O 
«2 


> 

3 

et! 


S 
o 
CO 


S 


5 

CO 


I— 1 

o 

ce 


'A 

S 
S 

CO 


y, 

o 
Cß 




Stîige II 








Ani. 1 


A ni. 2 


Aul. 3 


Anl. 4 


Anl. 5 


Anl. 6 


Anl. 7 


Anl. S 


Anl. 9 


Stage m 








Tub.l 


Tub.2 


Tnb.3 


Tub.4 


Tub.5 


Tub.6 


iSegmental duct. 


Stage IV 










Tub.2 


Tub.3 


Tub.4 


ïub.o 


Tub.6 


Segmental duct. 


Stage V 










ïub.2 


Tub.3 


ïub.4 


Tub.5 


Segmentai duct. 


Stage VI 












Tub.3 


Tiib.4 


Tub.5 


Segmental duct. 



19. în older embryos of Stage in, the visceral layer of the 
iiephrotome is folded out, and is called the coelomic projection 
which resembles the coelomic pocket described by Price for 
Bdellostoiiia ; however, in Bdellostoma, the fold is derived from 
iliö pariclal and visceral layern of the lateral plate and is after- 
wards converted into the Bowman's capsule, whereas the coelomic 
projection is the product of only the elsceral layer of the nephro- 
toiiie ; it gives rise to the radix of the mesentery which offers 
materials to the mesonephric tubules and to the gonads. 

20. The topographical position of the pronephros becomes 
first definite in Stage iv. It is situated in the chest cavity, dorso- 
lateral to the heart, forward of and dorsal to the liver, extending 
along either side of the chorda. This position is somewhat 
changed as the development proceeds ; the pronephros comes, 
in later stages, in front of the liver. 

21. During Stage iv, a structure, which I have called 
above the peritoneal partition, is observed as an outgrowth 
of the peritoneal wall and disappears during the same stage. 



MORPHOLOGY OF CYCLOSTOMATA. Ô/Ô 

This horizontal partition is found at three levels. The most 
dorsal is well developed, the ventral is a mere trace, and the 
middle is intermediate between the above two. I can not state 
at present anything definite as regards the significance of this 
striictare. 

22. The convolution of the pronephric tubule takes place in 
Stage IV. With tlie growth of the myotome, the collecting 
duct is prolonged ; consequently the connecting points of 
the tubules with the duct are farther removed from one another 
than before, whilst the nephrostomes retain their original 
position ; so that, the two posterior pairs of the tubules are placed 
in an oblique direction from dorsal and caudal to ventral and 
cranial. Each tubule is, then, convoluted in a cranio-caudal 
direction between the heart and the lateral peritoneal wall. In 
older stages, the tubules are coiled in all directions, until the 
chest cavity becomes filled up with the convolution of the tubules. 

23. Up to Stage vr, the nephromeres and the myomeres 
exactly coincide one above the other in position. This period is 
ver}'- long in comparison Avith other Craniota. As the develop- 
ment proceeds further, the pronephic tubules are however shifted 
gradually backward, so that, in Ammocoetes 10 mm. long, the 
myotomes are already not situated upon the tubules pertaining 
to each of them. In later stages, nothing of the relation can be 
traced. 

24. The segmental duct is looked upon as being brought 
about by the union of a series of the abortive pronephric tubules 
in about 12 somites lying posterior to the eighth somite^\ The 
Anlage is laid in the parietal layer of the nephrotome iu exactly 
the same manner as in the pronephric tubules of the glandular part. 

1) Iu the 9th somite, the aborted tubule actually forms the duct iu the segmeut. 



d/4 S. H ATTA : 

The difference is, that the tubules in the posterior region are 
soon cut off from tlie lateral plate and become the duct. 

2Ö. Between the epiblast on one hand, and the Anlage or the 
duct on the other, there exists always a space, and the duct 
has no connection with tlie epiblast except at its posteriormost 
end where the epiblastic cells might, as judged from the mitotic 
figures, contribute to the formation of the duct. 

26. In the somites posterior to about the twentieth somite, 
tlie Anlage of the duct is represented by a few cells in each 
segment probably detached from the dorso-lateral angle of the 
nephrotome. These cells multiply and are transformed into 
the segmental duct in the posterior part. 

27. In Stage ii, the Anlagen of the segmental duct are cut 
off from the mother layer in a few somites ; in Stage in, the 
duct is formed as far as about the eighteenth somite, while in 
Stage IV, it breaks out into the cloacal cavity. The cloacal 
opening of the segmental duct is found at a point where the 
hypoblastic cloacal wall is reflected into the epiblast, these two 
layers forming a diverticulum on either side. 

28. In stage iv, the primitive genital cells become apparent 
in the nephrotomes of the posterior 10 or more somites ; they 
can not be discriminated from other mesoblastic cells in the 
next advanced Stage. 

29. The blood-vessels, which specially supply the pronephros, 
acquire definite form in comparatively later stages, viz., at about 
Stage V. The dorsal aorta pours out the blood into two pairs of the 
blind vessicles, which are formed by the folding of the parietal 
peritoneum and are found between the first and second pairs, 
and between the second and third pairs, of the persistent tubules 



MOKPHOLOGY OF CYCLOSTOMATA. O/O 

respectively. The venous blood is carried away tlirongli the an- 
terior cardinal veins which penetrate the pronephros. 

20. These Idood-spaces are thus segmental in arrangement 
'<\m\intn'xoiintic in position. The two anterior pairs of them soon 
nndo]-go atrophy, Imt the posteriormost pair persists, becoming 
enlaro-ed and sacculated at the distal extremitv. This sacculated 
part of the vessel is filled up with free-cells and is called the 
glomerulus of the pronephros, and, therefore, there is only a pair 
of glomeruli in Petromyzon. 

II. Historical Review and Conclusions. 

As is well known, ]Max Schultze ('-56) was the first who dis- 
covered the pronephros in Petromyzon. Having investigated the 
larvœ of P. lyhineri, the author describes the structure as 
"Drüsenanlage" and homologised it with the "Urnieren (Wolf'sche 
Körper) " of the frog's larva. His statements on this body are as 
follows : " Nicht lange nach der Bildung dieser Drüse (Thymus) 
entstellt die Anlage einer zweiten, aus dem unter der Chorda 
dorsalis ans-ehäuften Blastem über dem Herzen. Aus der durch 
Pigmentanlagelungen früh schon sehr undurchsichtig w^erdenden 
Masse wachsen nämlich nach unten, gegen das Herz zu, 3 oder 
4 kurze Fortsätze hervor, welche eine eigenthüraliche Wimperung 
zeigen" (p. 30). 

The stage spoken of probably corresponds to Stage v, or vi, 
of my embryo. 

Our knowledge on this subject received important additions 
by the noted investigations of W. Müller and Max Für- 
BRIXGER. MuLLER ('75) uoticcd the first traces of the pro- 
nephros in a very young embryo, which had yet only four pairs 



376 s. HATTA : 

of gill-slits. This Anlage gives rise to a much coiled gland, which 
opens into tlie body-cavity, at first through only one ciliated 
funnel, but afterwards tlirongh four. The gland passes over 
posteriorly to a pair of ducts, which run along the chorda on 
either side and open into the cloaca. INUtller has homologised 
the structure with the "■' V ornière " of jlfjjxinc and called the 
duct " Urnierengang " (pp. 121-122). He found a pair of glo- 
meruli projected on the median surface of the gland and lined 
with the peritoneal epithelium. 

Max Furbringer ('78) studied the larva3 of Ammococles 
planen', which varied from 4.5 to 180 mm. in length. His state- 
ments essentially confirm Muller's. In his account we find the 
following sentences : '' Die auf allen Präparaten ausgebildete 
Vomiere, die ich im Wesentlichen ganz wie ]\1Üller fand, bildet 
einen nahmentlich bei den mittleren Stadien voluminösen und 
durch 4-5 Myokommata erstreckten Complex von Windungen, 
die vorn durch mehreren Peritonealcanäle (Wimpertrichter) in 
Bauchhöhle münden nnd hinten in den Vornierengang übergehen. 
Diese auf die 2-3 ersten Myokommata beschränkten Trichter 
ragen in unregelmässiger Folge bald ventral-medial, bald ventral- 
lateral in die Bauchhöhle vor und wurden (von Carlberla 
und mir) meist zu fünf gefunden. Die von rundliche Epithel- 
zellen bekleidete Glomerulus verhielt ganz wie Mïjller beschreibt" 
(p. 42).. 

The larvse of Ämmocoetcs in cjuestion seems to correspond 
probably to Stage v, or later stages of ray list ; in such a stage, 
I could not find more than three (or rarely four) pairs of the 
tubules, or of the nephrostomes.^^ 

1) See the foot-note on p. 358. 



3I0IIPH0L0GY OF CYCLOSTOMATA. 377 

The authors who have investigated the development of 
Petromyzon embryos step by step, are W. Scott, Goette, Shipley, 
and V. KuPFFEE. Their opinions are, however, somewhat 
divergent. Scott ('82) derives the pronephric tubules from the 
segmental duct which is, according to liim, brought about by 
the difiereutiation in situ of the cells forming the proxhnal 
margin of the lateral plate. The process takes place in the 
whole extent at the same time. At certain points (segmental ?) 
of the duct thus formed, evaginations are produced out of it ; 
these evaginations subsequently open into the body-cavity and 
establish the nephrostomes which are, according to Scott, 
found from two to three pairs in number. At about the stage 
in which the funnels are formed, he observed a pair of 
glomeruli. 

" In most respects," Shipley's observations ('87) " confirm 
his " (Scott's). But " on the origin of the ciliated funnels, the 
results differ from Scott's " and agree witlî those of Furbringer 
(Amphibian pronephros ?). According to Shipley, '* in the region 
of the heart, where the body-cavity has already appeared, its 
origin {i.e., of the segmental duct) seems to ])e somewhat different. 
The lumen of the segmental duct here becomes continuous with a 
groove in the parietal peritoneum, lying near the angle where 
the somatopleure and the splanchnopleure diverge. When this 
groove closes it leaves four or five openings which persist as the 
openings of the ciliated funnels" (p. 20). 

v. KuPFFER^^ ('88) observed, in P. planeri, the three pairs of 
the tubules arising from three distinct evaginations of the parietal 



1)1 know this paper uuly by Llie abstruct iii : JahresbericliL ii. die. Fortschr. d. Anal, 
u. riiysiul., Ed. 17. ISbO. 



378 s. HATTA : 

layer of the lateral plate ; the segmental duct is looked upon as 
of the epiblastic origin. 

GoETTE ('90) worked the development of Petromyzon fluvia- 
tilis ; his results with respect to the pronephros show some agree- 
ment with mine, especially those concerning the later stages. The 
author derives also the whole system of the pronephros (including 
the segmental duct) solely from the mesoblast. But we diverge in 
some important points from each other ; he has found the earliest 
traces of the structure at a time when the rudiment of the heart 
first becomes apparent (his vi. Periode) (p. 64). From the account 
given in the foregoing pages it is clear that this period belongs 
to a later stage in wdiich the pronephros has already made a 
considerable progress in development ; his figures 99, 103, &c., 
which are spoken of as representing the first appearance of the 
structure, approximately correspond with my figures 82, 83, &c., 
and with those of even older stages. 

The pronephros is, according to Goette, not of a separate 
Anlage in its first appearance, but arises in a form of a longi- 
tudinal furrow formed, on each side, by an evagination of the 
parietal layer of the mesoblast ; the lips of the furrow being fused 
at certain points, there remain three openings ; these are 
converted afterwards into three tubules and ciliated funnels. 
The tubules are added by stages until there are usually five, or 
more rarely four or six ; but how these are multiplied, he can 
not say with certainty. The tubules have, it seems to him, no 
relation to the metameres of the body ; for 3 to 5 tubules are 
found in the extent of 2 to 3 metameres (loc. cit., pp. 64-65). 

Th(:' .segmental duct originates, according to Goette, in pre- 
cisely the same way as tlie pronephros proper ; the only diflerence 
is the complete constriction of it from its mother-layer just as 



MORPHOLOGY OF CYCLOSÏOMATA. 870 

I have made out. From the region of the liver-aulage backward 
the development of the duct is irregular ; he says : " Auf der 
einen Seite zeigt sich seine Anlage noch rinnen förmig, während 
sie auf der andern Seite schon vollkoniinen röhrenförmig ab- 
geschnürt ist. Endlich wechselt dies Verhalten auch auf derselben 
Körperseite, so dass derselbe Gang, von der Lebergegend rück- 
wärts verfolgt, bald rinnen-, bald röhrenförmig, geschlossen oder 
mit offener Lichtung sich darstellt " {loco cit.^ p. ÖQ). The hind 
end of the duct opens in the cloaca (Afterdarm) by the fusion of 
their walls and by the communication of the lumen of the duct and 
the diverticulum of the cloaca. I have not observed in any stage 
of my embryos examined the numerous convolutions of the seg- 
mental duct demonstrated by Goette in the region immediately 
behind the " ursprüngliche Kopfniere." 

Goette has made out the three " peritoneale Scheidewände," 
as he calls them : two respectively in the anterior and the posterior 
end of the pTonephros, and the third on either side of the liver. 
Later, the first contributes, according to him, to the formation of 
the hind wall of the branchial pouch (Kieraentasche) ; the second 
is converted into " eine Venenbrücke zwischen dem Sinus 
venosus and der Leibeswand," while the tliird disappears without 
leaving a trace. They are, according to Goette, homologous 
with the " Schlussplatte " of the pronephros in Teleostei ; con- 
sidered phylogenetically, nevertheless, they have no intimate 
relation to the pronephros in Petromyzon {loco cit., pp. 56-61). 
This structure is, as stated on p. 349, doubtless the same as the 
uppermost peritoneal partition which I have found in my em- 
bi'yos. I have nothing to communicate on its signiücauce ; 
but I feel sure tliat his statement is not accurate when he says 
the structure appears earlier than the ])runeplirüs ; for his ligs. 



380 s. HATTA : 

96 and 97, to which his statement refers, represent a stage con- 
siderably later than the first formation of the pronephros itself. 
And the peritoneal partition is not confined to these three points, 
but is continuous throughout the whole extent of the pronephros ; 
moreover, beside tlie " peritoneale Scheidewände," there are found 
two other partitions of a similar character as above stated. 
Also, as to the fate of the structure my results differ from his : 
I have not been able to observe at all any such contribution 
to the formation of the hind wall of the branchial chamber and 
of the " Venenbrücke," as is affirmed by Goette. 

Kabl ('96) says in his recent extensive work on the Selachian 
nephric organ, that in quite young larvœ of Pelromyzon flaviatilis 
the pronephros also begins in the seventh somite, in which the 
first of the four ostia are found, as in Pridmruiy}^ His larvœ 
are, however, oOl hours or 20 days and 21 hours old ; such 
larvœ correspond to my embryos in Stage vi, and upwards, in 
which anteriorly two pairs, and posteriorly, one pair of the 
tubules disappeared and only three persistent tubules are seen. 
His first nephrostome represents the foremost of the persistent 
nephrostome. 

The accounts cited above all agree with the results given 
in the present paper in deriving both the pronephros and the 
segmental duct from the mesoblast alone, with the single excep- 
tion of V. KuPFFEE, who assumes the epiblastic origin of the 
segmental duct. They difter from the account given in the fore- 
going pages in the mode of the formation and in the number of 
the tubules formed. The first point of difference is due to the 

1) See the reference under Selaeliia (p. ÖUO). 



MORPHOLOGY OF CYCLOSTOMATA. 381 

fact that the authors probably overlooked the earliest phases of 
foruiation, which take place, as shown above, in a stage very 
young, but not youuger in comparison than that in other Anarania ; 
for the formation follows the metamerio î<cgiii<niation of the uuxohlast 
ill the anterior i-ogion. Tn later stages, tlio tu1)nles and the inter- 
soinitic portion of the collecting duct repeated in sections of a 
series appear, indeed, lila^ tlie cross-sections of a longitudinal 
furrow or groove of the lateral plate, the lips of which are fused 
at certain points, as described by Shipley and Goette (see my 
figs. 66-74). 

The number of the tubules and nephrostomes varies accord- 
ing to the stages of development. And if some stage or stages 
are overlooked, it must necessarily lead to an erroneous conclu- 
sion. This is the probable reason why the statements of the 
writers with reference to the number differ. 

Indeed, the anterior extremity of the pronephros has already, 
from the first appearance, the features of a rudimentary organ ; the 
first pair of the tubules can not be observed at the same time 
with the following five pairs, except by extremely good luck. In 
some embryos of Stage in, we see occasionally the collecting duct 
alone in front of the first tubule, so that we are led to infer 
that there were some pairs of tubules in front of the present 
first pair, which have degenerated during the course of tlie 
ancestral history. ^^ 

As is seen above, all investigators who have been occupied 
with the study of the development of Petromyzon agree in de- 
scribing only one pair of glomeruli. Shipley says ''there is only 
one glomerulus on each side, stretching on each side of the 

1) I have stated above that in the earliest part of Stage iii, the anterior extremity of 
the left collecting duct presents a conical protuberance (see the footnote on p. ,'Î29)- 



382 s. HATTA : 

alimentary canal extending tlirougli about the same space as llie 
glandular part of the kidney. Each glomerulus is a diverliculum 

of the peritoneum, which generally becomes sacculated; " 

(p. 1^1). The statements l)y Goette confirm SiiirLEY's, and ray 
results also agree witli theirs. However, this is not all of the 
vascular system of the pronephros but represents a posterior 
portion of it, the anterior part having disappeared entirely 
(see p. 3G8). 

No previous writer on Petromyzon has described such early 
stages as given above in the development of the pronephros, nor 
has any one remarked the temporary existence of the i^ronephric 
tubules in the branchial region as well as in the reoion of the 
segmental duct, I will, therefore, extend the comparison over the 
allied groups such as jMyxinoids and Amphioxus, and higher 
Cran iota to verify the new facts. 

With reference to the development of the nephric organ in 
Myxinoids, there is a great deal of information which we owe 
to the unwearied labors of W. Müller, Bemon, Weldon, and 
others^'. They had, however, no opportunity to observe the earliest 
stage of the embryos. Recently our knowledge on this subject 
has been greatly augmented by the new works of Price, Dean, 
and Maas. 

Price ('97) worked out the early development of the prone- 
phros observed in a few embryos at different stages of Bdellostoma 
i<touti. According to him " the first indication of the system 
occurs here in the eleventh segment (of spinal ganglion), and 
consists of a simple thickening of the somatic layer of the coel- 

])I liave not seen tlie paper by J. Mullfr. 



MORPHOLOGY OF CYCLOSTOMATA. 383 

omic epithelium, which extends through seven section?, 

the thickening has not been caused by a proliferation of cells, 
but certain cells having assumed the form of columnar epithe- 
lium, while the adjoining cells retained the form of flat epithelium. 

later an evagination will liere take place, to form a 

segmental tubule " (p. 209). These evaginations are connected 
with one another " by a streak of columnar epithelium, which 
in transverse section lesembles the first tubule anläge, except that 
there is no concavity on the lower surface ; this is the segmental 
(collecting) duct." " The union between the duct and tubules is," 
in another place he says, '' primary and not secondary " 
[loco cit., p. 210). 

The pronephros in Bdellostoma comprises, according to Price, 
69 segments (spinal ganglions). As it begins at the transverse 
plane opposite the eleventh spinal ganglion, it is inferred 
that the pronephros in Bdellostoma is extended over the whole 
length of the branchial region. But " the excretory system 
disappears through the greater part of this region before the 
gills are formed " {loco cit., p. 217). 

The segmental duct (in .^. .s/r.) is, according to the author, 
brought about by the rudiments of the hinder 20 degenerated 
tubules (in his Stage C) ; the number of the declining tubules 
increases by stages : in Stage A, there are two ; in Stage B, 
nineteen ; and in Stage C, twenty. 

This account is thus in close agreement with that given in 
the present work, excepting a slight difference as to the origin 
of the collecting duct and as to the number of the tubules. In 
Bdellostoma, the collecting duct develops out of the Anlagen 
independent from that of the tubules, while in Petromyzon, as 
stated in the foregoing description, the Anlage in a mesoblastic 



384 s. HATTA : 

somite develops solely into the tubule, and by tlie secondary 
union of the tubules' ends, the collecting duct is brought about. 

As regards the number of the tubules, tliere are, in 
Pelromyzon, only two pairs in the branchial region instead of 
twenty in Bdrllxtoma. The number is, however, of secondary 
importance ; it varies with the stages of embryos and possibly 
with individuals, and naturally more with tlie embryos of 
different families. This nuuierical variation is readiW explained 
by the degenerating tendency of the tubules. 

Price has made out the segmental evaginations of the dorsal 
corner of the coelomie cavity corresponding to the nephromeres ; 
they are called by him the " coelomie pockets." In Fetromyzon, 
I have found a series of solid knobs on the visceral layer of 
the intermediate cell-mass, which are transformed into the 
segmental folds of epithelium, forming then the direct continua- 
tion of the peritoneum. Thus the coelomie pocket in Bdellostoma 
and the coelomie projection in Petromyzon are apparently very 
similar structures; the two, however, differ from each other in origin 
and in fate. The former (coelomie pocket) is constructed by the 
parietal and visceral layei's of the lateral plate, while the latter 
(coelomie projection) is the product of only the visceral layer 
of the nephrotorae, the ventral half of the segmented part of the 
mesoblast. The coelomie pockets become the Malpighian body, 
and the coelomie projections give origin to the radix of the 
mesentery, from which the gonad-cells and the mesouephric 
tubules are derived. Nevertheless, these two structures are, I 
believe, homologous. Price's statements on the derivation of the 
coelomie pocket from the two peritoneal layers, are not as clear as 
is desirable, and its partition fiom the body-cavity might, it 
seems to me, represent the uppermost peiitoneal partition which 



MORPHOLOGY OF CYCLOSTOMATA. 385 

soon disappears, in Pelromyzon, without any definite significance. 
At any rate, the structure represents " parts of the original 
segmental coelome, that is, the nephrotoine," an unmistakable 
fact which is denied by Price. 

The embryos of ^lyxine which formed the materials of the 
valuable works by Maas are too old to be compared with those 
of Petromyzon used in the present work. But the results ob- 
tained by the author differ from those of Price in an important 
point, namely, in the derivation of the mesonephros. 

The pronephros and mesonephros are, according to Price, 
different parts of the same organ. " If the organ in question could 
only be a pronephros alone, or mesonephros alone," says Price 
'' I should unhesitatingly pronounce in favour of its being a pro- 
nephros " [loco cit., p. 120). And he proposes to call " the 
entire embryonic kidney holonepkro'<.''^ With Pabl, Maas, and 
others, I hesitate to accept Price's conclusion ; for there are, 
as may be inferred from his statements, great gaps not only 
between the Stages B and C, but also between Stage C and the 
adult. The formation of the mesonephros takes place in Petro- 
myzon only at a stage much advanced, in which the processes of 
the formation and degeneration of the pronephros go on in much 
the same manner as in Bdellodoma, and it is open to doubt if the 
mesonephros might not appear in later stages which were lacking 
among Price's materials. 

Up to the oldest embryo observed by Price there were 
neither glomeruli nor bloodvessels of a definite form, although there 
were found in the splanchnopleure some vessels whose position 
seemed to suggest their corresponding to the glomeruli of Selachia 
and Amphioxus ; " but they do not have any relation to the 
openings of the tubules, nor have they any direct connection 



386 s. HATTA : 

with the aorta " {loco cit., p. 213). The glomerulus figured in 
his Taf. 17, fig. 12 {(jl) corresponds, I think, with a part of the 
ü'lomerulus in Petrorniizony 

As is very well known, the independent studies of Weiss 
('00) and Boveri ('02) on tlie branchial chamber of Amphioxus 
gave a new direction to the morphological investigation of this 
field. A number of external openings of ciliated tubes is found 
at the dorsal corner of the peribranchial chamber of Amphioxus. 
Through the morphological study and physiological experiments 
this organ-system is demonstrated to be the excretory apparatus, 
or " Nierencanälchen," as Boveri calls them, of Amphioxus. 

Boveri counted 01 "Nierencanälchen" in an individual 4 cm. 
in length and possessing 183 gill-bars on the right side. In the 
adult he counted about 180 of the '' Nierencanälchen "; the number 
is, however, by no means constant, but varies within a certain limit. 

In the middle region of the branchial chaml^er, a " Nieren- 
canälchen " has 3 or 4 " Seitentrichter," and 2 " Endtrichter ;" 
such is the most complete one. It becomes gradually simplified 
both anteriorly and posteriorly, until it is at last represented by a 
short single tubule, as seen in Taf. 33, figs. and 13, given by 
Boveri. The tubules in the anterior and posterior part of the 
system thus show a sign of degeneration, as in the case of the 
pronephros of Cyclostomata. 

1) Dkan has published two ijapers on the development of the Californian Hag ('98 and 
'99) ; these excellent works contain merely the general account of the course of the develop- 
ment in surface view- VVe may expect that tlie full account will throw much light on the 
ontogeny of Craniota. There stand, in the account given by him in these works, the im- 
portant facts that the " pronephric tubules are apparent in connection with all the meso- 
blastic somites" ('93, p- 274) and that the pronephros is extended far backwards, beyond the 
anal region, into the tail ('99, p. 27-)- It would be liighly desirable to observe the pronephric 
tubules of the Hag in relation tu the myomere, and not to the spinal ganghia alone, as PracE 
has done. 



MOEPHOLOGY OF CYCLOSTOMATA. 387 

These nephric tubules receive, according to Bovert, the blood 
from the aorta, which gives two branchlets to each nephric 
segment. These branchlets form in each sesrment a network in 
the neighbourhood of, and winding around, the nephric tubule ; 
it is this network that Boveei calls <jlomerulus. 

From the structure, the position, the segmental arrangement, 
the physiological function, and the relation of the blood- vascular 
system to this system of organs, Boveri regards the latter 
as a primitive form of Vertebrate nephric organ and horaologised 
it particularly with the pronephros of Cran iota. The points 
of difference which exist between the " Nierencanälchen " of Ä}ii- 
phloxus and the pronephros of Craniota, have been smoothed away 
by the author's masterly arguments. The first of these points 
is the want of the segmental duct in Amjjhioxiis ; but this is re- 
presented, according to Boveri, by a part of the peri branchial 
chamber. The second is the relation of the nephric segments to 
other systems of organs. The " Nierencanälchen " is branchio- 
meric while the pronephros of Craniota is myomeric, in arrange- 
ment. But this difference is looked upon by him as only apparent ; 
for the number of gill-slits first formed agrees with that of the 
muscle-segments in the same region ; this is sufficiently demons- 
trated by the figure given by Weiss {foco cit., fig. 3). 

Thus the author has brought the " Nierencanälchen " of 
Amphioxus into perfect harmony with the pronephros of Craniota. 
Some additional light is now, I believe, thrown from the side of 
Craniota by the facts obtained in Cyclostomata, the lowest 
class of Craniota. This harmony will be brought out more in 
discussing the development of the pronephros in Selachia, Teleostei, 
and Amphibia, which will be treated further on. 



388 s. H ATTA : 

Thanks to the labors of many eminent investigators, tlie 
early development of tlie Selachian pronephros has been, as is well 
known, fully studied, so that the facts gathered from this field 
are well adapted to be compared with those from other groups. 
I have found, in the present investigation, many important points 
running parallel with the development of the Selachian pro- 
nephros. I may then be allowed to compare my own results in 
Petromyzon with those already arrived at in Selachia. Reference 
will, however, be limited to those works which are sufficient to 
verify the points I wish to bring out. 

Through the excellent work of Kuckert ('88) we can best 
learn the origin of the pronephros in Selachia. " Die erste Anlage 
der Vorniere " is recognised " in Form einer gegen den Ectoblast 
gerichteten Vorbuch tung des parietalen Mesoblasts." This Anlage 
is first brought about by the thickening of the parietal layer of 
the mesoblast, which is found " in den Bereich des segmentirten 
Mesoblasts, d.h. Somiten " (p. 209) ; this thickening is called by 
the author '' Segmentalwulst." The foot-note also runs as follows : 
" Der Ursprung des Segtnentalwulstes reicht ventral bis zu der 
Stelle herab, wo die Somiten in den unsegmentirten Mesoblast 
der Peritoneal wand übergehen " (p. 209). The " Segmentalwulst " 
is so called because it is noticed as the segmental thicken- 
ing of the parietal mesoblast of which E,uckert recognised, 
in his Stad. ii^\ six for Toqjedo and four for Prisimrus, stretching 
over a corresponding number of the myotomes. The first indication 
of the pronephros is expressed, in Selachia also, segmentally in 
the segmental part of the mesoblast at the stage in which the 
metameric segmentation of the mesoblast is still going on, and 

l)Tlie embryos iu tlie stage have 25-27 somites. 



MOKPHOLOGY OF CYCLOSTOMAÏA. 389 

tii(3 myotome is not yet cut off from the lateral plate, just as in 
Petroiinjzoii. The foremost of them is found in the hind part of 
the third or fourth body-somite. The development of the Anlage 
in each segment agrees also with that in Pelromyzon ; for he says: 
" Der Segmentalwulst zeigt in vorliegende Stadium (Stad. i) 
regelmässig die stärkste Entwicklung in seiner mittleren Ab- 
schnitt, also ungefähr im Bereich des dritten des ihm gehörigen 
Somiten, und verjüngt von da allmählich nach seinem vorderen 

und hinteren Ende zu " {loco cit., pp. 210). 

The account given by Ruckert is essentially confirmed by 
later investigators such as van Wyhe ('89)^^ Rabl ('96), and others, 
although they differ from one another in the interpretation of the 
facts and in some unimportant points. Rabl looks upon the 
Anlage of the pronephros (his Vornieren wulst) as the ventral 
poition of the somite just as Ruckert does, ^Yhile it is, ac- 
cording to VAN Wyhe, the product of the lateral plate (his 
Hypomer). This is, as it seems to me, not a contradiction in the 
facts, but in the terms used ; for van Wyhe states : " Da nun der 
Pronephros, wie, spätere Entwicklungsstadien zeigen, ein Produkt 
der Seitenplatte ist, während der unmittelbar dorsal davon liegende 
Theil des Mesoderms zur Mittelplatte gehört, ist die Segment- 
irung des Mesoderms bei Selachiern also nicht auf die jMyotom- 
platte beschränkt, sondern erstreckt sich auch auf die Mittelplatte 
und den dorsalen Theil der Seitenplatte " {loco cit., pp. 474-47Ö). 
The fact is, therefore, no other than that the portion of the meso- 
blast dorsal to the ventral limit of the Anlage of tlie pronephros 
undergoes segmentation, and the portion ventral to this point 
remains unsegmented, constituting the lateral plate. I will, in 

1) The em})i7o, in which tlie first traces of the pronephros is seen, is, according to 
VAX Wyhe, in a stage witli 27 somite-, whereas Rabl has seen in an embryo o{ PriMiiiius 
with 25 somites. 



390 s. HATTA : 

this place, not go further, but return in future pages to the 
discussion of this point. It is, however, safe, I believe, to regard 
this portion of the mesoblast as a part of the somite. 

A'^AN" Wyhp: found the foremost pronephric segment in the 
third body-somite (his Rumpfsegment), and IIabl states that the 
Vornierenwulst begins in the seventh somite formed (his 
Gesammtsegment). According to Rabl, however, vax Wyhe's 
third Rumpfsegment corresponds to his seventh Gesammtsegment. 
To verify this fiict Rabl has extended the comparison over 
Peiromyzon, and found that in this case also the pronephros 
begins in the seventh somite ; but the pronephric tubule in that 
somite is, as noticed above (p. 380), not the anteriormost of the 
tubules in his sense, but of the persistent tubules. 

Vais^ Wyite noticed hv^e of the pronephric segments for 
Raja, and three for ScylUiün and Pristiurus ; while Rabl counted 
eight Yornierenwiilste for Raja, and four for Pristiurus, The 
results in Petromyzon, therefore, best agree with those made out 
by RÜCKER T in Torpedo. 

The authors agree in deriving the collecting duct from the 
lateral extremities of the pronephric Anlagen, where they 
become confluent. 

RuCKERT has observed, in Pristiurus, as well as in Torpedo, 
the secondary connection of the Segmentalwulst with the epiblast, 
which has led him to believe in some contribution of epiblastic 
cells to the formation of the pronephros, while vax Wyhe 
and Rabl deny tliis. I have found the same connection in 
Petromyzon, but I have found no sign of the contribution 
of epiblastic cells to the formation of the pronephros. The 
phenomenon is temporary in both Selachia and Petromyzon ; 
it takes place in Selachia, according to RÜckert, in his Stad. ir, 



MORPHOLOGY OF CYCLOSTO:\rATA. 391 

and already in his Stad. in, a space is seen between these two 
structures. 

The deg-eneration of the tuhules in Sehichia runs a course 
parallel with that mentioned under Am2:)hioxii.s and C^'clostomata. 
As seen above, tlie Anlagen of the pronephros are developed most 
vigorously in tlie middle part of the pronephros, as in the case 
of A'ii}pkwxu>^ and Cyclostomata ; and degeneration begins at the 
cranial and caudal extremities as there. 

Vax Wyhe says that the degeneration consists in a confluence 
(Verschmelzung) of the ostia. According to Ruckekt, the 
Vornierenfalte becomes simply flattened out in the cranial part 
of the pronephros. The reduction in the caudal part is noteworthy : 
the Anlao-en are here constricted ofl" from the mesoblast and con- 
verted into the anteriormost section of the segmental duct. 
Only the middle (the third) diverticulum (in Torpedo) persists 
in communicating with the bod^^-cavity and becomes the ostium 
ahdominale. 

In Petromyzon, I have unfortunately failed to observe accu- 
rately the manner of degeneration of the tubule in the cranial 
part. It is however probable that it begins either from the 
blind tip of the tubule (the first tubule), or by obliteration 
of the nephrostome (the second tubule). In the caudal part, the 
collecting duct is constricted off from the lateral plate by 
obliteration of the tubule and constitutes the foremost section of 
the segmental duct, in precisely the same manner as in Selachia. 
The difierence is : in Petromyzon the communication with the 
body-cavity is retained l»y the three middle nephrostomes, while 
in Selachia, it is through only the middle one, that is, the ostlunt 
ahdominale. 

The segmental duct becomes apparent in an embryo with 



392 s. HATTA : 

35 (van Wyhe), or 34 to 3ö (Rabl) somites. The anterior 
small section of tlie duct is formed, as just stated, in the same 
manner in Petromyzon and Sekichia. The mode of formation of its 
posterior hirger portion in Selachia differs from that of Petromyzon. 
RÜCKERT ('88) and van Wyhe ('88, '89, '98) believe that 
it is the product of the epiblast^\ while Eabl maintains its 
purely mesoblastic origin. At any rate, the posterior tip of the 
duct or the cord is sharply pointed and connected firmly 
with the epiblast throughout its growth until it opens into the 
cloacal cavity, which is effected, according to van Wyhe and 
Eabl, in the embryo with 83 to 84 somites. It can be inferred 
from VAN Wyhe's figs la and lb, that this communication is found 
in a plane vertical to the thirty-eighth Eumpfsegment'-^ In 
Petromyzon, the duct, being formed of a series of abortive pronephric 
tubules, has no genetic relation to the epiblast except in the 
cloacal region where the duct seems actually to receive cells from 
the epiblast, as fully stated above (p. 366). 

The nephric arteries of Selachia which were discovered by 
Paul Mayer without reference to their relation to the pronephros, 
were studied by Euckert and their true nature was pointed out 
by him. There are six of them in Torpedo corresponding to the 
number of the nephric segment ; they are, however, not somitic 
but intersoinitic in position. The vessels not only pass through the 
nephric fold, but throw a solid process, the interior of which 
consists of round or spindle-shaped cells. This is, according to 
EÜCKERT, the equivalent of the pronephric glomerulus of Am- 
phibia described by FÜrbringer. The development and decline 
of these vessels go on parallel with those of the pronephric diver- 

1)1 will return to tills point again in future pages. 

2) Accoi'ding to Rabl's counting, tliis somite corresponds to Iiis forty-secnnd somite. 



MORPHOLOGY OF CYCLOSTOMATA. 393 

ticula. The vessels in the cranial as well as in the caudal part 
of the pronephros are weaker than those in the middle (the third 
and fourth) ; only the latter vessels develop further and become the 
vitelline artery. Van AYyhe confirms Rückert's account and 
has described three vessels in Friüluriis. In addition to these, 
VAN Wyhe has pointed out the very small segmental vessels on 
the left side, which go not to the intestine, but to the body- 
wall. They are not equivalent to Ihe intestinal vessels on the 
opposite side. One of them gives a branchlet to the glomerulus 
which sends out, in its turn, a branchlet to the cardinal vein. 
The homologous vessels on the right side are to be seen coming out 
of the root of the vitelline artery. Boveri remarks that the 
vessels of Paul Mayer present many points of harmony with 
the branchial vessels in Amphioxus. Rabl agrees essentially wnth 
the account given by Rückert and van Wyhe, but denies the 
existence of a glomerulus. According to Rabl, the structure 
called the glomerulus by Rückert does not fulfil the condi- 
tions of being a glomerulus ; he says : " Eine einfache Ausbucht- 
ung einer Arterie ist noch keine Gefässschlinge, geschweige 
denn ein Glomerulus " {loco cit., p. 668). 

Most of the early investigators, who observed the develop- 
ment of the Teleostian pronephros, believe it to be mesoblastic 
in origin. There are very few writers as Ryder ('87), and Brook 
('88), who derive the segmental duct from the epiblast. According 
to Oellacher ('73), Goette ('75 and '88), Fürbringer ('78), 
and Hoffmann ('86), the first Anlage of the pronephros is 
brought about by the evagiuation of the parietal layer of the 
mesoblast at the level of the junction of the somite with the 
lateral plate, formiug thus a longitudinal groove on each side. 



894 s. H ATT A : 

which is subsequently constricted off from the body-cavity. 
This takes place at first in the middle region of the body, whence 
it proceeds both anteriorly and posteriorly. 

Oellacher observed that the Anlage is converted into a 
longitudinal canal or the segmental duct, being completely shut 
off from the body-cavity in both the anterior and posterior 
parts. The anterior section of the duct is much swollen and 
transformed into the pronephric chamber. From the dorsal 
aorta, a pair of branches is given off which pushes into the 
pronephric chamber, pressing against its median wall and giving 
rise to a pair of the glomeruli. This j^ortion of the duct becomes 
coiled up and constitutes the pronephros. 

Goette's view somewhat differs from the account given 
above : the anterior end of the longitudinal groove is not com- 
pletely closed from the body-cavity, but leaves awhile the com- 
munication Avith the latter, wdiich is, according to Goette, the 
morphological equivalent of the nephrostomes of the Amphibian 
and Petromyzon pronephros. Opposite this nephrostome, he 
says, the glomerulus is formed by evagination of the visceral 
peritoneum and projects freely into the body-cavity. This portion 
of the peritoneum together with the nephrostome is constricted 
from the rest of the peritoneum ; the coelomic cavity thus shut 
off is converted into the joronephric chamber. 

This view is essentially confirmed by subsequent writers such 
as Fuebkinger ('78), Hofemann i^^^), and others, although 
Hoffmann differs in his view of the mode of the formation of 
the glomerulus. 

According to the results recently arrived at by Felix'^ in 

1) I know ]iis paper only I)y the iihstract in the Jahreaberichte über die Fortschritte der 
Analomiexund Physiologie, N.F. Bd. 111. '97. 



MOKPilOLOGY OF CYCLUöTOMATA. 395 

the embryos of Salmonidse, the earliest traces of the pronephros 
consist, in embryos with 11 pairs of the somites, of five solid 
proliferations of the lateral plate which is already cut off from 
the somite. These proliferations, being coincident with the caudal 
half of the third to seventh somites, are strictly metameric 
in arranoement and are reofarded bv the author as the rudimentarv 
pronephric tubules. These tubules soon become confluent with 
one another to form a single outgrovvth of the lateral plate, which 
is called by the author the " primäre Vornierenfalte." The 
*' primäre Yornierenfalte," which passes over into the parietal and 
visceral layers of the lateral plate, undergoes a longitudinal 
constriction (the " sekundäre Vornierenfalte ") by which it is 
divided into the dorsal and ventral parts. From the former, 
the anterior section of the segmental duct originates, while the 
latter is transformed into the pronephric chamber. By stages, 
the dorsal part woanders laterally, and the ventral part travels 
medianwards. At the same time, these parts are separated from 
each other, leaving the communication at only one point, which 
is called the " Pseudonephrostom." 

This phase of the development of the pronephros observed 
by Felix is, as I believe, undoubtedly earlier than that looked 
upon by the previous authors as the earliest indication of the 
pronephros. 

At the time when the Anlagen of the pronephros are con- 
verted into the " Vornierenfalte," the Anlage of the caudal contin- 
uation of the segmental duct, becomes apparent in the eighth 
to the tenth somite; it is brought about by the division of the 
primary lateral plate (lateral plate in the ordinary sense) into (1) 
the secondary lateral plate (lateral), (2) the segmental duct (middle), 
and (3) the Anlage of the "Stammvenen" (median). This pro- 



39G s. HATTA : 

cess proceeds posteriorly until the duct comes to lie close to the 
rectum (Enddarm). 

Felix thus observed the segmental Anlage of the pronephros 
in its glandular part, and derives the rest of the system from 
the proximal margin of both tlie parietal and visceral layers of 
tlie secondary lateral plate ; he has observed neither the posterior 
growth, nor the epiblastic origin, of the segmental duct. 

The pronephric chamber which results from the confluence of 
the five pronephric tubules, is not homologous, according to 
Felix, with that of Amphibia, in which the chamber should be 
a constricted part of the body-cavity into which the tubules 
open. 

Quite recently, 8\vaen and Brächet ('99) have published a 
paper on the early development of the mesoblastic organs in 
Salamonidœ. Although my manuscripts were nearly finished, 
when I saw this interesting paper, I must here refer in a few 
words to it^\ 

The authors found the first traces of the pronephros under 
the fifth somite, of two embryos, one of which was in the stage of 
11 somites, and the other of 13 somites. It is not the product 
of the parietal layer of the lateral plate only, but is formed, 
as Felix believes, by the proximal portion of both the parietal 
and visceral layers of the secondary lateral plate (l'extremé interne 
de la plaque latérale secondaire"^). The internal cavity enclosed 
by the pronephros is, therefore, not the diverticulum, but a 
part of the body-cavity. 



1)1 am much indebted to my fiieiid, Dr. A. Oka, who read the paper iV)r uie- 
2)Accordhig to the authors, the " plaque hiterale primitive " is divided into the "plaque 

latérale secondaire" and llic "masse intermédiaire;" tlierefoie, the "plaque latérale 

secondaire" corresponds to the hiteral plate \t6L'\t' oi' Pel rum i/zon. 



MORPHOLOGY OF CYCLOSTOMATA. 397 

The Anlage of tlie pronepliros is laid in exactly tlie same 
manner from the fourth somite to the cloacal region. Under the 
anterior three somites from the fourth to tlie sixth, the Anlagen 
are developed into tlie pronephric chamber ; the Anlagen posterior 
to these are all transformed into the " canal excréteur," as they 
call the segmental duct, and they have come to the conclusion 
that the '' canal excréteur " of the pronei)hros has the morpho- 
logical value of a rudimentär}^ pronephric chamber. 

Tlie facts given in the last two papers, are thus in close 
accordance with one another as well as with those given by 
myself in the foregoing pages. Differences between their 
results and mine are that the authors derive the system from the 
lateral unsegmented mesoblast, and that both the parietal and 
visceral layers of it partake in the formation of the system. As 
has been stated in the descriptive part, this derivation is only 
apparent ; a little further study shows that only the parietal layer 
gives rise to the system, and this part of the layer belongs to 
the somite. Indeed, this part appears to form, for some time, the 
proximal portion of the lateral plate, being early cut otf from 
the rest of the somite. It must be remembered that this separa- 
tion is not the separation of the lateral plate from the somite, 
but that of the Anlage of the pronephros from the rest of the 
somite; or, the result of the development of the pronephros. It 
is merely for a physiological reason that this development or 
separation of the pronephric Anlage goes on earlier than, for 
instance, in Selachia, it performing in Teleostei the actual ex- 
cretory function. This will be understood easily, when a com- 
parison with other groups is made further on. 

It has been a well known fact that the development of 



398 s. HATTA : 

Amphibia shows, in several respects, a parallel course with that 
of Feiroiiiyzon. Careful observations on the development of 
AmjDliibian pronephros, adduced by recent investigators, have 
intensified this similarity with the exception of a few points which 
are, however, probably of secondary importance. 

Most authors wlio have worked on the Amphibian develop- 
ment agree in deriving the entire system of the pronephros from 
the parietal layer of the mesoblast only, and in regarding it as 
arising originally as a common pouch, the anterior part of 
which is divided secondarily, l)y a partial closure of the 
peritoneal communication, into a number of the pronephric 
tubules. 

This view has been advanced by earlier authors such as 
W. MÜLLER ('7Ö), GoETTE ('75), FÜebrtnger (78), Hoffmann 
('86), and others. The stage at which the pronephros appears 
coincides exactly with that in Pctromyzon, as Max Füerringer 
says in his well known work : " Die erste Entwicklung der 
Vorniere und ihres Ausführungsganges findet hier nach der 
Scheidung des Mesoderms in Urwirbel und Seitenplatten statt 
und folo't unmittelbar der beo;innenden Sonderuns; der ersten in 

r5 o o 

einzelne Urwirbel amd der Spaltung der letzteren in Haut- und 
Darmfaselplatten. Embryonen von Rana temporaria von circa 
2.0 Mm. Länge und von Triton alpestris von ca. 2.0 Mm. L. 
entsprechen diesen Stadium " (p. 3)^' 

MoLLiER ('90) has made out the segmental Anlage of the 
Amphibian pronephros, having worked with the embryos of Triton, 



l)The nephrostomes are found, according to tlie author: 

2 in Sal aman drina mandata, 3 in Rnna tempnraria, 

2 in Triton alpestris, 3 in Bombinata- ignens (Goette), and 

2 in Siredon pisciformis, 4 in Coecilia rosfrata (Spengel). 



MORPHOLOGY OF CYCLOSTOMATA. 399 

Bufo, and Rana. His accounts confirm, as a whole, those given 
by RÜCKERT for Sehichia above referred to, but differ some- 
what from those of most other authors who have worked on 
Amphibian pronephros. Mollier states as follows : " Wir sehen 
hier ebenfalls zuerst eine solide, von dem Mesoblast ausgehende 
Anlage, deren Structur anfänglich schwer zu erkennen ist und 
erst mit dem Hohlwerden, wie bei den Selachiern, klar hervor- 
tritt. Dann linden wir, dass hier zwei resp. drei getrennte 
Cimälclien vorhanden sind, die von den Somiten in conver- 
genter Kichtung ausgehen und erst nachträglich untereinander 
vereinigen zu einem Längscanal, von dem aus die Vornierent- 
richter in die Leibeshöhle führen " [loco cit., p. 229)/^ The 
author derives in this wise the pronephric tubules, exactly as 
in the case of Pctromyzon, from the segmented part of the 
mesoblast only. 

Mollier's accounts are for the most part in close accord 
with the results given by Field ('91, p. 282), who, one year 
later independently of Mollier, began with Anura, and 
extended the wT)rk over Urodele Amphibia. In one point, their 
results differ Avidel y ; but " the difference is," it seems to Field, 
" apparent rather than real." According to Mollier, the 
nephrostomes communicate with the cavity of the myotome, the 
myocœlome of van Wyhe ; this is denied by Field, who believes 
that *' the pronephric tubules have to do with the ventral seg- 
ment of the mesoderm " {loco cit., p. 283). It seems to me 
that this ** ventral segment of the mesoderm " corresponds to the 
pronephrotome of van Wyhe in Selachia or to " I'extreme interne 
de la plaque latérale secondaire " of Swaen and Brächet in 



1) It seems that tlie earliest traces of the pronephros are perceived in an embryo 
vounsrer than that with 7 somites. 



400 s. HATTA : 

Teleostei, and I agree with the view of Ruckekt, here represented 
by tliat of MoLLiER. 

MoLLiER and Field agree with each other in assigning three 
pairs of the tubules for Rana and Bufo, extending from the 
second to the fourth somite, and two pairs for Tnlon (Mollier) 
and Amhlystoma (Field), covering the third and fourth somites/' 
In addition to these, Mollier observed occasional occurrences of 
the third tubule in Triton, ^Yhicll is, according to Field, not 
equivalent, as Mollier maintains, to the third tubule in Bufo 
and Rana, because the additional third tnbule in Tnlon is fonnd 
in tlie fifth somite, while the third tubule in Rtana and Bufo is 
under the fourth somite. According to Semox, there are ten 
pairs of the tubules on either side of the body in Ichthyoj)hi>^. 

A pair of glomeruli has l)een made out in Amphibia ; 
the structure is connected by special vessels with the dorsal 
aorta on one hand and with the cardinal vein on the other. 
This branch of the aorta is believed by Field to correspond 
to a part of Mayer's vessels in Selachia. Beside these, there is 
no vessel arranged segmentally or otherwise. 

The section of the body-cavity corresponding to the pro- 
nephric stretch is gradually expanded, and is shut off temporarily 
from the rest of the cavity by a close contact of the parietal 
and visceral layers of the cœlome ; this part of the cavity is, 
according to Goette, homologous with the pronephric chamber 
in Teleostei and with the homologous structure in Peiromyzon, 
wdiich is called by him the '* peritoneale Scheidewände." 

The so-called ventral portion of the Amphibian pronephros 
is, according to Mollier, brought about by the separation of 

1) According to FielDj Mollier's first body-segment in Tiiton corresponds to his third 
somite in Amblyslonia. 



MORPHOLOGY OF CYCLOSTOMATA. 401 

the ventral portion of the pronephric Anlagen from the dorsal, 
'which latter is differentiated into the tubules and constitutes the 
dorsal portion of the pronephros. The ventral part of the An- 
lagen separated from the dorsal retains anteriorly its connection 
with the anterior most tubule and posteriorly with the segmental 
duct. It is prolonged and bent out anteriorly in front of the 
dorsal part. " Mollier's description is," Field says, " sub- 
stantially in accord with my own observation," [loco cit., 

p. 286). This feature of the duct shows, it seems to me, a close 
resemblance to the auteriormost section of the Teleostean segmental 
duet which is, as above referred tu, bent in the same fashion. 

The segmental duct arises, according to previous writers, as 
a longitudinal common furrow oft he parietal peritoneum, which 
lurrow is later constricted off from the mother-layer and becomes 
converted into a long canal. Molliee has observed the segmental 
duct transformed directly from the mesoblast, just like the 
glandular part of the pronephros, in the two somites behind the pro- 
nephros. Whether the greater remaining part of the duct is formed 
likewise by differentiation of the mesoblast, or by a backward 
growth of the hind end of the duct iirst formed, he could not decide 
with certainty ; but the observations of Field elucidate this point. 

"The segmental duct arises," Field says, "throughout its 
entire length by a proliferation in situ of the somatopleure " 
{loco cit., p. 223). The author has observed neither its epiblastic 
origin nor a free growth of its posterior end, except in the cloacal 
region Avhere it " grows across the cloaca free from adjacent tissue " 
{loco cit., p. 223). In Stage v, the cloacal opening is seen. This 
opening is found, in Fuuia and Bufo, in the vertical plane with 
the middle of the twelfth somite, whereas it is below the twentieth 
somite in Anihlyslonia (Field). 



402 s. HATTA : 

The duct is segmental in origin. Field says : " I believe 
I am justified in concluding that the segmental duct between 
Somites v, and ix, arises in situ from a thickening of the somato- 
pleure serially equivalent to that from which in the anterior region 
the pronephros is developed " [loco cit., p. 219). There are no 
other A'ertebrata which agree more with Petroinijzon with reference 
to the development of the segmental duct, than Amphibia. Indeed, 
here as there, the segmental duct is of segmental origin and is to 
be looked upon, as seen in Petromyzon, as the continuation of 
a series of abortive pronephric tubules in the posterior region. 

Authors who have observed the epiblastic origin of the 
segmental duct in Amphibia are very few. vox Perexji ('87) 
has published the results of his study on liana cscidcnta, but his 
note is unfortunately very short'- . This view is opposed, so far 
I am aware, by almost all recent observers. After bringing 
the results by him into harmony with those by KÜckekt in 
Selachia, Mollter says : " Im einen Punkte weichen die Amphibia 
von Selachiern ab, dass die Vomiere mit dem Ektoblast in 
keine nähere Beziehung tritt. Allerdings heftet sie besonders in 
den Stadien, in welchen sie voluminöser erscheint, dem Ektoblast 
oft in allallend inniger AVeise an. ''■' ''' '■' Doch lässt sich 
stets eine scharfe Grenze beiderlei Blätter ziehen, wenigstens bei 
Bufoj wo die Ektoblastelemente durch ihren Pigmentgehalt deut- 
lich gekennzeichnet sind " (loco cit., p. 229). 

Tlie historical review undertaken in the foregoing pages 
shows the agreement to a large extent of the results arrived at 
in several groups of Anamnia. Some points of disagreement are 
naturally met with ; but these are, I believe, only apparent. 

1)1 have not seen the jiaper ]iy I'.rouk. 



MORPHOLOGY OF CYCLOSTOMATA. 403 

III the groups above referred to, the first indication of the 
excretory system becomes apparent at a stage in which some 
mesoblastic somites are formed and the metameric segmentation 
of the mesoblast is going on. This is the Anlage not of the 
segmental duct, l)ut oî the pronephros. A single exception is found 
in Bdellostoiiia, in which the early traces of the system become 
visible, as we learn from Price, at a stage much more advanced 
than in other Anamuia, that is, at the stage in which the sclero- 
myotome is cut off from the rest of the mesoblast and mesen- 
chyraatous cells fill up the spaces between organs and organ-systems. 

I have endeavoured to reconcile the points, in which the 
views of the previous authors diverge from one another, under 
the following three headings : — 

A. — The Anlage of the Fronephr'ic Tubule is the Product 
of the Ilesoblastic Soinîle and not of the Lateral Plate. 

The view that derives the pronephros from a single common 
groove formed either of only the parietal, or of both the parietal 
and visceral layers of the unsegmented mesoblast (the lateral 
plate), is advocated by most of the authors who have worked 
on the development of Petromyzon, Teleostei, and Amphibia. 
This is due probably to the early separation of the sclero-myotome 
from the rest of the mesoblast in these groups. In them the 
Anlagen of the pronephros (or the nephric segments) together 
with the lateral plate are cut ofï from the sclero-myotome and 
form, for some time after this separation, the proximal portion 
of the lateral plate. It must be borne in mind that this se- 
paration is not the separation of the lateral plate, but of the 
nephrotome, from the sclero-myotome. This is, therefore, a 



40-1 s. HAÏTA : 

step ill the diÖeren dation of the mesoblastic somite, and because 
the distal (ventral) portion of the hitter happens for a time to be 
continuous witli the lateral plate, we are not justified in concluding 
that it is derived from the lateral plate, which, as we know, never 
undergoes seo-mentation. 

It is a significant fact that in Selachia and Amniota, in 
which the pronephros does not function as the actual excretory 
organ, this separation of the mesobhist into the sclero-myotome 
and the nephrotome is not effected so earl}^ as in the above 
groups, but takes place only at later stages, with the first dif- 
ferentiation of the mesonephros. This consideration makes it 
reasonable to conclude that the early separation of the mesoblastic 
somite into the proximal and distal portions is caused by 
physiological necessity and has no morphological significance^^ 

The case of Lacerla ((gilis is very instrnctive. According to 
Hoffmann" ('80), the Anlagen of the pronephros in this animal 
are, in the most anterior segment, cut off from the myotome 
(sclero-myotome) and remain connected with the lateral plate 
just as in Petromyzon, Teleostei, and Amphibia ; whilst in all the 
following portion, they are the actual diverticula formed segment- 
ally in the parietal layer of the lower part of the somite, as in 
other Keptilia (pp. 2G4 and I'oö). We thus see the two modes 
of separation in one and the same animal. 

All recent authors airree in thinkinir that the Anlage of the 
pronephros is expressed in itself segmen tally and is strictly 
myomeric. Now the question arises : How many parts are to be 
distinguished in the mesoblast, and to what part of it does the 
Anlage of the pronephros belong? 

Van Wyhe ('89) has distinguished, in Selachia, three por- 

l)Tliis view is grounded upon the suggestion of 1'eof. Mixsukuri. 



MORPHOLOGY OF CYCLOSTOMATA. 405 

tions of the mesoblast which are called by him the " Epimer," 
" Mesomer," aiul " Hypomer " respectively. The epimere of 
VAN Wyhe corresponds solely to the myotome ; his mesomere 
comprises the Anlage of the mesonephros and the sclerotome ; 
and the hypomere consists of the Anlage of the pronephros, 
the genital gland, and the lateral plate. The epimere, mesomere, 
and the dorsal part of tlie hypomere undergo the metameric 
segmentation, while the remaining portion of the hypomere 
remains unsegmentcd. According to van Wyhe, the dorsal seg- 
mented part of the hypomere is, therefore, the 2:)roduct of the 
lateral plate (see p. 389). It seems to me that this division of 
the mesoblast does not agree with the facts observed in Fetromyzon 
and the other Anamnia above referred to ; for the mesoblast in 
these groups consists, in early stages, of two portions : (1) the 
segmented, and (2) the unsegmentcd, and of nothing more, just as 
RÜCKERT ('88) and Eabl ('88, '96) have remarked. According to 
RiJCKERT and Rabl, the segmented portion — the somite — com- 
prises the myotome and the sclerotome ; the pronephros and the 
mesonephros are derived from its ventral (distal) portion, which is 
called by Eückert the '' Nephrotom " ('88, p. 272). 

In Petromyzon, these two portions of the mesoblast, the 
segmented and the unsegmentcd, are, in early stages, clearly dis- 
tinguished, being histologically different (see p. 31Ö). The meso- 
blast in such an undifferentiated state is almost entirely occupied by 
the segmented portion, while the unsegmentcd jiortion is very 
small, being represented by the loose tissue of a few cells. Such a 
mesoblastic segment exactly corresponds to the somite of Buckert 
and Rabl. The proximal half of the segmented portion coincides 
with the sclero-myotome of those authors. It consists not only 
of the myotome, but also includes the sclerotome. And it is the 



400 s. H ATT A : 

distal lialf of tins segmented portion which folds out in each 
segment to give rise to the Anlage of the pronephric tnbule on 
one hand and to the cœlomic projection on the other, and, 
therefore, corresponds to the " Nephrotom " of Rückert. 

The nephrotome, therefore, constitutes, in both Petromyzon 
and Selachia, precisely the same part of the mesoblast, viz. the 
distal (ventral) portion of the somite, through which the sclero- 
myotome is connected with the lateral plate. 

The above early stage in the differentiation of the mesoblast, 
in Petromyzon corresponds also to the " Ursegment " of ^m- 
pliioxus in Hatschek's sense ('88). By further development of 
it the unsegmented mesoblast is brought into light, and we can 
then distinguish the *' Urwirbel " and the " fSeitenplatte " of 
Hatschek ('88). And the ventral half of the Urwirbel constitutest 
in Petromyzon, the connecting canal between the unsegmented 
cœlomic cavity and the sclero-myotome, that is to say, the 
nephrotome. Let us now examine what part of the Ursegment of 
Ämphioxim represents the nephrotome of the Craniota. 

In his excellent work on '' Die Nierencanälchen des Am- 
phioxns," BovERi ventures to solve this important question. 
After a discussion he comes to the conclusions : 

(1) That the " Gononephrotom " in Craniota must correspond 
to a part of the " Urwirbel " of Hatschek ; 

(2) That the " Gononephrotom " of Craniota is homologous 
with the genital chambers of the adult Amjjhioxus. 

But these chambers *' sind ursprünglich die segmentale 
Yerbindungscanäle zwischen der unsegmentirten Leibeshöhle und 
der Sclero-Myotom gwesen " ('92, p. 493). I can, therefore, 
ascribe no other significance to the ventral half of the segmented 



MORPHOLOGY OF CYCLOSTOMATA. 407 

mesoblast in Pelromyzon than that it is the morj3hological 
equivalent of the " segmentale Yerbindungscanäle." 

It thus follows that the distal half of the segmented mesoblast 
in Petromyzon undergoes exactly the same fate as that in Am- 
phioxus : it is transformed into the pronephros and the cœlomic 
projection or the " dorsal segmental cœlome," which latter gives 
rise, just as Boyeri suggests in Amphioxus, to the mesonephric 
tnl)ules and, in the hinder region, to the genital gland. 

As has been pointed out in the historical revie\Y, the rela- 
tion of the Anlage of the pronephric tubule to the mesoblastic 
somite is the same for Teleostei and Amphibia, as in Petromyzon. 

It iiiny, tlicrcforCy safely he stated, that the segmented portion 
of the mesohlad constitutes in these groups a single integral 
structure until the separation of the nephrotome in cwitinuo with 
the lateral plate from the sclero-myotome. This separation is, as 
ahove stated, not the separation of the somite from the lateral plate , 
hut the differentiation of the somite into the .^clero-myotome and 
the nephrotome, preparatory to the development of the urogenital 
system. The reason ichy the separation take>< plaee earlier in some 
groups than in others, rests only on physiological grounds. 

B. — The Whole System of the Pronejihros of Cyclostomata, 

Teleostei, and Amphihia is Homologous with the 

Nierencan'dlchen of Ampthioxus (Boveri) and 

not perfectly Homologous loith the Selachian 

Pronephric System. 

I have already stated above (pp. 386 and 387) that Boveri 
has brought the pronephric system of Craniota in harmony with 
the system of the " Nierencanälchen " of Amphioxu-^, basing his 
arguments on the structure, the position, the myomeric arrange- 



408 s. HATTA : 

ment, the physiological function, and the relation of the vascular 
system to the organ. His comparison is, however, almost entirely 
limited to Selachia on the side of Craniota, owing perhaps to the 
scantiness of the literature at that time. I accept in the main 
this homology, and I may perhaps extend this comparison a little 
further. 

I will begin with the homology of the pronephros of Cyclo- 
stomata with the " Nierencanälchen " of Amphioxm. 

It is well known that the starting point of the hepatic 
diverticulum from the enteric canal demarcates, in the Chordata, 
the respiratory section of the canal from tlie nutritious section of 
it ; and, as Gegenraue ('78, pp. 563 — 581), Balfoue^^ {'^b), and 
others affirm, the œsophagus and stomach in the higher forms are a 
part of the former section, which is called the fore-gut. And the 
homology of the hepatic cœcum of Amphioxus with the liver of 
the Craniota, lias been much strengthened by recent mor- 
phological studies and physiological experiments-^ The results 
of my present study also confirm this view. I will use, therefore, 
this fixed point as the landmark of comparison of the two 
organ- systems, the pronephros and the " Xierencanälchen," and of 
the pronephros in different groups of Craniota. 

l)Froin the account of Balfouk, I will cite the following lines:— 

' Jn Amphioxus tlie respiratory region extends close np to the opening of the hepatic 
diverticulum, and therefore to a position corresponding with the conimencenient of tlie 
intestine in higher types. In the craniate Vertebrata the number of the visceral clelts has 
become reduced, but from the extension of the visceral clefts in Amphioxus, combined with 
the fact that in the higher Vertebrata the vagus nerve, which is essentially the nerve of the 
branchial pouches, supplies, in addition tlie walls of the oesophagus and stomach, it may 
reasonably be concluded, as has been pointed out by Gegenbaur, that the true respiratory region 
primitively included the region which in the higher types forms tlie oesophagus and 
stomach' (Vol. ii, p. 7nS). 

Balfour has also shown that the solid cord of the œsophagus in Elasmobrancliii and 
Teleostei, is the remanent of the gill-rudiments in tlie ancestry {loco cit., pp. CI and 78). 

2) J. A. Hammar, '99, '98, and Guido Schneider, '99, 



MORPHOLOGY OF CYCLOSÏOMATA. 40*J 

The " Nierencanälclien " of Amphioxu><, according to Boveki 
('92), extends over and is limited to, the whole extent of the 
branchial region, the posterior larger part of which covers the 
hepatic cœcum. The pronephros of Cyclostomata extends 
from the anterior body-somite to the cloaca. The anterior section 
of the system constitutes afterwards the glandular part represented 
by the pronephric tubules and is found in front of, and over, 
the Anlage of the liver, or in the region of the fore-gut ; a certain 
number (two in Fetromijzon, twenty in Bdellostonia) of the anterior 
nephric segments are found in the branchial region, and the 
posterior one or two segments of the glandular part {Petromyzon) 
cover the liver-Anlage. It follows that the six^^ to twenty 
or more pronephric tubules correspond to as many " Nieren- 
canälchen " in about the middle one tliird'-^ of the branchial resiun 
0Î AmpJiioxu><, and that the " Nierencauiilchen " lying posterior to 
this point are represented, in Cyclostomata, by a number of the 
rudimentary tubules which are converted into the segmental duct. 

The '' Nierencanälchen " are not put in communication with 
one another by the collecting duct, as in the pronephros of Cyclo- 
stomata, but open to the exterior segmeutally. I have stated in 
the descriptive part (p. 333) that the free extremities of the 
pronephric tubules in Pciroiiiyzon are brought into close contact 
with the epiblast, so that the latter is pressed out by the enor- 
mous growth of the tubule and that this is especially the case in 
the first and second tubules. This fact throws light upon the 
homology of the pronephric tubules in Pelroniyzon with the 
" Nierencanälchen " of Ainphioxus : in other words, the condition 

l)TIie glandular part nf the pronephros in Fetromyzon, are represented by the six 
pronephric tubules. 

2) The branehiomeres in the posterior section of the gill-basket of Amphioxus are after- 
wards added (see pp. 41U-4I1). 



410 s. Il ATTA : 

seen in the '* Nieren canal clien " would be brought about, if tlie 
tubules in Pviroinyzon came to open to the exterior, boring 
through the epiblast by the further growth of their free extremity. 
This intimate contact of the tubule-end with the epiblast takes 
place, as above mentioned, in the middle of Stage it, where 
the tubules have developed a little beyond the mere Anlnge. 

The pronephric tubules of Feironiyzon are, in this stage, 
already united with one another by the intersomitic solid cord ; 
this union is, however, not primary, but secondary. This stage 
presents, I think, the phylogenetic stage, in which the " Niereu- 
canälchen" with separate external segmental openings, and the pro- 
nephric tubules with the collecting duct, diverge from each other. 

By this assumption, it is not meant that in the ancestry of 
Chordata the tubules were closed blindly inside the epiblast ; for 
the Anlage of the pronephric tubule might have been, in the 
ancestral form too, brought about by the folding of the mesoblast, 
to break out finally to the exterior. This perforation would 
become unnecessary when the secondary union of the tubules had 
been acquired. 

Since a certain number of the '"' Nierencaniilchen " in front 
of the base of the hepatic cu:;cum, is represented by the pronephric 
tubules of the glandular part in Cyclostomata, those lying over 
it will be homologous with the pronephric tubules which are 
found over and posterior to the hepato-pnncreatic Anlnge and 
converted into the anterior section of the segmental duct, being 
secondarily united with one another by the confluence of the free 
extremities of the tubules. 

There is not to be seen tlie post-hepatic " Nierencanälchen " 
in Ainqyhioxus. We learn from Laxkestee ('89) and Willey ('91) 
that in Amphioxus, the new branchial slits are added, by stages, 



MORPHOLOGY OF CYCLOSTOMAÏA. 411 

to the posterior end of the phaiynx, so that, in hiter sta^^es, the 
coincidence of the number of the slits with that of the myotomes 
is lost ; and that this addition of the slits continues throughout 
life. The nephrotomes in these new slits have been, I think, 
originally coincident, in each segment, with the m3^otomes, from 
"which the}' were cut off in early stages and have remained 
undeveloped until the new appearance of the added slits. It 
seems, therefore, probable that the branchial region of AinphloxiDi 
once extended over the largest portion of the enteric canal, 
while a very small section in the posterior part of the canal 
performed the nutritious function, as is seen now in the Ascidian''. 
The " Nierencanälchen " in this hinder part may represent the 
pronephric tubules in the post-hepatic section of the segmental 
duct of Cyclostomata'-'. 

The pronephric system of Fctrcniiyzon comes to have the same 
relations with the epiblast as the " Nierencanälchen " oî Aniphioxus 
at three different points : the free ends of the two anterior 
pronephric tubules and the hind end of the segmental duct 
(probably the hindmost pronephric tubule). Whilst in the greatest 
section of the system the comnnmication with the exterior has 
been lost, these three points might have preserved it to a 
considerably later phylogenetic stage : the two anterior tubules 
playing the same ph^^siological part as the " Xierencanälchen " 
of Amphioxus, and the posterior being employed as the only 

excretory pore of the system secondarily established by the union 
of the tubules. 

In main points (with exception of the presence of the 
tubules in the branchial region, of the contact or connection of 



])BalJbiu" says: "In Ascidiaus the respiratory sack is homologous with tlie respiratory 
tract of Amphioxus"' {loco cit., p. 758.) 
2) See p. 108. 



412 s. H ATTA : 

tlieni with the epibhist, &c.), the proiiephric system of Teleostei 
and Amphibia shows, as stated in the liistorical review, the same 
characters as that of Cyclostomata, so that the facts established 
in Cyclostomata have the same significance for the Teleostei and 
Amphibia. 

Although such is the case in those Craniota and Aiiiphio.uis, 
the pronephros of Selacliia is quite otherwise : the Anlagen of 
the pronephros are here formed in tlie mesoblastic somites posterior 
to the Anlage of the liver (see below), and only one or two 
segments of them are converted into the segmental duct (Rïjckert). 
These pronephvic Anlagen in Selachia are, therefore, the mor- 
phological equivalent, not of the glandular portion of the pronepliros, 
but of those whieh are converted into the segmental duet in the 
Craniota just mentioned. 



0. — The Segmental Duct in Selacliia is not the Blorphologieal 

Equirale)it of the Dud of the Same Name in C>/clostomafa, 

Teleostei, and Amphibia. 

Contradictory views arc met with in the derivation of the 
segmental duct. The results arrived at in Cyclostomota, Teleostei, 
and Amphibia, well agree in making it of the mesoblastic origin ; 
there are a few authors who Ijelieve in the epiblastic origin of the 
duct in these groups, but their papers are not more than mere 
notes. In Selachia, the circumstance is reversed ; I am not aware 
of any recent author other than Kabl, who advocates the meso- 
blastic origin of the Selachian segmental duct. The facts given by 
Rabl are, however, not the same as those observed in the 
groups just referred to. In these, as stated above, the duct is 



MORPHOLOGY OF CYCLOSTOMATA. 413 

clifferentiated, so to speak, in sUu from the mesoblast in its whole 
length, and as recent anthors agree, is composed of a series of 
the abortive tubules formed in each nephrotome. This is not 
the case in Sehichia ; here it is brought about, as Rabl states, by 
the posterior growth of the collecting duct wdiicli is formed by 
the confluence of the lateral extremities of the pronephric Anlagen. 
It is not easy to bring these tw^o ^videly divergent modes of 
formation into harmony with each other. 

A few morphological considerations, however, would, I believe, 
enable one to derive one type of the system from the other. I 
may be permitted to state here some of these considerations. 

I will start with the question : Is the segmental duct of 
Selachia the morphological equivalent of that of Peirom.yzon, 
Teleostei, and Amphibia ? I believe the question can be answered 
safely in the negative, if we consider (1) the position of the 
pronephros first formed, and (2) the origin of the duct. 

In the first place, the pronephros in Selachia appears, as 
we learn fiom Rabl, in the mesoblastic somites lying 'posterior 
to the Anlage of the liver ; thus the Anlage of the liver lies 
under the fourth and fifth somites, and that of the pancreas 
under the sixth, while the pronephros covers the seventh to 
tenth somites ('96, p. 667). On the contrary, in Petromyzon^^ the 
pronephros originates in the mesoblastic somites anterior to the 
hepato-pancreatic Anlage, only the posterior one or two nephro- 
tomes covering the liver. Such being the case, the pronephric 
segments in Selachia correspond to the same number of the 
abortive tubules in Petromyzon and the other Craniota above 



1) As we learn from Goette ('75) and Oellaclier ('78), the anterior section of the 
pronephric system in Amphibia and Teleostei, is also found in the mesoblast opposite to the 
posterior section of the fore-gut. 



414 s. H ATTA : 

mentioned, wliicli are converted into the anterior section of the 
segmental duct in these groups. It follows that the segmental 
duct in this part of Peiromyzon and the two other groups, is 
not the morphological equivalent of the duct of the same name, 
but of the pronephros itself, in Selachia. 

In the second place, let us consider the mode of growth 
of the segmental duct. Whichever view may he taken of its 
origin, whether epiblastic or mesohlastic, this duct in Selachia 
does not arise segmentally as in the otlier Craniota just referred 
to. It is a backward growth produced either by delamination 
from the epiblast, as Ruckert and van Wyhe affirm, or by 
cell-multiplication within the structure of the mesohlastic collect- 
ing duct itself, as Kabl states. Hence it can not be homologous 
with the duct of the same name in Petroiiiyzon and the two 
other groups, which is derived segmentally from the rudimentnry 
pronephric tubules. The Selachian segmental duct is, in its 
whole length, represented, as I believe, by the posterior small 
section of the segmental duct in Peiromyzon and Amphibia. 

In Peiromyzon, the hind end of the duct comes into an 
intimate connection with both the epiblast and the lateral diver- 
ticula of the cloaca, filling up the space between them, and fusing 
with both of them. We may suppose that the direct communication 
of the duct with the exterior, if such truly existed in the an- 
cestral history, may have been at this point of the epiblast. 
This fused condition of the duct and the epiblast reminds us of 
the early stages of the Selachian duct at the stage when it has 
been produced only a little posteriorly from the pronephric region. 
I believe that if the duct is to be compared in Peiromyzon and 
Selachia, a stage such as the above ought to be taken. The 
largest part of the Selachian duct is represented by a free hind- 



MOEPIIOLOGY OF CYCLOSTOMAÏA. 410 

ward growtli formed after such a condition is passed, and its 
homologue can not be found anywhere in Pctromyzon. I believe 
that the same can be stated of Amphibia which is, to judge from 
Field's account, very much like Petromyzon in this respect 
(see p. 401). 

According to tax AVyhe and Rabl, the duct in Selachia 
appears in the seventh to the tenth (in Rabl's sense) somites of 
embryos with 34-3Ö somites and, when it is later connected with the 
cloacal wall, the connection is found, — as can be inferred from 
fig. 7/; of van Wy'HE, — in the thirty-eighth (forty-second of Rabl) 
Rumpfsegment (or further backwards) of Fristiurus embryos with 
80 (van Wy'he) to 87 (Rarl) mesoblastic somites. There is found 
in f^elachia, therefore, a number of the mesoblastic somites in 
the region back of the pronephros, which do not give rise either 
to the pronephric tubules or to the segmental duct, and the duct 
grows backwards, free from the mesoblast inside, during the 
period in which the somites increase from 34-35 to 80-87, and 
for the space reaching from the eleventh or twelfth to the forty- 
second (in the sense of Rabl) somite. Such a considerable 
prolongation of the duct during this period is not observed in 
Petromyzon^* and in the two other groups of Craniota mentioned. 

And furthermore, it is questionable whether, during this 
posterior growth, the duct in Selachia receives the constituent cells 
from the epiblast along its whole length, as Rückert and tax 
Wyhe believe ; or only at the point of the epiblast overlying the 
hind end of the pronephros, with which the duct is connected, 
and posteriorly to this point grows free from both the epiblast 



1) At about this stage (Stage iv), there is no space left behind the pronephric system 
segmentally formed; for the embryo ot Petromyzon, is retort-shaped and lias the anus situated 
in the ventral median line of tlie bulb of the retort. 



41G s. HAïïA : 

and the mesoblast. The latter view seems probable to me. The 
figures ('80, figs. 5 a-c) given by van Wyhe to illustrate his view 
of the epiblastic origin of the duct, are from the vertical j^lane 
of the eighth Eumpfsegment of a Scyllium embryo with 37 
somites, which corresponds to the twelfth Gesammtsegment of 
Kabl. The figures ('96, figs. 9a, Ob, 10a, and 10b) given by Rabl 
to the negation of van Wyiie's view, a]-e from the vertical plane 
of the twenty-second Gesammtsegment of an embryo with the 
()3 somites. These two cases are, I suppose, the two ends of the 
same duct in different stages ; the anterior end, being the 
equivalent of the hind end of the segmental duct in Pdromyzon, 
actually receives cells out of the epiblast, as the figures by \'ax 
Wyhe show ; the other end, which is seen in Rabl's figures, is 
the point of mere contact with the epiblast, along which it is 
shifting backwards. 

If the above comparimn he eorrect, the segmental duct in 
Selaehia is, except the anterior very ><ni all section xvlùcli is formed 
directly of the abortive tulnilex, not hoiiioloyous with the duet of 
the same name in Petromyzon, but is a dructure secondarily acquired. 

From the above account, it may be safely concluded that in 
its primary phylogenetic stage, the pronephric system of the 
Craniota above referred to consisted of a number of segmentally 
arranged tubules, which were directly formed, in each mesoblastic 
segment, from the distal (ventral) portion of the mesoblastic 
somite, and opened independently to the exterior ; that the lateral 
extremities of these tubules were afterwards secondarily united 
with one another, thus constructing the collecting and segmenta^ 
duct, the hind end of which opened directly to the exterior ; and 
tliat the acquisition of an opening of the duct into tlie cloaca 



MORPHOLOGY OF CYCL0^JT03IATA. 417 

was the tertiary stage of changes in the system. Such a course 
of the phylogenetic development of tlie system is, Jiowever, no 
other than that advanced by Kückeiit {'SS, p. 200). 



In the present paper, the historical comparison will he 
limited to the groups of Yertebrata stated above ; the review of 
Amuiota and some other theoretical considerations will be reserved 
to a future paper, in which I propose to deal v\'ith the further 
fate of the pronephros and the development of the mesonephros 
in Petromyzon. 

Having compared the results arrived at in the present work, 
with those in difterent classes of Anamnia, I may be justified 
in drawing the following conclusions. 

In Petromyzoii, the first indications of the pronephros be- 
comes apparent at a stage earlier than those hitherto regarded 
as the starting point, that is, at a stage in which the mesoblast 
in the anterior rcL'ion has undergone the metameric segmentation 
but the lateral plate is not yet cut ofi' from the somite. 

The tissue giving rise to the pronephros is the parietal layer 
of a small section of the mesoblast, which forms the distal 
(ventral) half of the mesoblastic somite. This section of the 
mesoblast exactly corresponds to the "Nephrotom" of Kuckert 
in Selachia. 

The Anlage of the pronephros in all the groups of A^ertebrata 
above referred to is produced by the evagination of the parietal 
layer of the nephrotome which theoretically ought to contain a jwrt 
of the cœlomic cavity. Tn Cyclostomata, such a cavity is 



418 s. HATTA : 

actually preseiii^^* in other groups, the Anlagen are mere thicken- 
ings. 

As the pronephric tubules are derived, in each segment, from 
the distal (ventral) half of each mesoblastic somite, the prone- 
phros is, from the first, of a segmental arrangement, being 
strictly myomeric. In fact, the separation of the sclero-myotome 
from the lateral plate is eöected on account of the differentiation 
of the Anlage of the pronephros or of the nephrotome. 

In Petroiiiyzon, the pronephric tubules which constitute the 
glandular part of the system and the anterior section of the 
segmental duct, are formed in the region of the fore-gut and some 
of them are detected in the region where the gill-pouches are 
afterwards formed ; these latter disappear entirely before the gills 
come into view. 

The segmental Aidagen of the pronephric tubules are 
secondarily connected Ijy the duct formed out of tw^o adjacent 
pronephric Anlagen and put in connnunication with one another. 

Tlie degeneration of the pronephric tubules takes place from 
both the cranial and caudal extremities of the system. In the 
cranial part, the tubules disap]3ear without leaving any trace ; 
while in the caudal, they are converted into the anterior section 
of the segmental ikict. The remaining part of the system func- 
tions for some time as the excretory organ. 

The pronephric Anlagen in the hinder region do not develop 
beyond a certain point, but are em]3loyed solely to give rise to 
the segmental duct just as in the somites having degenerated tubules. 

From whit has been said, it is, I venture to think, no rash 
conclusion to regard the pj-onephric iuhules in Peivomyzon as having 
once extended over the hody -segments from the branchial region 

]) According to ISwAtx and Buaciikt, tlic same fact is seen in Teleuslei. 



MORPHOLOGY OF CYCL0ST03IATA. 419 

to the chacal part, and as having been, in the anterior rer/ion, 
replaced by gilh and, in the posterior, converted into the .segmental 
duct. 

Ill the two anterior segments which l^elong to the hranchial 
region, the free ends of the tubules are brought into close contact 
with theepiblast but this germinal layer has, in this region, no 
share in the formation of the system. The hind extremity of the 
segmental duct, however, strikes against the epiblast and has every 
appearance of receiving some cells out of it. These facts allow 
us to infer that all the tubules once had each an independent 
external opening until they were secondarily united with one 
another by the intcrso initie duet. 

The visceral layer of the nephrotome becomes evaginated 
medianwards and forms a series of segmental pouches on either 
side of the subchorda ; but this feature is temporary, and the struc- 
ture is soon smoothed by their Ijecoming confluent with one another. 
This series of pouches is, I l^elieve, the remnant of the primitive 
segmental cœlonie, and gives rise to the gonads and the meso- 
uephros. 

If the accounts given above be correct, the primary rnesoblast 
is, during early development, divided into two distinct portions : («) 
the larger proximal portion ivhich is segmented, and {b) a ><mall 
distcd jyortion which is unsegmented. The former is differentiated 
into the sclero-myotome and the nephrotome, and the latter forms 
simply the jjeritonecd linings. 

The pronephric vessels acquire their définitive form in much 
later stages ; when established, they are intersomitic in position. 
The posterior part is transformed into a pair of the glomeruli 
of the pronephros. 

Feiromy.:o/i has for a long time been looked upon as being 



420 s. HATTA : 

peculiar and standing apart from other Vertebrata in the develop- 
ment of the pronephros. But the results brought out in the 
present work speak for a complete parallelism between this genus 
and the representatives <jf other classes of Vertebrata. 

Biological Laboratory, 

The College of Peers, Tokyo. 

November, 1899. 



Postscript. 



By the kindness of Peof. Watase, I have been enabled to 
look through Dk. Wheeler's paper on " The Development of the 
Urogenital Organs of the Lamprey "^' which has just been published. 
I lind a general agreement of his results with mine. The most im- 
portant point is the discovery of the earliest traces of the pronephros 
as given in the foregoing pages. As to the formation of the 
segmental duct, his views are somewhat different from mine ; this 
and some other points of divergence are, as I believe, due to 
gaps in his materials. Thus, his fig. 1, wdiich represents section 
through an embryo in his Stage 1, corresponds to my fig. 1, while 
the next older stages (Stages 2 and 3) in his series, spoken of as 
representing Goetïe's fig. 1) where the heart is already formed, 
coincide with the oldest embryo of my Stage iv. As seen in the 
foregoing description, most of the important processes in the 



l)Zool. Juhrbüclier, ALtlieil. für Anat. and Ontog., Bd. xiii, '99. 



MOKPIIOLOGY OF CYCLOSÏOMAÏA. 421 

development of the pronephros and the segmental duct take place 
in this interval of time, ^vhich AVheeler has unfortunately 
omitted to study. But in the main, his results confirm mine. 
This agreement arrived at independently naturally a fiords a good 
evidence of the correctness of the facts lilven. 



422 s. IIATTA 



Literature. 



('79) Balfour : — A Monograph on the Development of Elasraobranch 

Fishes. 

^'35) : — A Treatise on Comparative Embryology. 

('87) Beard: — On the Origin of the Segmental Dnct in Elasmobranchs : 

Anat. Anz., Bd, ii. 
('92) BovERi: — Die Nierencaniilchen des Amphioxns : Zool. Jahrb., Bd. v. 
('88) Brook:— Note on the Epiblastic Origin of the Segmental Duct in 

Teleostean Fishes and in Birds : Proceedings of the Royal Society of 

Edinburgh, Vol. xiv. 
('98) Dean: — On the Development of the Californiau Hagfish (Bdellostoma 

stouti Lockington) : Quart. Journ. Micr. Sc, N. S., Vol. xl. 
('99) : — On the Embryology of Bdellostoma stonti : Festschr. z. 

siebe nzigsten Geburtst. v. C. v, KnpfFer. 
('82-'90) DoHRN : — Studien zur Urgeschichte der Wirbelthirki^-pers : Mittheil. 

a. d. Zool. Stat. z. Neapel. 
('91) Field : — The Development of the Pronephros and Segmental Duct in 

Amphibia: — Built. Mus. Comp. Zool. Harv. Coll., Vol. xxi. 
CSG) Flkmming : — Die ektoblastische Anlage des Urogenitalsystems bei 

Kaninchen : Arch. f. Anat. n. Physiol., Anat. Abtheil. 
('97) Felix : — Die PRiCE'sche Arbeit ,, Development of the Excretory 

Organs of a jMixynoid (Bdellostoma stoutij'' und ihre Bedeutung 

ffir die Lehre von der Entwickehmg des Harnsystems : Anat. Anz., 

Vol. xin. 
('78) Fürbringer: — Zur Vergleichenden Anatomie und Entwicklungs- 
geschichte der Excretonisorgane der Vertebraten : Morph. Jahrb., 

Bd. IV. 
('78) Gegenbaur : — Grundriss der Vergleichenden Anatomie, 2. Auflage. 
('75) Goette : — Die Entwicklungsgeschichte der Unke, Leipzig. 

(^'88) :— Ueber die Entwicklung von P. planeri : Zool. Anz., Jahrg. vi. 

(^'90) : — Entwicklungsgeschichte des Flussneunauges (P. fluviatilis), 

Hanburg und Lei[)zig. 
('97) Gregory: — Origin of the Pronephric Duct in Selachians: Zool. Bullet. 

Vol. i. 
('81) Hatschkk: — Studien tÜ)er Entwickehmg des Amphioxns: Arb. a. d. 

Zool. Inst. d. Univ. Wien., Tom. iv. 



MORPHOLOGY OF CYCLOSTOMATA. 4zO 

('98) Hammar : — Zur Kenntniss der Leberentwickeliiiis; bei Amphioxus: 

An at. Anz., Bel. xv. 
('97) : — Ueber einige Hanptzüge der ersten embryonalen Leberent- 
wickelnno;: Id., Bd. XIV. 
('88) Hatsohek : — Ueber den Schichtenban von Amphioxns : Id., Bd. in. 
('91) H ATTA : — (3n the Formation of tlie Germinal Layers in Petromyzon : 

Jour. Coll. Sc. Imp. Univ., Jap. Yol. v. 
('97) :— Contributions to the Morphology of Cyclostomata. I. On 

the Formation of the Heart in Petromyzon: Id., Vol. x. 
('97) • : — Preliminary Note on the Development of the Pronephros in 

Petromyzon: Annot. Zoul. Jap., Vol. i. 
('90) Heri'wig : —Lehrbuch der Entwicklungsgeschichte des Menschen und 

der Wirbelthiere, 3. Auflage. 
('86) Hoffmann: — Zur Entwicklungsgeschichte der Urogenitalorgane bei den 

Anamnien: Zeitschrift f. wiss. Zool., Bd. xuv. 
(^'89) : — Zur Entwicklungsgeschichte der LTrogenitalorgane bei den 

Reptilien: Id., Bd. xlviti. 
('88) V. K[TFFEu: — Lieber die Entwickelung der Neunaugen: SB. d. k. bayer. 

Akad. AYis. zu München, Bd. I. (Ptef. in Jahresb. Fortschr. Auat. 

Phys., '89). 
('90) : — üie Entwickelung von P, planeri : Arch. f. mikrosk. Anat., 

Bd. XXV. 
('89) Lankrster:— Contributions to the Knowledge of Amjihioxus lanceolatus 

Yarrell : Quart. Journ, Micr. Science, N. S., Vol. xxix. 
('90) Lankester & WiLLEY : — The Development of the Atrial Chamber of 

Amphioxus : Id., xxxi. 

('97) Ma ASS : — Ueber die Entwickkingsstadien der Vorniere und Urniere bei 
Myxine: Zool. Jahrb., Bd. x. 

('88) Martin : — Uelx;r die Anlage der Urniere beim Kaninchen : Arch, f. 
Anatomie and Entwicklungsgeschichte. 

('87) Mater, (P.): — Ueber die Entwickelung des Herzens und der grossen 
Gefässstämme bei Selachiern: IMittheil, Zool. Stat. Neapel, Bd. vir. 

('88) MiTSUKURi :— The Ectoblastic Origin of the Wolffian Duct in Cheolonia 
(Prelim. Notice): Zool. Anz. Jahrg. xi. 

('90) MoLLiER : — LTeber die Entstehung des Urogenitalsystems bei Amphi- 
bien: Arch. Anat. Phys., Anat. Abth., Bd. in. 

('75) Müller (AV.) : — Das Urogenitalsystem des Amphioxus und Cyclostomen: 
Jen. Zeitschr. Naturw., Bd. ix. 



424 s. HATTA : 

('72 & 73) Oeij.acher : — Beiträge zur Entwicklungsgeschichte der Knochen- 
fische : Zeitsch. f. wis. Zool, Bd. xxii n. Bd. xxiir. 
('70) OwsjANNiSKOw : — Die Entwicklungsgeschichte der Flussneunaugen : 

Melanges Biologie, Tom. vii. 
('87) V. Perenji: — Die ectoblastische Anlage des Urogenitalsysteras bei Rana 

escnlenta und Lacerta viridis : Zool. Anz., Jahrg. x. 
('97) Price : — Development of the Excretory Organs of a Mvxinoid (Bdello- 

storaa stonti Lockington): Zool.^Tahrb. Bd. x. 
('88) Rabl: — Ueber die Differenzirung des Mesoderms: Anat. Anz., Bd. in. 

('87) : — Theorie des Mesoderms: Morph. Jahrb., Rd. xv. 

(^'9G) : — Ueber die Entwickelung des Urogenitalsystems der Sela- 

chier (zweite Fortsetzung der " Theorie des Mesoderms "): Id., Bd. xxiv. 
('88) PtücKERT : — Ueber die Entstehung der Excretionsorgane bei Selachiern: 

Arch. Anat. u. Phys., Anat. Abth. 
('89) : — Zur Entwickelung des Excretionssystems der Selachier; eine 

Erwiederung an Herrn van Wyhe: Zool. Anz., Jahrg. xii. 
('82) Ryder : — A Contribution to the Embryography of Osseous Fishes with 

Special Reference to the Development of the Cod : Report U. S. 

Fishcommission. 
('99) Schneider, (Guido) : — Einiges über Resorption und Excretion bei Am- 

phioxus lanceolatus : Anat. Auz., Bd. xvi. 
C^)()) ScHULTZE, (Max) : — Die Entwicklungsgeschichte von Petromyzon planeri, 

Haarlem, 
('82) Scott : — Beiträge zur Entwicklungsgeschichte der Petromyzonten. 

Morph. Jahrb., Bd. vii. 
('81) Sedgwick : — On the Early Development of the Anterior Part of the 

Wolffian Duct and Body in the Chick : Quart. Journ. Micr. Sc, N.S., 

Vol. XXI. 
('9G) Semon : — Das Excretionssystem der Myxinoiden in seinen Bedeutung 

für die raorphologishe Auffassung (Ref. Jahresber ü. d. Fortsch. d. 

Anat. u. Phys., N. F., Bd. ii). 

('97) :— Das Excretionssystem der Myxinoiden : Anat. Anz,, Bd. viii. 

('97) : — Vorniere und Urniere, Id., Bd. xiii. 

('75) Semper: — Das Urogenitalsystem der Plagiostomen und seine Bedeu- 
tung für das der übrigen Wirbelthiere : Arbeiten a. d. zool.-zootom. 
Inst, in AVürzburg, Bd. ii. 

('87) Shipley: — On Some Points in the Development of Petromyzon 
fluviatilis : Quart. Journ. Micr. Sc, N.S., Vol. xxyit. 



3I0IIPH0L0GY OF CYCLOSTOMATA. 425 

('95) tSüBüTA : — Die Eatwickelung der Voruiere der Salmoniden : x4.nat. 

Anz., Bd. VI. 
('84) Spee, (Guaf): — Ueber directe Betbeiligung des Ektoderms an der Bildung 

der Urnierenanlage des Meerschweinchens : Arch. f. Anat. und Phys., 

Anat. Abth. 
('99) SwAEN et Brächet : — Entiide sur les premières phases du development 

des organs dérivés du mésoblaste chez les poissons téléostéeus : Arch. 

de Biologie, Tom. xvi. 
('90) Weiss: — Excretory Tubules in Amphioxus lanceolatus : Quart. Journ. 

Micr. Sc, N. Ö., Vol. xxx. 
('84) Weldün: — On the Head-kidney of Bdellostoma, with a Suggestion as 

to the Origin of the Suprarenal Bodies : Quart. Journ. ]\Iicr. Sc, N.S., 

Vol. XXIV, 
('83) Wiedershedi : — Lehrbuch der vergleichenden Anatomie der Wirbethiere. 
('91) Willey: — The Later Larval Development of Amphioxus: Quart. Jour. 

Micr. Sc, N.S., Vol. xxxi. 
('94) :— Amphioxus and the Ancestry of Vertebrates, New York and 

London. 
ÇSB) V. Wyhe:— Die Betbeiligung des Ectoderms an der Entwickeluug des 

Vernierenganges : Zool. Anz. Bd. ix. 
('89) :— Ueber der Mesodermsegmente des Rumi)fes und die Ent- 

wickelung des Excretionssystems bei Selachiern : Arcli. f. mikr. Anat., 

Bd. XXXIII. 
('98) :— Ueber die Betheiligung des Ektoderms an der Bildung 

des Vornierengang bei Selachiern: Verband], d. Anat. Gesetsch. 

auf. d. zwölften Vorsamml. in Kiel. 
('88) Ziegler : — Der Ursprung der mesenchymatischen Gewebe bei den 

Selachiern : Arch f. mikr. Anat., Bd. xxxu. 



I'l.ATE XVTI. 



Plate XVII. 



[The magnification is the same for all figures: Cx2, with the single 
exception of fig. 4 which is Ex 2.] 



a.pn.1-6, Anlagen of ])ronephric 
tubules from the first to sixth. 
a.sd., Anlage of segmental duct. 
ed., collecting duct. 
cli., chorda doi'sali.s. 
cut., cutis-layer of myotome. 
d., dorsal row of mesohlast. 
cp., epiblast. 
Idj., hypoblast. 
l.m.^ lateral [)late of mesoblast. 



m., median row of mesoblast. 
m.jj., parietal layer of mesoblast. 
mes., mesoblast. 

mt.I, II, &c., the first, second, &c. 
myotome. 

mus., muscle-layer of myotome. 
m.v., visceral layer of mesoblast. 
n., neural cord or canal. 
v., ventral row of mesoblast. 



Fig. 1. A transverse section through the dorsal region of an embryo inter- 
mediate between ytaü;es i and ii. 

Fig. 2-7. From a series of transverse sections through a younger embryo 
of Stage 11. 

Figs. 8-17. From a series of transverse sections through an older embryo 
of fcîtage II. 

Figs. 18 ar.d 19. Two sections from a series of cross-sections through a little 
more advanced embryo than the last. 

Figs. 20-29. From serial cross-sections througli the most advanced embryo 
of »Stage 11. 



Kv.l. 



Aï-yô. 




l'ùl.ll. 



Fiq.in. 



V- 



Fill. ta. 



,.y. '*^"- 



Fiy.S. 



Jour. Se. Coll. Vol. Xm. PI. XVII. 
Figi. 





Fio.6. 






,;.--^Vn f 




.^■■■' 


.\ — ^-.^__... 





t'ifl. '. 



Fig.S. 









-^ 



Fig-a. 



Fi.l /-'■ 



Pig.lS. 



Fit). 20. 



i 









é?^ ~^~\ 




T'lLATE XVIII. 



Plate XVIII. 






a.jm.1-6, Anlagen of pronephric 
tubules from the first to sixth. 
a.sd., Anlage of segmental dnct. 
ed., collecting duct, 
eh., chorda dorsalis. 
c.p., coelomic projection. 
cut., cutis-layer of myotome. 
fg., fore-gut. 
ep., epiblast. 
hy., hypoblast. 

J.m., lateral plate of mesoblast. 
m.p., parietal layer of mesoblast. 



ms., mesoblast. 

')ni.T, II, &c., the first, second, &:c. 

myotome. 
mvs., muscle-layer of myotome. 
m.v., visceral layer of mesoblast. 

n., neural cord or canal. 

pp.c, pleuroperitoneal cavity. 

pt.1'6, prone[)hric tubules from the 
first to sixth. 

.sc/^, subchorda. 

sd., segmental duct. 



Figs. 30-31. From the same series as figs. 20-29 of the last plate. 

Figs. 32-50. From a series of transverse sections through a younger embryo 

of Stage 11 r. 
Fig-^. 51-58. From a series of transverse sections through an embryo of 

Stage HI. 
Fig. 59. A section through an older embo3'o of Stage iir, the posterior 

continuation of which is shown in the next following plate (figs. 60-63). 



^-i. 

.^:^#"-^^" 



W 



•mmmm 



^^*' 






r.fi.. 



/.f 




Jour. Sc. Coll. Vol. XIII. PI. XVIII. 



!^t£- 



"■ .. IL 








*: ' 




Fiy.Xl 



"■ ^, 



" ^ 



/•vv/..;/. 






)',:„. :;-. 



.^^ 



"äj! 




/■■<«..«. 




1.1 



FliATT] XIX. 



Plate XIX. 



Vil., collectino; dnet. 

cil., chorda dorsal is. 

f'.j)., coelomic projection. 

cul ., cutis-layer of myotome. 

(1., dorsal row of mesoLlast. 

ep., epiblast. 

((J., fore-giit. 

/., Anlage of liver. 

hy., hypoblast. 

Lm., lateral plate of mesoblast. 
■hl., median row of mesoblast. 
mcli., mesenchymatous cells. 
'iiif^., mesoblast. 
m.p., parietal layer of mesoblast. 



nii.I, IT, AT., the first, second, cIt. 

myotome. 
onus., muscle-layer of myotome. 
m.v., visceral layer of mesoblast. 
n., neural canal. 
nst.2-3, nephrosti^me the second 

and third. 
pp.1-3, peritoneal partition. 
pp.c., pleuroperitoneal cavity. 
pt.1-6, pronephric tubules from the 

first to sixth. 
seit., subchorda. 
S(l., segmental duct. 
sg., spinal ganglion. 
V. ventral row of mesoblast. 



Figs. 60-63. From the same series as, and the posterior continuation of, 

fig. .59. 
Figs. 64-76. From a series of transverse sections through an oldest embryo 

of Stage III. 
Figs, 77-81. From a series of transverse sections through an embryo of 

Stage IV ; hence the body of embryo in the |)resent stage is twisted, 

the sections pass through unavoidably ol)lique planes. 



Fù).60. 



eù).6i. 



Fiff.tii. 



Fi,,, i;:;. 



.<iA,\ ë 



Jour.Sc^JColl. Vol. XIII. PI. XIX. 



ri„. <;.<>. 



/A \-w\ " 




C^^l>1 



x^^««« /••'• M 






h'ifß. 7ii. 




Fifl. 7'~. 



f 




h'i<,. r.v. 



Fù/.7,S. 



■^/s:. — ^' C 



^ ,„,.« 



Fiff..S(l. 



..é-'h 




„, * 



/■VV/..VA 



- ï%;. 



Ilr 




^ 1 I •/ 



PLATE XX 



Plate XX. 



«..<?(/,, Anla,2;e of segmental duct. 

bp., blaRtopnre. 

hrg., brauchial region. 

cc, cloacal cavity. 

ed., collecting dnct. 

eh., chorda dor.salis. 

c.dv., div(irticnlnm of cloacal 

cavity. 
co.sd., cloacal opening of seg- 

]n(Mital duct. 
c.p., coolomic projection. 
dl.hp., dorsal lip of blastopore. 
fg., fore-gnt. 
ep., epiblast. 
(jc. genital cells. 
liy., hypoblast. 
ivt., intestine. 
K, Anlage of liver. 



l.m., lateral plate of mesoblast. 
well., mesenchymatons cells. 
ms., mesoblast. 

m.jo., parietal layer of mesoblast. 
rat.T, IT, &c., the first, second, &c. 
myotome. 

m.V'., visceral layer of mesoblast. 

n., nenral canal. 

nd.Ô, fifth nephrostome. 

per it., peritoneal membrane. 

pp.l-S, ]ieritoneal partition. 

pp.c, plenro]ieritoneal cavity. 

pt.l-G, ])ronephric tubules from 
the first to sixth. 

sch., Hul chorda. 
sd., segmental dnct. 

yc, yolk-cells. 



Figs. 82-80. Sections from tlu^ same series as fig. 81, lying posterior 
to it. The section shown in fig. 87 ])asses througii somewhat frontally 
owing to the bending of th(^ body-axis of the embryo ; the neural canal, 
which is Ijent in the sauK^ manner as tlie axis meets with two time 
in section. 

Figs. 90 and i)L. Two sections passing through in the same way as in 
fig. 87 ; in lig, 90 the dorsal li[) of the b1asto])ore, and in fig. 91, the 
ni)p(n' (dorsal) ]iortion of it, is cnt through. 

Figs. 92-9f). Froui a series of transverse sections through a little older 
embryo than the last ; the embryo is twisted in the same way as it. 

Fig. 97. Frontal section through an embryo abont the sam(^ stage as the 
last, the bodv (»f which has been straightened befon^ cut throuiiii. 



.'^^. 




Jour Sc. Coll. Vol. X/H. PI, / ,' 



m^^. 



Fifl./il. 




/•;■ 



I 





•:>-:^';i^v 



"4^ 





IPI.ATE XXI. 



Plate XXI. 



a.cv., anterior cardiiial vein. 

OM., auditory pit. 

J)l)., ventral li]) of blasto[)ore. 

I>rg., brancliial region. 

bs., Llood space. 

ed., collecting duct. 

eh., chorda dorsali«. 

ce, cloacal cavity. 

c.dv., diverticuliun of cloacal 
cavity. 

co.sd., cloacal opening of seg- 
mental duct. 

cîo., wall of cloaca. 

cp., e[)iblast. 

(/L, glomerulus of proueplinjs. 

/'., dorsal tin. 

/(J., fore-gut. 

h., heart. 

hy., hypoblast. 

/., liver, or Anlage of liver. 



Lw., lateral plate of mesoblast. 
mch., mesenchymatous cells. 
7n.p., i)arietal layer of mesoblast. 
perit., peritoneal membrane. 
mt.T, II, &c., the first, second 

myotome, &c. 
m.v., visceral layer of mesoblast. 
n., neural canal. 
nst.2-6, nephrostomes from the 

second to sixth. 
pp.c, pleuroperitoneal cavity. 
pjt. 2-5, pronephric tubules from 

the first to fifth. 
r.m., radix of mesentery. 
sell., subchorda. 
sd., segmental duct. 
t.a, tract of aorta. 
t.ac. tract of anterior cardinal 

vein. 
tr.a. truncus arteriosus. 



Figs. 98-106. From a series of transverse sections through an embryo of 

Stage V. 
Figs. 107-110. From a series of transverse sections through an embryo 

of Stage VI. 
Fig. 111. Transverse section through the cloacal region of an older embryo 

of Stage VI. 
Figs. 112-114. A series of sagittal sections through a younger embryo in 

Stage IV. 
Fig. 115. A. frontal section through an embryc» a little more advanced 

than that of Stage vi. 



A 



'■' \ 




Piff. too. 





Jour. Sc. Coll. Vol. XIII. PI. XXI. 



Fi,,.ior>. 










*-**S?t. 



m^. 



%"" 



Beiträge zur Wachstumsgeschichte der 
Bambusgewächse, 



Von 



K. Shibata, Rigahishi. 
Mit Tai ein XXILXXIV. 

I. Einleitung. 

Unsere Kenntnisse über Bau und Lebensweise der Bambus- 
gewächse waren bisher sehr mangelhaft gewesen, obwohl in neueren 
Zeiten einzelne merkwürdige Erscheinungen auf dem Gebiete der 
Physiologie dieser eigenartigen Baumgräser durch interessante Beo- 
bachtungen^) einiger die Tropen besuchender Botaniker zu Tage 
gefördert wurden. Bekanntlich gehören die meisten Bambuseen 
zu wärmeren Gegenden der alten und neuen Welt, mit Ausnahme 
von einigen kälterem Klima angepassten Formen, wâe z. B. 
Bambusa Kurihmis, die in einer nördlichen Insel Japan's 
bei 46° n. B. gedeiht. Unser Land besitzt eine Reihe von 
Bambusformen, welche unserer Pflanzenphysiognomik ein charac- 



IjG. Kraus, Pliysinlogisclies aus den Tropen. I. Liingenwaclistum der Bambusrohre. 
Ann. d. Jard. Bot. d. Buitenzorg. Vol. XII, p. 196. 

H. Mol isch, Über das Bluten tropischer Ilolzgewächse im Zustand völliger Belaubung. 
Ann. d. Jard. Bot. d. Buitenzorg. 1898. 2 tes Snppl. p. 28. 



428 K. SHIBATA : 

teristisclies Aussehen verleihen. Einige hochwüchsige Formen aus 
der Gattung Phyllostachys sind bei uns überall häufig cultiviert, 
hauptsächlich für die mannigfaltigste Verwendbarkeit der Rohre 
und auch für ihre Frühjahrsschösslinge, die ein beliebtes Gemüse 
darbieten, während andere Arten aus Arundinaria und Bamhusa 
als Zierpflanzen in unseren Gärten gemein sind. 

Nun stellte ich es mir hier als Aufgabe, in erster Linie die 
Natur und das Verhalten der wichtigen Baustofife während der 
verschiedenen Vegetationsperioden zu studieren, da mir besonders 
die ungemein rasche Entwicklung der Schösslinge etwas interes- 
santes in Bezug auf Stoffwanderungs- und Stoffumwandlungsvor- 
gänge darzubieten schien, oder in anderen Worten die 
Wachstumsgeschichte der Baumgräser mit Berücksichtigung der 
Bauverhältnisse näher zu verfolü'en. 

Die früheren Angaben über die Systematik, Verbreitung und 
äussere Lebensweise der Bambusgewächse haben eine Zusammen- 
stellung in einem Werk Schröter's') erfahren, und es schien mir 
überflüssig dieselbe hier wiederzugeben. Was die Physiologie der 
Bambusgewächse anbetrifft, besitzen wir abgesehen von älteren 
Beobachtungen über das Wachstum der Schösslinge nur die ein- 
gangs erwähnten Arbeiten von Kraus und Moli seh. Über die in 
Bambuspflanzen vorkommenden Stoffe besitzen wir ebenfalls spär- 
liche Angaben. Colnr) studierte ,,Tabaschir" in seiner klassischen 
Arbeit. Kozai'^) stellte chemische Untersuchungen über die stick- 
stofflialtigen Bestandtheile des Schösslings von Phyllostachys mitis 



])C. Schröter, Der Bambus und' seine Bedeutung iils Nutzpflanze.^; Basel, 1895. 
Vergleiclie ferner : 

E. Hackel, Bandjusaccre. Engler's*Die natürlichen Pflanzenfamilien. II, 2. p. 89. 

A. et C. Kivière, Les Bambous. Vegetation, culture et multiplication. 1878. 

2) F. Cohn, Über Tabaschir. Beiträge z. Biol. d. Pflanzen. Bd. IV, p. 365. 

3)Y. Kozai, On the nitrogenous non-albuminous Constituents of Bamboo shoots. 
Bulletin of tlie ddlege of Agriculture. Vol. I, Nr. 7. 



WACHSTUMSGESCHICIITE D. EAMBUSGEWAECIISE. 429 

an, unci dabei fand er das Vorkommen von Tyrosin und Asparagin 
neben kleineren Mengen der Nucleinbasen. Die Angaben über 
die Bauverhältnisse der Bambusgewächse finden wir in Arbeiten 
von Schwendener, Haberlandt, Strasburger, Hohenauer, 
Güntz, Ross und Magnus. Näheres über die Litteratur wird 
noch an geeigneten Stellen ErAvähnung finden. 

Die vorliegenden Studien ^Yurden im Laufe des acaderaischen 
Jahres 1898-1899 im botanischen Institute der Kaiserlichen 
Universität zu Tokio ausgeführt. An dieser Stelle spreche ich 
meinem hochverehrten Lehrer Herrn Prof. Dr. Miyoshi meinen 
wärmsten Dank für seine vielfache Belehruni»- und Anreüunii' 
aus. Es ist mir auch eine angenehme Pflicht Herrn Prof Dr. 
Matsumura für seine gütige Unterstützung hier meinen herz- 
lichsten Dank auszudrücken. 



II. Untersuchungsmaterial und Methodisches. 

Als Untersuchungsobjecte dienten mir die im botanischen 
Garten der Universität cultivierten Bambusarten, insbesondere 
Phyllostachys mitis, Piv., die sich in dieser Gegend in voller Ent- 
wicklung findet und im hiesigen botanischen Garten auf ziemlich 
grossem Grund gepflanzt ist. Die Wachstumsgeschichte der 
Schössliuge der obengenannten Art wurde von der ersten Anlage 
bis zum mehrere Meter hohen Halmzustand verfolgt. 

Auch die Schössliuge folgender Arten wurden zum Yer- 
gleichungszweck untersucht : 

im April austreibende — Bambusa palmata, Bambusa Veitchii', 

im ^lai austreibende — T'hijlloducUya 'puherula, Arundinaria 
japonica ; 

im Juni austreibende — Phyllostachys bambiisoidcs ; 



430 K. SHIBATA : 

im Jul i-A ug u s t austreibende — Arundinaria Simonis Aï'undi- 
naria Hind si i ; 

im O c 1 b e r - November austreibende — Arundinaria 3£atsu- 
77iurce, Arundinaria quadr annularis, Arundinaria Hindsii 
var graminea. 

Die Entwicklung der Ehizoraspitze wurde bei folgenden 
Arten im Herbst untersucht : Phyllostachys mitis, Phyllostachys 
hambusoides. 

Für andere Arten, die ich in meiner Untersuchung gezogen 
habe, verweise ich auf das am Ende dieser Arbeit beigefügte 
Artenverzeichniss. 

Um die Umwandlung und Wanderung der Stoffe in Reserve- 
stoffbehältern und in wachsenden Theilen zu verfolgen, 
bediente ich mich unter nöthigen Cautelen der üblichen micro- 
chemischen Methoden. Darüber sei hier folgendes bemerkt : 

Stärke. Meyer 'sehe Chloralhydratjodlösung^) wurde mit 
Vortheil benutzt. 

Glycose. (Reducierender Zucker). Meyer'sche") und 
Schi m per 'sehe"') Methoden wurden neben einander angewandt, 
dabei hat sich die letztere zur Nachweisung der kleineren 
Menge geeigneter erwiesen. Obwohl diese üblichen Methoden 
auch zu unserem Zweck völlig ausreichten, habe ich noch Sicher- 
heits wegen eine andere Reaction ausgeführt. Ich habe nämlich 
die Wasserauszüge von jungen Hahnen, Wurzeln, Rhizomen und 
Scheideblättern und auch den Blutungssaft mit essigsaurem 
Phenylhydrazin erwärmt, und dabei erhielt ich stets charac- 
teristische gelbe Nadelkrystalle von Glucosazon. 



])Vergl. Slrashurgcr, lîotauisclies Pratticuiii. III. Aiiöage. p. 1277. 
2)A. Meyer, Microchemische Reaction zum iS'acliweis tier reducireiiden Zuckerarten. 
Ber. d. D. B. G. 1885. p. G32. 

3)A. Zimmermann, Die botanisclie Microfechnik. p. 7-5. 



WACHSTÜMSGESCHICHTE D. BAMBUSGEWAECHSE. 431 

Rohrzucker. Die Unzuverlässigkeit der bekannten Saclis'- 
sclien Methode wurde vielfach von Autoren betont. Ich habe 
nur ausnahmsweise diese Reaction benutzt, während in den meisten 
Fällen ich mich der neulich von Hoffmeister^) aufgestellten 
Invertinmethode bediente.-) 

Gerbstoffe. Die Eisensalzlösungen, insbesondere die ätherische 
Lösung des Eisenchlorids, und auch die Kaliumbichromatlösung 
wurden angewandt. 

Fette. Alkannatinctur und 1-procentige Osmiumsäurelösung 
wurden benutzt. 

Asparagin. Die bekannte Borodin'sche Methode^) hat sich 
zweckentsprechend erwiesen. Zur Erkennung der erhaltenen 
Krystalle als Asparagin diente mir hauptsächlich die AVinkel- 
messung. Vielfach wurde das Borodin'sche Verfahren mit 
gesättigter Asparaginlösuug angewandt. Ferner diente mir 
Diphenylamin-Schwefelsäure zur Unterscheidung von Asparagin 
und Salpeter. 

Tyrosin. Die nach Borodin'scher Methode behandelten 
Schnitte ergaben eine reichliche Ausscheidung von eigenthümlichen 
Krystalleu, die ohne Schwierigkeit mit Tyrosin identificiert 



1) C. Hoffmeister, Über den microchemîsclien Xacliweis von Kohrzncker in pflanz- 
lichen Geweben. Jalirb. f. wiss. Bot. Bd. XXXI, p. 688. 

2) Die von den „ Ebisu "-Brauerei bezogene Hefe-Beinkultur wurde zur Darrfteliung von 
Invertin verwendet, dabei habe icli wie folgt verfahren: Die Hefemasse wurde mittelst 
Filtration von der Kulturflüssigkeit befreit und nach wiederholtem Auswaschen mit Wasser zum 
dicken Brei angerührt. Der Hefebrei kam nach dem Verreiben im Mörser in den Thermostat 
bei 50° C, in welchem er für 10-12 Stunden gelassen wurde. Hiernach wurde die abfiltrierte 
Flüssigkeit mit 90^» Alcohol versetzt, und der dabei entstandene voluminöse Xicderschlag 
auf Fil trirpapier gesammelt, welcher, nach wiederholtem Auswaschen mit iWfô Alcohol und 
dann mit absolutem Alcohol auf Schwefelsäure getrocknet wurde. Die wässrige Lösung der 
erhaltenen weissen kreideartigen Substanz, die allein niemals Fe h ling' sehe Lösung reduciert, 
zeigte ein energisches Inversions vermögen. Bei wiederholten Versuchen habeich 
ferner in keinem fall die Beimengung von diastatischen und oellulosespaltenden Enzymen 
gefunden. Weitere Verfahren mit den Schnitten genau nach Hoffmeister. 

8)A. Zimmermann, Die botanische Microtechnik, p. 80. 



432 K. SHIBATA : 

wurden^) (Fig. 58). Bel zun g' sehe Glycerin-Methode-) wurde 
auch mit Vortheil benutzt, wobei sich schöne Nadelbüschel in 
den Zellen bilden (Fig. 61). Ich habe ferner zur Erkennung der 
Vertheilung des Tyrosins in eiweissarmen Gewebetheilen 
Mi lion's Reagens benutzt, und dabei wurden die tyrosinreichen 
Zellen schnell blutroth gefärbt. Die Färbung bleibt nach dem 
Auslaugen der zuvor mit absolutem Alcohol behandelten Schnitte 
mit dem warmen Wasser für 10-20 Minuten so gut wie gänzlich 
aus. Daher kann diese Rothfärbung niemals von Eiweissstoffen 
herrühren. 

Eiweiss. Biuretreaction und Ras pail's Reaction wurden 
vornehmlich benutzt. Mi 11 on 's Reagens kam zur Anwendung 
erst nach dem Ausziehen von Tyrosin in oben beschriebener 
Weise. 

Mineralstoffe. Die von Sc hi m per") empfohlenen Reagentien 
wurden verwendet. Die Controllversuche wurden öfters ausge- 



1) Dies geschah aus folgemloii Ciiüiiden: 

1. Die Gestalt der KrydaUe. Die feine jSI'adelbüschel in dendritisclier Gestalt oder 

Doppelpinselforni bietet ganz dasselbe Aussehen wie reines Tyrosin. 

2. Das optiüche Vakdteri. Im durchfallenden Licht erscheinen die Krystalle bräun- 

lich und im auflliUenden Licht weisslich seidenglänzend. Im polaiisierten Licht 
zeigen die Krystalle starke Doppelbrechung. 

3. Die LUsUchkeitsverhälfnisse. Die Krystalle sind unlöslich in kaltem Wasser, aber 

löslich in kochendem "Wasser, Ammoniak und verdünnter Salzsäure. Ferner 
sind die Krystalle unlöslich in heissgesättigter Tyrosinlösung. 

4. Das Verhalten beim Erhitzen. Wenn man den mit Tyrosinkrystallcn besetzten 

Objectträger auf der Flamme erhitzt, bis nebenbei vorhandene Asparaginkrystalle 
sich zu braunen Schäumen verwandeln (ca. 200° C), so sieht man, dass die 
Nadelkrystalle ganz unverändert bleiben. 

5. Das Verhalten gegen Millon's Reagens. Die Krystalle lösen sicli im Mi lion's 

Reagens ndt einer prachtvoll rothen Färbung der umgebenden Flüssigkeit. 
Die oben angeführten Merkmale reichen aus, die Krystalle microchemiscli als Tyrosin zu 
erkennen. 

2) Beizung, Eecherclies chimiiiue sur la Germination. Ann. d. Sc. nat. Bot. Ser. VII. 
T. 15, p. 209. 

o)A. F. W. Schimper, Zur Frage der Assimilation der Mineralsalzs durch die 
grüne Pflanze. Flora. Bd. 73. 1890. p. 210. 



WACHSTÜMSGESCHICHTE D. BAMBUESGWAECHSE. 433 

führt, um die Keinlieit der Eeagentien zu prüfen. Die micro- 
chemischen Reactionen wurden sowohl an frischen Schnitten als 
an auf Deckgläsern geglühten Aschen vorgenommen. 



III. Die Bauverhältnisse. 

Die Bauverhältnisse der Bambusen sind bisher spärlich 
und nur gelegentlich Gegenstand der anatomischen Forschung 
geworden. Strasburger^) hat den Bau des Geflissbündels von 
Bambusa vulgaris kurz geschildert. Das mächtig entwickelte 
Bastgewebe in Bambushalmen wurde vielfach von Schwenden er) 
Haberlandt"^) u. A. erwähnt. Ross^) bemerkte den anomalen Bau 
der Wurzeln. Die Betrachtungen über die Blattstructur finden 
wir in den Arbeiten von Güntz'), Magnus''), Haberlandt') 
und Schwendener'^). übrigens liegen uns noch einige kurze 
Angaben von Hohenauer') und Möbius^°) vor. Nun schien 
es mir wünschenswerth die Bauverhältnisse der Vegetationsorgane 
der Bambusgewächse einem genaueren Studium zu unterwerfen, 
damit für die physiologische Forschung dieser interessanten Pflan- 
zengruppe eine festere Grundlage geschaffen werde. Meine 



1) Strasburger, Über d. Bau n. Verrichtungen d. Leitungsbahnen. 1891. p. 363. 

2) Seh wendener, Das mechanische Princip in anat. Bau d. Monocotylen. p. 65. 

3) Haberlandt, Entwickhingsgtschichte des mech. Gewebesystems d. Pflanzen, p. 23. 

4) Ross, Beiträge z. Anatomie d. abnorm. Monocotylenwurzeln. Ber. d. Deutsch. Bot. 
Gesells.'.Ed. T, p. 338. 

5)Güntz, Unters, üb. d. anat. Structur d. Gramineenblätter, p. 37, 41, 44, 48, 64 etc. 

6) Magnus, Einfalzungen d. Zellmembran.' (Just's Jahresber. d. Bot. I, p. 367). 

7) Ha|berland[t, Vergl. Anat. d.jissim. Gewebesystems d. Pflanzen. Jahrb. f. wiss. Bot. 
Bd. XIII, p. 100. 

8) Schwendener, Die Mestomscheide der Gramineenblätter. Ges. Bot. Mitt. Bd. II, 
p. 178. 

9) Hohenauer, Vergl. anat. Unters, üb. d. Bau d. Stammes bei d. Gramineen, p. ôôC 

10) Möbius, Üb. d. eigent. Blühen von Bambusa vulgaris, (Ref. in Bot. Centralbl. 1899. 
Nr. 51, p. 479). 



434 K. SHIBATA : 

diesbezügliche Untersuchungen erstreckten sich auf sieben und 
zwanzig vorwiegend einheimische Formen, die sich in die vier 
Gattungen von Phyllostachys, Arundinaria, Bambusa und 
Dendrocalamiis vertheilen. Die wesentlichen Ergebnisse will ich 
in folgenden Zeilen kurz darzustellen versuchen. 

Das Rhizom. 

Die Khizome^) von Phyllostachys- und Arundinaria-Arieia. 
sind bekanntlich kurz gegliederte horizontal verlaufende Stengel- 
gebilde mit einer rundlichen oder rundlich-ovalen Querschnitt- 
form. Die iuternodiale Markhöhle ist stets stark reduciert, meist 
nur einige mm breit und kommt nicht selten zum gänzlichen Ver- 
schwinden. Wir wollen zunächst beispielweise ein Rhizominter- 
nodium von Phyllostachys mitis ins Auge fassen. 

Nächst unter der Epidermis kommt ein 1-3 schichtiger 
Ring von den englumigen langgestreckten sklerotischen Paren- 
chymzellen, deren Querwände öfters etwas schief gestellt sind. 
Das darinnen liegende 20-25 schichtige büudelfreie Parenchym 
wird als die primäre Rinde") aufgefasst. Es geht ohne scharfe 
Grenze zum Grundgewebe des Centralcylinders über, in welchem 
in üblicher Weise die collateral gebauten Gefässbündel zerstreut 
liegen. Sämmtliche Parenchymzellen sind verholzt und mit 
zahlreichen ovalen Tüpfeln versehen. In diesem Gefässbündel 
erblickt man ein typisch gebautes Gramineenbündel. Die grosse 
Lumenweite der Siebröhren ist dabei höchst auffallend^) ; es wurde 
oftmals einen Durchmesser von O.lö mm erreicht^), während 



1) Vergl. A. et C. Eivière, Les Bambous, p. 68, p. 2r;ß. 

2) Falkenberg, Vergl. Unters, üb. d. Vegetationsorg. d. Monocotylen. p. 163. 

3) Strasburger, Leitungsbahnen, p. 363. 

4) Die Angaben über die Lumenweite der Siebröhren einiger Pflanzen findet man bei 
Lecomte (Ann. d. Sc. nat. Bot. Sèr. VII, T. X, p. 242), 



WACHSTUMSGESCHICHTE D. BAMBUSGEWAECHSÈ. 435 

die grösste Parenchymzelle 0.09 mm und die Geleitzelien meist 
nur 0.005 mm weit sind. Die beiden seitlichen Gefässe, deren 
maximale Lumenweite 0.2-0.3 mm beträgt, communicieren mit den 
einschichtigen Belegzellen durch regelmässig angereihte quer- 
gestreckte Tüpfel. Die stark entwickelten Bastbelege um die 
Gefässbündel sind seitlich an der Grenze zwischen Hadrora und 
Leptoai und oft auch unterhalb der seitlichen Gefässe unter- 
brochen. Dadurch kommen ein oder zwei Paar Durchlassstellen 
zu Stande, die, wie zuerst Schwendener^) bemerkt hat, einen 
Stoßaustausch zwischen Bündelelementen und Grundgewebe 
ermöglichen. 

Wenn man die Querschnittbilder der Ilhizominternodien der 
übrigen Arten in Betracht zieht, so lassen sich unter denselben 
folgende drei Typen unterscheiden, nämlich : 

Erster Typus. Die äussersten Bündel, welche direct an 
die Binde grenzen, stehen vollkommen isoliert von einander. 
Hierher gehören Phyllostachys mitis, Phyllosiachys hamhusoides, 
Phyllostachys puberula und Arundinaria Narihira. 

Zweiter Typus. Die Bastbelege der äussersten Bündel 
verschmelzen sich häufig unter einander und auch mit den 
Baststrängen zu unregelmässigen Bastbändern. Als Beispiele 
dienen Arundinaria japonica, Bambusa borealis, Arundinaria 
Tootsik, A. Simoni, A. Hindsii etc. In den zwei letztgenannten 
Arten befindet sich jedoch oft ein nahezu vollkommener Bastring. 
(Fig. 3). 

Dritter Typus. Der echte subcorticale Bastring-), an wel- 
chen die Mestombündel innen angelehnt sind, befindet sich bei 



1) Schwendener, Das mechanische Princip. p. 107. 

2) Vergl. H a her land t, Entwicklungsgeschichte des median. Gewehesystems, p. 28.; 
Physiologische Pflanzenanatomie, p. 157. 



436 K. SHIBÀTA : 

Bambusa pahnata, B. Veitchii, B. paniculata, B. nipponica, 
B. ramosa, B. nana, Arundinaria quadrangular is, A. 3Iaisumurœ, 
A. variabilis, Arundinaria pygmœa und ferner Phyllostachys 
Kumasasa. (Fig. 2). 

Der Bastring der letzterwähnten Arten, welcher je nach 
Species verschieden stark ausgebildet ist, geht entwicklungs- 
geschichtlich aus einem entsprechend continuirlichen Cambium- 
ring^) hervor. Schwendener sagt"): ,, Für Bambuseen ist die 
Querschnittform des mechanischen Systems, wie ich sie früher 
beschrieben habe, characteristisch genug, um jede nähere Verwandt- 
schaft mit den Festucaceen oder irgend einem anderen Tribus 

auszuschliessen. Ein Bastring ist nicht vorhanden, " Diese 

Bemerkung Seh wen den er 's passt aber nach obigem Befunde 
auf die Rhizome nicht.^) Der subepidermale Sklerenchymring ist 
nur schwach entwickelt, es ist bei den meisten Arten nur 1-2 
Schichten dick. Die Dicke der primären Binde ist meist un- 
ansehnlich und variirt zwischen 4-35 Zellschichten. 

Der Bastring wird stets vielfach unterbrochen in den Knoten, 
um hier den neu eintretenden Blattspursträngen Platz zu machen. 
Ferner ist es als die Kegel hervorzuheben, dass innerhalb des 
Knotens der Bastbeleg des Mestombündels eine bedeutende Re- 
duction erfährt, und sich meist nur auf eine dünne Sichel um 
das Leptora beschränkt. Die sämmtlichen Elemente des Bündels 
sind hier kurzgliedrig, und die Seitenwände der Siebröhren sind mit 
den ausserordentlich zahlreichen Siebtüpfeln versehen. Bekannt- 

1) Haber land t, Entwicklungsgeschichte, p. 2S. 

2) Seh wen den er, Die Mestomscheide der Gramineenbliitter. p. 183. 

3) Die mittleren Durchmesser der Rhizome dieser drei Typen stehen ungefähr im Ver- 
hältnisse 6:3:1. Die Ausbildung der Bastplatte resp. des Bastrings in Khizoraen entspricht 
wohl den mit der Diinnheit steigenden Aufforderungen für die Biegungsfestigkeit. Jedenfalls 
gehört hier die Anordnung des mechanischen Systems in Khizomen nicht zum sogenannten 
taxonomischen Merkmalen. 



WACHSTUMSGESCHICHTE D. BAMBUSGEWAECHSE. 437 

lieh findet sich in dem Knoten die Vereinigung der Blattspur- 
stränge unter einander und mit den Achselknospenbündeln statt^). 
Die hier eintretenden Knospenbündel verbreiten sich in dem 
Knotengewebe nach allen Richtungen hin. Die Gefässe der Knos- 
penbündel setzen sich in üblicher Weise unter starker Krümmung 
an die der Blattspuren an"). Dennoch verdient die Art und 
Weise, wie der Übergang des Leptoms erfolgt, eine besondere 
Beachtung. Das Leptom des Knospenbündels ist bei der 
Ansatzstelle an Blattspuren so stark angeschwollen, dass ihr 
ganzer Umriss mit einer Spindel zu vergleichen ist. Figur 4 
stellt ein derartiges Gebilde dar. Diese angeschwollene Partie 
weicht von dem üblichen Bau des Leptoms in so fern ab, dass 
sie die Differenzierung ihrer Elemente in Siebröhren und Geleit- 
zellen nicht mehr aufweist, sondern aus lauter gleichartigen 
feinen ca. 5-6 fx breiten cambiformartigen Elementen") zusam- 
mengesetzt ist (Fig. 4 u. 8), und folglich im Querschnittbild ein 
regelmässiges englumiges Maschen werk darstellt (Fig. 5 u. 7). 
Die Anordnung dieser feinen cambiformartigen Elemente bietet 
eine grosse Eigenthümlichkeit dar. Die etwas schräg gestellten 
Endflächen der seitlich an einander stossenden Elemente scheinen 
ungeßihr auf derselben Querebene zu liegen, und demgemäss stellt 
der ganze spindelförmige Theil im Längsschnitt ein der Länge nach 
an einander angereihtes meist 5-7 fâches Stockwerk von cambi- 
formartigen Elementen dar. Die Elemente in den mittleren 1 



1) Vergl. Falkenberg, Yergl. Unters, d. Vegetationsorgaue. p. 187. 

2) Strasburg er , Leitungsbahnen, p. 353 u. p. .365. 

3) Selbst bei den relativ weiteren (z. B. bei Arundinaria Slmoni) überschreitet die Breite 
kaum 9 \j.. So weit ich unterriclitct bin, wurde derartige Structur l^isher in keiner anderen 
Pflanzen beobachtet. Zum Beispiel finden wir keine diesbezügliche Angabe bei verschiedeneu 
Monocotylen, die von Falkenberg {loc. cit.) und Strasburger {loc. cit.) gründlich 
untersucht wurden. Ob sie auch bei anderen Pflanzen vorkommt muss dtshalb zur Zeit 
dahingestellt bleiben. 



438 K. SHIBATA : 

oder 2 Etagen dieses Stockwerks sind sehr langgestreckt und 
besitzen fein undulierte Seitenwände mit zahlreichen grossen 
ovalen tiq^felartig verdünnten Stellen, die zum Beispiel bei 
Fhy llostachy s müis eine VI ehe von 5x3 /i erreichen (Fig. 6). In 
einem Ende dieses spindelförmigen Theils vermitteln die etwas 
breiteren Elemente, — die sich zu 4-5 je einer feinbetüpfelten 
Endfläche der Siebröhren und einzeln auch den Geleitzellen 
anschlissen, — den Uebergang zum normal gebauten Leptom des 
Knospenbündels (Fig. 10). Das andere Ende der Spindel setzt 
sich in verschiedener Neigung und oft sogar rechtwinkelig an das 
Leptom der Blattspuren an (Fig. 8). Einige Schichten der 
Scheideelemente grenzen gewöhnlich den spindelförmigen Theil 
vom Grundparenchyra ab. Die Zellwandbeschaffenheit der 
spindelartigen Theile weicht kaum von der des Leptoms ab ; sie 
zeigen nämlich ebenso starke Cellulose-Beaction mit Chlorzinkjod 
oder Schwefelsäure-Jod, und sie werden auch mit Anilinblau, 
Congoroth u.a., im nahezu gleichen Farbenton wie Siebröhren 
gefärbt. In späterem Alter tritt jedoch oft eine Spur Holzreaction 
in den Elementen der oben erwähnten mittleren Etagen ein. Was 
den Plasmagehalt dieser Theile anbetrifft, so scheint es nur auf 
einen zarten AVandbeleg beschränkt zu sein, wie es bei Siebröhren 
stets der Fall ist. In den ersten Entwicklungszuständen habe ich 
constatiert, dass diese Anschwellung aus den entsprechend ver- 
mehrten Längstheilungen der procambialen Zellen an der betreffen- 
den Stelle hervorgeht. In solchen früheren Stadien zeichnet sich 
diese Anschwellung durch besonders reichlichen Plasmagehalt 
und auffallend grosse Zellkerne aus, wie es Fig. 9 zeigt. 

Derartige spindelförmige Leptomanschwellungen in Knospen- 
bündeln habe ich regelmässig in Rhizomknoten sämmtlicher von 
mir untersuchter Arten gefunden, aber man findet sie am stärksten 



WACHSTU3ISGESCHICHTE D. BAMBUSGEWAECHSE. 439 

ausgebildet in den Rhizomknoten der Phyllostachy s- Arteü, wobei 
ihre Quersclmittgrösse sogar einem grossen Mestombündel nahe- 
kommt. Da die an Rhizombündel sieh ansetzenden Achselknos- 
penstränge in ihrer Gesammtheit ein physiologisches Analogon des 
haustorial Saugorgans darstellen, so ist es nicht unmöglich, dass 
dieses Gebilde in dem Sinne ausgebildet ist, dass es eine specifisch 
absorbierende Wirkung auf die Siebröhren der Mutterrhizome 
auszuüben vermag. Allerdings besitzt es in seinen anatomischen 
Merkmalen vieles gemein mit den üblichen Absorptionsgeweben^j. 
So lange aber die Stofftransportmechanik im Leptome noch nicht 
in allen Hinsichten aufgeklärt ist-), möge die nähere Erörterung 
der physiologischen Vorgänge, die sich in diesem abweichend 
gebauten Leptomtheile abspielen, auf eine künftige Gelegenheit 
verschoben werden. 



Der Halm. 

Die Bambushalme^) sind bekanntlich mit den hohlen Inter- 
nodien versehen, die bei PhyUostachys mitis oft eine ansehnliche 
Dicke von 20 cm erreichen. 

Die primäre Rinde ist stets weit schwächer entwickelt als 
in dem Rhizome ; diese Verhältnisse, wie sie schon von Falken- 
berg'') und Rothert'^) für andere Pflanzen nachgewiesen wurden, 
gehen noch aus den folgenden Beispielen deutlich hervor : 



1) Haberlanclt, Physiologische Pßauzeuanatomie. p. 186. 

2) C z a p e k , Über d. Leituiigswege d. organischen Baustoffe in Pflanzenkörper, p. 2-1 ; 
Lecomte, Etude du Liber des Angiosperme?. Ann. d. Sc. nat. !;èr. YIL T. X, p. 303. 

3) Vergl. Ei viere, Les Bambous, p. 134. 

4) Falkenberg, I.e. p. 134. 

5) Rother t, Vergl. anat. Unters, üb. d. Differenzen im prim. Bau d. Stengel u. Rhizome. 
p. 92. 



440 



K. SHIBATA : 





HALM 


EIIIZOM 


Durchmesser 
des 
Central- 
cylinders 


Dicke 

der 
Einde 


Q^^^ 


Durchmesser 

des 

Ceutral- 

cyl Inders 


Dicke 

der 
Einde 


33 


Phyllostacliys onitis 


130.0 


0.315 


412 


23.8 


0.728 


Phyllostachys hambmoides 


34.0 


0.059 


575 


20.0 


0.611 


33 


Phyllostachys pubenda 


56.0 


0.049 


1142 


22.0 


0.933 


23 


Phyllosta cJiys Ku m asasa 


2.7 


0.023 


120 


5.6 


0.780 


7 


Arundinaria Simoni 


14.0 


0.049 


285 


8.0 


0.494 


16 


Arundinaria japonica 


17.0 


0.050 


340 


8.4 


0.286 


29 


Arundinaria liindsii 


22.0 


0.059 


373 


18.0 


0.364 


49 


Arundinaria quadrangularis 


23.0 


0.045 


511 


8.0 


0.260 


31 


Arundinaria 3Iaisiimurœ 


2.5 


0.018 


139 


2.9 


0. 20 


15 


Arundinaria 2^yg'}iicea 


2.3 


0.036 


64 


4.5 


0.325 


14 


Arundinaria Narihira 


14.0 


0.049 


285 


21.0 


0.468 


45 


Bamhusa horealis 


5.5 


0.045 


122 


6.5 


0.212 


31 


Bamhusa pahnata 


10.0 


0.063 


159 


7.1 


0.624 


11 


Bamhusa Veitchii 


4.5 


0.023 


200 


2.9 


0.143 


20 



*Q. = Das Verhilltniss des ersteren zur letzteren. 

Die äussersten 1-2 Scliicliten Rindenparenchymzellen sind 
oft sklerotisch verdickt (Fig. 12) und unterscheiden sich von den 
übrigen nur durch eine grössere Länge ; folglich haben wir 
hierbei keineswegs mit einem Bastring, wie Haberia n dt einst 
annahm/) zu tliun. Die Bastbelege der peripherischen Gefäss- 
bündel und die dazwischen liegenden Baststränge verschmelzen 
mit einander zu un regel massigen Bastbändern, besonders häufig 
in dünneren Halmen von A.rundinaria 31atsumurœ, A. pygmœa 
etc. Dennoch begegnet man hier in keinem Fall dem echten 
Bastringe. 



1) Haberl andt, Entwicklungsgeschichte, p- 23. 



WACHSTUMSGESCHICHTE D. BAMBUSGEWAECHSE. 441 

Die zuerst von Scliwendener^) bei einigen Bambusaarien 
entdeckte eigenthümliche Parencliymlamelle, die quer in dem 
innenseitigen Bastbelege inseriert ist, habe ich auch in den 
Halmen aller echter Bambusasuten {B. vulgaris, B. nana und B. 
stenostachya), Dendrocalamus laliflorus und bei 2 Arundinaria- 
arten {Ä. Himhil, A. quadrangular is) aufgefunden, überdies 
habe ich die Fälle beobachtet, dass die Lamelle nur an einer 
Seite in das Grundparencliym übergeht, und dass sogar das Paren- 
chym in der Mitte des Beleges allseitig von Bastzellen umschlossen 
liegt (Fig. 17 u. 18). Nach den letzterwähnten Thatsachen 
erscheint es a priori sehr wahrscheinlich, dass dieses Parenchym- 
gewebe erst nachträglich aus einem Theil des Procambiums des 
Bastbeleges hervorgeht, wie es Haberlandt") schon vermuthet hat. 

In der That konnte ich in einem Procambialstrang in jun- 
gen Internodien zuerst nichts von dieser Parencliymlamelle 
erkennen. Sie differenziert sich erst später aus einer Strang- 
anlage, in welcher alle Formelemente schon fertig angelegt sind, 
derart, dass die langgestreckten Procambialzellen in betreffender 
Stelle successive Quertheilungen erfahren und zum Epen um- 
gewandelt werden (Fig. 21 u. 22). Die feinkörnige Stärke, die 
später dem Zucker Platz macht, tritt sogleich in diesem Gewebe 
auf (fig. 20 u. 22) und bleibt in demselben während der weiteren 
Ausbildung des Stranggewebes. In dieser AVeise dient die 
Parencliymlamelle den dem Mestom unmittelbar anliegenden 
Bastzellen als ein Speicherungsort der nötigen Baustoffe. Die 
durch diese Parencliymlamelle vom Mestom abgetrennte Bast- 
masse bleibt gewöhnlich in ihrer Ausbildung sehr zurück, wie 
das Fig. 19 zeigt. Nun schien es mir berechtigt diese Paren- 



1) Sch wendener, Das mechanische Princip. p. G5. 

2) Haberlandt, I.e. p. 23. 



442 K. SHIBATA : 

cliymlamelle als eine im Innern des Stranggewebes eingeschobene 
,, Stärkescheide"^), die in ihrer physiologischen Kolle der 
gewöhnlichen strangumgebenden gleicht, aufzufassen. Derartige 
Einrichtungen würden vielleicht zweckentsprechend sein, bei 
einem mit so starkem Bastbelege versehenen Bündel, wie es bei 
Bambuseen angetroffen wird^). 

Die schon erwähnten spindelförmigen Leptomanschwellungen 
des Knospenbündels kommen auch in dem Halmknoten vor, 
jedoch meist in schwächerer Ausbildung. 

Die dünneren Blatttragenden Zweige stimmen in ihrem Bau 
mit den dickeren Halmtheilen wesentlich überein. Bei einem 
solchen (mit einem Durchmesser kleiner als 1 mm) wird das 
Eindenparenchym zu 1-2 Schichten reduciert und mehr oder 
minder verdickt. Die aussenseitigen Bastbelege der peripherischen 
Bündel stossen oft direct an die Epidermis, so dass eine Art 
Bastrippe zu Staude kommt"). Derartige Eippenbildung konnte 
ich jedoch bei einigen Arten, wie Arundinaria Matsumurce, selbst 
in den dünnsten Zweigen (0.7 mm dick) nicht nachweisen. 
Was den Gefässbündelverlauf in den Halmen sowie in den 
Ehizomen anbetrifft, so gehört er dem Palmentypus^) an, and 
habe ich durch successive Querschnitte und Eängsschnitte in 
der Spitzen region constatiert, dass die grossen medianen Blatt- 
spurstränge 5-6 Internodien zurücklegen müssen, bevor sie sich 
an andere Blattspurstränge ansetzen. 



1) Vergl. Heine, Über physiologische Function der Stärkescheide. Ber. d. D. B. G. 
1885, p. 189. 

2) Allerdings wurde die mechanische Bedeutung, die De tief sen (Üb. d. Biegungselasti- 
cität V. Pflanzentheilen. Arb. d. Bot. Inst. Würzburg. Bd. III, p. 182.) diesen Parenchym- 
lamellen zuzuschreiben versuchte, von Sc h wenden er (Zur Lelirc v. d. Festigkeit d. Gewächse. 
Ges. Bot. Mitteil. Bd. II. p. 19-20.) genügend widerlegt. 

3) Vergl. Schwendener, Die Mestomscheide. p. 183. 

4)Pe Bary, Vergleichende Anatomic d. Vegetationsorgane, p. 271 ff. 



WACHSTU3ISGESCHICIITE D. BAMBUSGEWAECHSE. 443 



Der Stiel. 

Die vielen untersten Internotlien des Scliösslings von Phyllo' 
stachys mitis vereinigen sieb sehr früh zu einem verholzten soliden 
Gebilde, welches im fertigen Zustande 2-4 cm lang und nur 
1.0-1.5 cm dick ist. Das Gebilde, das ich hier ,, Stiel"^) nenne, 
lässt keinen Unterschied mehr zwischen Internodien und Nodien 
im inneren Bau erkennen. 

Die Einde dieses Theils besteht aus etwa 20 Schichten 
parenchymatischer Zellen. Die Bastbelege der äussersten Bündel 
verschmelzen sich zu einem vielfach unterbrochenen unregel- 
mässigen Band. Nach innen liegen zahlreiche Bündel, welche 
die ganze Länge des Stiels hindurch gedrängt verlaufen (Fig. 13) 
und dann in Bhizomknoten eintreten, um sich dort an die 
Blattspuren anzusetzen. So überwiegen im Querschnitte des 
Stiels sehr stark die Bündel, und das dazwischen liegende Paren- 
chym ist zu einem 2-4 schichtigen schmalen Gewebe reduciert. 
Diese Verhältnisse entsprechen wohl der Function des Stiels als 
Leitungswege und nicht als Speicherngsort. Die Querschnitt- 
form des Bündels mit Bastbelege ist in den meisten Fällen 
rundlich oval und es ist vollkommen von 3-6 schichtigen 
Bastzellen umo-eben. Daher ist der Stoffaustausch zwischen 
leitenden Elementen und Grund parenchym so gut wie gänzlich 
ausgeschlossen. Das Leptom nimmt die äussere Hälfte des Bündels 
ein und besteht aus einer Anzahl 0.04-0.05 mm breiten Siebröhr- 
en und englumigen Geleitzellen. Das Hadrom besteht aus nur 
einem (seltener zwei) grossen Gefässe (oft bis 0.15 mm weit), 
welches von einigen Schichten kleinzelligen Hadromparenchyms 



1) Dieser „Stiel" stellt also eine einzige Stoffleitungsbahn zwischen dem vom Schösslinge 
B'ch entwickelnden Halme und dem Rhizome dar. 



444 K. SHIBATA : 

umgeben ist (Fig. 14). Dazu kommen nocli einige Traclieiden 
mit netz-, spiral- oder ringförmigen Wandverdickungen. Bei den 
"übrigen Arten sind die ebenso schmalen Stieltheile in genau 
derselben Weise ausgebildet und sie stimmen in ihrer inneren 
Structur mit dem oben beschriebenen gänzlich überein. 



Die Wurzel. 

Die zahlreichen Wurzehi^) befinden sich radial angeordnet 
an den Bhizomknoten und den unterirdischen Hahnknoten ; sie 
erreichen bei Phyllostachys- Arten eine maximale Länge von 70 
cm mit einem Durchmesser von 4 mm. Zunächst will ich hier 
den anatomischen Bau der Wurzelrinde von Phyllostachys- und 
Â7'undinaria- Arten näher betrachten. 

Die äusserste Zellschicht der Rinde lässt sich als die Aus- 
senscheide unterscheiden, indem die äusseren und radiären 
Zellvvände sehr stark verdickt sind (Fig. 24« u. 25). Damit 
spielt sie die Bolle der schützenden Oberhaut austatt der Epi- 
dermis, die sehr früh zerstört und abgeworfen wird. Nach innen 
folgt die verholzte Bastschicht (Fig. 24«). Das Bindenparenchym 
lässt sich in die äusseren aus unregelmässig polygonalen weit- 
lumigen Zellen zusammengesetzten Schichten und die inneren 
aus regelmässig in radialen und concentrischen Beihen angeord- 
neten Zellen bestehenden Schichten imterscheiden"). Die letzteren 
sind von einer Anzahl radialer Lufträume durchzogen. Die 
Zellen der Endodermis besitzen eine starke Verdickung von 
inneren und radialen Wänden, die in üblicher Weise verkorkt 
sind (Fig. 28). Die stark verdickte Wandung ist zierlich ge- 



1) A. et C. Ei viere, Les Bambous, p. 93. 

2) Die Zellscliiclitenzahl der inneren Einde ist stets kleiner als die der äusseren. 



WACHSTUMSGESCHICHTE D. BAMBUSGEWAECHSE. 445 

schichtet und von feinen verästelten Kanälen förmlich durchsetzt 
(Fig. 28). Im Querschnitt stellt demnach diese nach Aussen 
zugekehrte C-Scheide ein symmetrisches Bild mit der oben er- 
wähnten ebenso C- förmigen Aussenscheide dar^). Die 1 oder 2 
innersten unmittelbar derEndodermis anliegenden Kindenschichten 
sind bei Phyllostachys Kumamsa, Bambusa borealis und Ärundinaria 
quadrangularis als Verstärkungsriug") ausgebildet, indem die 
Zellen durch innenseitige C-förmig verdickte und stark verholzte 
AYände ausgezeichnet sind (Fig. 34). 

So weit es den Bau der Wurzelrinde betrifft, zeigen die echten 
Bambusa- Arten nämlich B. vulgaris, B. nana, B. stenostachya, 
B. arundinacea u. a. ein von dem oben beschriebenen ganz ab- 
weichendes Verhalten. Hier weisen die Zellen der subepidermal- 
schicht keine Wandverdickung auf und ebenso verhalten sich 
die persistenten Epidermiszellen. Darauf folgen 2-3 Schichten 
enger Bastelemente, welche nach innen scharf von den weit- 
lumigen Eindenparenchymzellen abgesetzt werden (Fig. 26 u. 27). 
Die äussere Kinde besteht nur aus einigen Zellschichten, während 
die von grossen Lufträumen durchzogenen inneren Schichten viel- 
fach dicker sind (Fig. 27). Die Endodermiszellen sind ringsum 
verdickt und bilden die sogenannte 0-Scheide^). Merkwürdig 
ist ferner der Bau des Verstärkungsrings. Die innersten 1-oder 
2-schichtigen Rindenparenchymzellen führen an ihren inneren 
Wänden eine Anzahl unregelmässig gestalteter aus reiner Cellulose 
bestehender Auswüchse, die häufig die äusseren Wände erreichen, 
so dass sie im Querschnitt die ganzen Zellen nahezu aus- 



1) Dasselbe Verhiiltniss wurde von Schwendener (Die Scliiitzscheideii und ihre 
Verstärkungen. Ges. Bot. Mitt. Bd. II, p. 120, 127.) auch bei einigen Orchideenluftwurzehi 
bemerkt. 

2) Schwendener, Die Schützscheide und ihre Verstärkungen. Ges. Bot. Mitt. p. 132. 

3) Vergl. Schwendener, I.e. p. 128, Tabelle. 



446 li. SHIBATA : 

zufüllen scheinen (Fig. 31 u. 32)^). Eine Anzahl einheimischer 
Arten, die wegen i lirer 6 Stamen bisher in Bambusa eingereiht 
wurden, z. B. B. Veitchii, B. ijalmata, B. horealis etc., weisen 
jedoch im Bau der Wurzelrinde eine vollkommene Übereinstim- 
mung mit Ärundinaria auf und so schien es mir berechtigt, unter 
Berücksichtigung noch anderer Merkmale, — vor allem : Fehlen der 
in Bastbelege eingeschobenen Parenchymlamellen, die Gestalt der 
Caryopse, die langkriechenden E-hizorae u.s.w. — diese Formen- 
gruppe von Bambusa loszutrennen und als eine neue Section in 
Arundinariece aufzunehmen. Die Aufstellung dieser neuen 
Formengruppe bietet uns doj)pe]tes Interesse ; denn einmal erweist 
dieselbe, dass die Eintheilung nach der Zahl der Stamen, aufweiche 
man in der Bambuseensystematik ein grosses Gewicht zu legen 
pflegt, nicht immer durchführbar ist. Andererseits kommt diese 
Formengruppe^) in ihrer Verbreitung auf Japan^) beschränkt vor. 
AVir gehen nun zur Betrachtung des Centralcylinders über. 
Die Anordnung der Leitbündel weicht, wie es von Ross^) 
nachgewiesen wurde, vom typischen Bau der Monocotylen ab. Zu 
den normalen peripherischen radialen Bündeln, die ausserordent- 
lich polyarch sind,^) kommt noch eine Anzahl der inneren iso- 
lierten Hadrom- und Leptomstränge hinzu. Im Querschnitte 
beliebiger junger Wurzeln bemerkt man innerhalb der dünnwandi- 
gen Endodermis ein oder zwei Schichten des ununterbrochenen 
Pericambiums (Fig. 23), dessen allgemeine Vorkommniss in Bam- 



1) Derartige Structur findet man nicht in S cli wendener' s Aufzählung der verschiedenen 
Verstärkungsformell (vergl. I.e. p. 132). 

2) Die Anzahl der bis jetzt bekannten hierhergehörigen Arten ist neun. Vergl. Makino. 
Bambusacepe Japonicse. Bot. Mag. XIV, Xr. 156, p. 20. 

3) Vielleicht auch in China. 

4) Koss, Beiträge zur Anatomie abnorm. Monocotylenwurzel. Ber. d. D. B. G. Bd. I, p. 337. 
5)Z. B. in einer 4 mm dicken Wurzel von PhyUoi<tachß mills habe ich mehr als 150 

gezählt. 



WACHSTUMSGESCHICHTE D. BAMBUSGEWAECHSE. 



447 



buseen um so mehr Beachtung verdient, als bei den meisten 
Gramineen die primordialen Gefässe nach van Tieghera^) direct 
der Endodermis anzustossen pflegen. Die inneren Hadromstränge 
kommen in etwa drei concentrischen Ringen angeordnet vor. 
Dazwischen liegen zerstreut die inneren Leptomstränge, welche 
jedesmal aus den 1 oder 2 Siebröhren und den euglumigen 
Geleitzellen bestehen (Fig. 43 etc). Bei echten Bajubusa- Arten 
besitzt die stets einzeln stehende Siebröhre einen regelmässig 
ovalen Umriss (Fig. 38). Die Gesammtanzahl der inneren Leptom- 
stränge beträgt in den meisten Fällen, w'ie folgende Beispiele 
lehren, eine Hälfte der peripherischen, aber bei echten Bamhusa- 
Arten kommen beide fast in gleich grosser Anzahl vor. 



Phyllostaclnjs mit is 

P. hamhusoides 

P. jmherula 

Avil ndina via jcqwnica 



Ä. 


Hinds ii 


A. 


Matsumurœ 


A. 


variai) il is 


A. 
Bamhusa 
B. 


jiygmœa 

palmata 

Veiichii 


B. 


raiiiosa 


B. 


paniculata 


B. 
B. 
B. 


nippon ic a 

vulgaris 

sienostachya 


B. 


nana 



Dendrocalamus laiiflorus 



Peripherisches 


Inneres 


Leptorn 


Leptoni 


84 


42 


83 


41 


47 


24 


75 


40 


181 


97 


27 


13 


30 


16 


43 


22 


41 


20 


29 


15 


20 


10 


39 


20 


33 


16 


70 


QS 


47 


41 


48 


45 


81 


55 



l)Van Tieghera, Les Eaciue. p. 123; Vergl. Mo rot, Eeclierche sur le Pericvcle. 
Ann. d. Sc. nat. Sèr. VI, T. 20, p. 233 und auch Falkenberg, Vergl. Unters, d. Vege- 
tationsorgane, p. 192. 



448 K. SHIBATA : 

Die übrigen Elemente des Ceutralcylinders werden, abge- 
sehen vom centralen Markparenchym, prosencbymatiscli zugespitzt 
und zugleich stark verdickt. So entsteht hier ein hohlcylindri- 
scher mechanischer Ring, in welchem sämmtliche Leitstränge 
eingebettet liegen. Hier muss noch eine Frage gelöst werden : 
In welcher Weise geschieht die Communication zwischen den 
einzelnen leitenden Elementen, die von einander getrennt im 
mechanischen Gewebe liegen? Zwar hat Reinhardt^) die in Frage 
kommenden Verhältnisse bei den anomal gebauten Wurzeln von 
Musaceen, Fandanaceen, Palmeen und Gydcmthaceen ermittelt 
und manch interessantes entdeckt. Betreffs der Communication 
zwischen einzelnen Leptomsträugen in unserem Fall muss vor 
allem bemerkt werden, dass die ausserordentlich stark verdickten 
und verholzten Pericambiumzellen als die Leitungswege zwischen 
den peripherischen Leptomsträugen kaum in Betracht kommen^). 
Wenn man nun die Zahl der in beliebigen zwei Wurzelquersch- 
nitten vorkommenden Leptomstränge sorgfältig mit einander 
vergleicht, so kann man leicht eine bedeutende Abnahme der- 
selben nach dem Wurzelspitze wahrnehmen, wie es aus einigen 
beigefügten Beispielen hervorgeht : 

Zahl der Leptoaistriliige in 
Proximalende Mitte Distalende 

22.5 cm langes WurzelstücF) 128 — 104 

12 „ „ „ 116 108 102 

Der Umstand beruht bloss darauf, dass die inneren Lep- 
tomstränge sich unter einander und mit den peripherischen im 
weiteren Verlauf allmählig verschmelzen, wie man sich durch 
Betrachtung successiver Querschnitte überzeugen kann. Die 

l)Eeinliardt, Das leitendegewebe einiger anomalgebauteii Monocotylenwurzel. Jahrb. 
f. wiss. Bot. Bd. XVI, p. 'iZ^. 
2) Reinhardt, I.e. p. 361. 
o] von Fhylloslachys bunibusoides- 



WACHSTÜMSGESCHICHTE D. EAMBUSGEWAECHSE. 449 

Figuren 39 und 40 zeigen einige Fälle der erwähnten Yersclimelz- 
ung. Dieser Modus des Leptomverkehrs ist nach Keinhardt'- 
schen Angaben auch bei anderen anomal gebauten Wurzeln häufig 
verwirklicht^). Der zweite Modus ist aber von mehr wirksamer 
und aufiiilliger Art. Bei jeder Ansatzstelle der zahlreich ent- 
springenden Neben wurzeln an Centralcylinder werden sämmtliche 
hier befindliche peripherische sowie verschieden tief liegende 
innere Leptomstränge in einem System förmlicher Anastomosen- 
bildung zusammengehalten, welche bei den meisten Arten sich 
auf die etwa 10 periperischen Leptomstränge hinüberstreckt und 
bei Bambusa-Arien sogar die Hälfte des ganzen Umfangs des 
Centralcylinders in sich umfasst. Die etwas schematisierten 
Figuren 35 und 36 illustrieren das obengesagte. Das hier die 
Verbindung zwischen einzelnen Leptomsträngen herstellende 
Gewebe besteht aus den plasmareichen parenchymatischen Zellen, 
die mit den ansehnlich grossen Zellkernen und den dünnen un- 
verholzten Wänden versehen sind (Fig. 43 u. 44). Die Versch- 
melzung zweier Gefässe habe ich nur selten gesehen, während 
bei der Ansatztelle der Nebenwurzel sämmtliche mechanische 
Zellen zufolge reichlicher Tüpfelbildung uud häufiger auftretender 
Querwände einen holzparenchymartigen Character annehmen und 
demgemäss dem Saftaustausch zwischen den eingebetteten Gefässen 
besser angepasst sind. Den directen Anschluss der Leptomelemente 
an Holzparenchymzellen, wie es von Reinhardt für llusaceen 
uud Gydanihaceen^) nachgewiesen wurde, habe ich auch häufig 
bei Bambuswurzeln angetroffen (Fig. 41). 

Die Basaltheile des Centralcylinders der Nebenwurzel sind 



1) Kein hard t, I.e. p. 364, p. 343 etc. 

Vergl. Ross, Beitr. z. Anat. abnorm- Monocot. wnrzel. p. 334. 

2)Eeinhardt, /.c. p. 343, p. 346 und p. 348. 



450 Iv. SHIBATA : 

aus dem stark verdickten porösen rechteckigen parenchymati- 
schen Zellen gebildet, durch welche die kurzen Basalglieder jedes 
Leptomstrangs abwärts verlaufen, um sich dem oben erwähnten 
Leptomanastomosencomplex der Hauptwurzel anzuschliessen 
(Fig. 37). Hingegen scheinen die Gefässe der Nebenwurzel 
basalwärts meist blind zu endigen, so dass sie nur selten in 
directen Zusammenhang mit denen der Hauptwurzel kommen. 

Die Nebenwurzeln zeigeu in ihrem Bau alle Merkmale der 
bezüglichen Hauptwurzeln, dennoch fehlen ihnen stets die inneren 
Hadrom- und Leptomstränge (Fig. 47 u. 45). 

Die Ansetzung der Wurzeln an die Staramorgane geschiet 
in der bei Monocotylen üblichen AVeise'). Die Elemente des 
mechanischen Kings des Wurzelcentralcylinders breiten sich 
scheibenförmig aus und verschmelzen sich mit den äussersten 
Bündeln der Stammgebilde. Einzelne losgelöste Wurzelstränge 
dringen noch weiter ein und schliessen sich den peripheren 
Stammbündeln an, wobei das Leptom der ersteren solch eine 
Umgestaltung erfährt, wie sie bei den Knospenbündeln beobachtet 
wird'). 

Es erübrigt noch einen interessanten Befund kurz zu 
erwähnen. Die Rindenparenchymzellen der Nebenwurzeln, mit 
Ausnahme von den innersten 2-3 Schichten kleinlumiger Zellen, 
sind gewöhnlicli von einem Pilz bewohnt, der in jedem Zelllumen 
ein ansehnliches Knäuel von dicken verschlungenen Mycelfäden 
bildet (Fig. 46 u. 47). Die verpilzten Wurzeln bieten trotzdem 
ein ganz normales und gesundes Aussehen dar. Die Mycelfäden 
treiben hie und da sogenannte Vesikulen aus und producieren 
oft massenhaft gelbe körnige Substanz von nicht genau bekannter 



1) Vergl. Falkenberg, Vcrgl. Unters, p. 196. 

2) Vergl. p. 437. 



WACHSTUMSGESCHICHTE D. BAMBUSGEWAECHSE. 451 

chemischer Zusammensetzung. Die Stärke verschwindet gewöhn- 
lich vom inficierten Gewebe. Es unterliegt also keinem Zweifel, 
dass wir in diesem Fall mit einem en dotrophi sehen Myco- 
rhiza^) zu thun haben. Der Wurzelpilz fehlte in keiner der von 
mir untersLichten Arten und ist sowohl in den epiderralosen 
Nebenwurzeln von Arundinaria- und Fhyllostachys-Arten als 
in den mit Epidermis versehenen Bambiisa-^ehenviuYzeln con- 
stant nachweisbar. Die Eindengewebe der Hauptwurzeln habe 
ich meist pilzfrei gefunden, abgesehen von den dünneren Wurzeln 
von Arundinaria variabilis, Bamhusa ramosa, etc. Die Lösung 
der Frage nach der physiologischen Kolle"), die dieser Pilzsym- 
biont in der Ernährung der Baumgräser spielt, wall ich mir 
für künftige Studien vorbehalten. 



Die Blattgebilde. 

Die in zwei entgegengesetzten Reihen gelegenen, breiten 
Scheideblätter'^) umliüllen übereinander den ganzen Schössling 
und auch die wachsende Spitze des Rhizoras. 

In der basalen, zum Knotengewebe übergehenden Region 
jedes Scheideblattes weisen die Leitstränge in ihrem Hadrom kein 
grosses getüpfeltes Get'äss auf, sondern sie besitzen nur zahlreiche, 
oft mehr als 15 Ring- oder Spiralgefässe, die mit einander man- 
nigfach anastomosieren. In dem in der mittleren Partie des 
Scheideblattes ausgeführten Querschnitte erblickt man parallel- 

1) Frank, Lehrb. d. Botanik. Bd. 1, p. 274; ITber neue^Mycorliiza-Fornien. Ber. d. D. 
B. G. Bd. V, p. 400. 

2) Es wurde neuerdings vielfach, die Ansicht geäussert, dass die|Pflanzen die mit Mycorhiza 
ausgerüstet sind, der Assimilation des freien Stickstoffs befähigt[^seien. Yergl. Jause, Ann. d. 
Jard. Bot. Buit. Vol. 14, p. 200, und auch Nob be, Landw._Versuchs-St. Bd. LI, p. 241. 

3) Ki viere. Les Bambous, p. 76-82, p. 231. 



452 K. SHIBATA : 

verlaufende, abwechselnd starke und schwache Leitbündel, die in 
ihrem Bau kaum von den dem Stammorgan eigenen abweichen 
(Fig. 48). Sie sind mit einander durch die aus einigen Siebröhr- 
en und Gefässen bestehenden Qneranastomosen verbunden 
(Fig. 53). Die Bastbelege auf der Leptomseite stossen gewöhnlich 
unmittelbar an die stark verdickte Epidermis der Aussenfläche an 
(Fig. 51), aber bei dicken fleischigen Scheideblättern der Phyllo- 
stach y s- Arten liegen fast alle Bündel mit iliren Bastbelegen ganz 
frei im Parenchym (Fig. 52). Entgegengesetzt den stärkeren 
Bündeln liegen die bandförmigen, meist 2-3 schichtigen Bast- 
stränge an der Blattinnenseite. Die letzteren kommen bei Ärundi- 
naria 3Iatsumurœ sonst auch an der Blattaussenseite hier und da vor 
(Fig. 49). Das Scheideblattparenchym besteht aus dünnwandigen, 
saftreichen Zellen,^) von denen einige subepidermale Schichten bei 
unterirdischen, harten Scheideblättern sclerenchymatisch verdickt 
sind. Bei den derben oberii'dischen Scheideblättern von Arundi- 
naria-KYiQw tragen die an die Intercellularräume angrenzenden 
Flächen der Parenchymzellen die eigenthümlichen bald kugel- 
förmigen, bald stäbchenförmigen Auswüchse, die starke Holz- 
reaction geben (Fig. 54). 

Die Spaltöffnungen kommen an der Ober- sowie Unterseite 
der Scheideblätter vor. 

Die laubblatttragenden Blattscheiden stimmen in ihrem Bau 
mit den oben geschilderten Niederblättern wesentlich überein. 

Die Laubblätter der Bambuseen sind schon wiederholt von 
vielen Forschern anatomisch untersucht worden. So haben 
Kareltschikofi'), Magnus und Haberlandt die Armpallisa- 

1) Die Parenchymzellen ansgewaclisenerSplieideblätter enthalten fast keine Stärke, sondern 
viel Glykose. 

2) Karel tschikoff', Üb. d. faltenförmiga Verdickungen in d. Z?llea einiger Gramineen, 
p. 180. (Referat), 



WACHSTUMSGESCHICHTE D. BAMBUSGEWAECHSE. 453 

dennatur des Assimilationso-ewebes erkannt. Bei Güntz finden 
wir Angaben über einige allgemeine Characteristik der Bambnseen- 
blätter, dabei führte das energische Auftreten der mechanischen 
Elemente ihn zur x\ufstellung des ,, Bambuseentypus " der 
Gramineenblätter^). Bei dieser Sachlaj;e würde es berechtigt sein, 
dass ich mich hier nur auf einige kurze Notizen beschränke. 

Um jedes Mestombündel bemerkt man zweierlei Scheiden"), 
d.h. eine farblose Parenchymscheide und eine innere verholzte 
Bastscheide (Fig. 55 u. 56). Die stets einschichtige Parenchym- 
scheide fehlt selbst bei kleinsten Bündeln nicht; bei stärkeren 
Bündeln ist sie oft dort stark verdickt und verholzt, wo sie an 
subepidermale Bastrippeu anschliesst. Die wenigstens um das 
Leptom stets vorhandene Bastscheide wurde von Seh wendener 
als Mestomscheide ausgezeichnet') und der echten Schutzscheide 
zur Seite gestellt. Dieselbe ist um die kleineren Bündel meist 
einschichtig, aber bei den stärkeren nicht selten mehr als 4 
Schichten dick. Ferner verhalten die Elemente dieser Scheide 
sich gegen Schwefelsäure kaum anders als gewöhnliche verholzte 
Bastzellen, während sich die un verholzte Parenchymscheide gegen 
dieses Beagens sehr widerstandsfähig erweist. So ist die in Rede 
stehende Scheide als eine vereinfachte Form der das Mestom 
vollkommen uraschliessenden Bastscheide, wie ich sie schon bei den 
Stielbündeln beschrieben habe, aufzufassen. 

IV. Der Entwieklungsvorgang der Schösslinge. 

Als Gegenstand der folgenden Darstellung diente mir Phyllo- 
stachys mitis. 

l)Güutz, Unters, üb. d. anat. Struct, d. Gramiueenbl. p. 64. 

2) Sc h wendener, Die Mestomscheiden der Gramineeublätter ; Vergl. Strasburger, 
Leitungsbahnen, p. 344. 

3) Sc h wenden er, Lc. p. 178. 



454 K. SHIBATA : 

Die auf jedem Knoten der wachsenden Rhizomspitze angelegte 
Knospe wird erst im nächsten Jahre zu einem kleinen Schössling 
mit dem schon differenzierten, verholzten, ca. 1 cm langen Stiel 
ausgebildet. Diesen letzteren nenne ich kurzweg den Schössling 
des 2ten Stadiums, während die dem Knoten dicht anliegende, 
stiellose Knospe als Istes Stadium von diesem unterschieden 
wird. A¥enn man einen Querschnitt in der oberen Region dieses 
kleinen Schösslings ausführt, so sieht man den Centralcylinder 
gesondert in einen peripherischen, schmalen, bündelführenden Ring 
und in umfangreiches Markgewebe, welches sich nach unten 
allmählis; verschmälert. Auf dem Länsisschnitt sieht man dicht 
unterhalb des Urmeristems vom Vegetationspunkt beginnend eine 
grosse Anzahl abwechselnd stärkereiche und stärkearme Zonen, 
welche letztere sich in späteren Stadien zu Internodien verlängern. 

Der Schössling des 2ten Stadiums nimmt im Laufe des Som- 
mers an Grösse zu und wächst im Spätherbst (October-November) 
schon zu einem mittelgrossen Schössling des oten Stadiums. In 
diesem Zustande verharrt er während des Winters. 

Anscheinend schon in März tritt eine rasche Zunahme an 
Grösse ein und im Anfang April erreichen die Schösslinge unter 
der Erde eine ansehnliche Grösse, die ich als 4 tes Stadium 
kennzeichnete. Der Schössling ist mit zahlreichen geräumigen, 
dicken Scheideblättern bedeckt. Der verholzte Stiel ist nun ca. 
2 cm lang und 0.9-1.2 cm dick geworden. Die unteren, an den 
Stiel sich direct anschliessenden, etwa zehn Internodien, deren 
mittlere Höhe 1-2 cm beträgt, sind mit zahlreichen 3-4 mm dicken 
und bis etwa lö cm langen Wurzeln dicht besetzt. Ueber den 
inneren Bau ist folgendes zu bemerken. Die Spitze, unterhalb 
des Urmeristems, besteht aus abwechselnd stärkereichen und 
stärkearmen Zonen, deren Zahl binnen G mm 40 beträgt. 



WACHSTUMSGESCHICHTE D. BAMBUSGEWAECHSE. 455 

Die Dicke der beiden Zonen nimmt nach unten zu, und erst in 
Entfernung von 1.5-2.0 cm vom Vegetationspunkt entstehen im 
internodialen Markgewebe sichtbare Querrisse, deren Weite und 
Höhe in den nachfolgenden Internodien immer zunehmen. Diese 
primordialen Markhöhlen erreichen in den unteren, mit Wurzeln 
besetzten Internodien eine maximale Höhe von ca 3 mm, dabei 
besitzt das Diaphragm eine Dicke von 2.5 mm. Die Bündel- 
anlage in der Spitzenregion besteht aus engen procambialen 
Zellen. Überhalb der Stelle, wo die erste Markhöhle zum 
Vorschein kommt, erfolgt schon die Differenzierung in Elemente 
des Bündels. 

Indessen tritt die Spitze des Schösslings allmählig auf der 
Erdoberfläche hervor ; vom Ende April ab erfolgt dann ein 
rasches Wachstum desselben. Schon in Mitte Mai erreichen 
mehrere Schösslinge eine Höhe von 8-10 Meter, und an den 
oberen Kodien findet die Entfaltung von blatttragenden Aesteu 
statt. Mehrere Internodien auf der Erde sind nur 6-9 cm 
lang und nach oben nimmt die internodiale Länge graduell zu. 
In mittlerer Höhe der Pflanze erreichen sie die Länge von 20 
cm und mehr. Bis auf diese Region haben alle Internodien ihr 
Längenwachstum vollendet. Mehrere darauf folgende Internodien 
besitzen basale Wachstumszonen. Die von dort nach oben lie- 
genden Internodien verjüngen sich allmählig zum Vegetations- 
punkt. Die noch in Streckung begriffenen Internodien sind stets 
mit Scheideblättern umhüllt. Der Schössling in diesem Zustande 
ist im 5ten Stadium. 

Hier lasse ich einige Zahlenangaben folgen : 



456 



K. SHIBATA 







Stadium IL | Stadium III. j Stadium IV. 




12 3 


1 


2 ! 3 


1 


2 


3 

48.5 


Länge 
in 
cm. 


Aeussere 
Curvatur 


4.4 


4.2 4.6 1 20.2 


19.3 


16.0 


56.7 


48.0 


Innere 
Curvatur 


3.7 


3.8 


3.4 


13.9 


13.4 


12.0 


Maximalumfang in cm. 


5.3 


4.9 4.4 


15.6 


14.8 


13.8 


34.0 


31.5 


34.5 


Gewiclit in Gr. 


4.18 


4.21 4.11 


173.0 154.0 1 116.5 

1 


1831.0 


1432.0 


2062.0 



Tägliche Zuwaehsmessimgen 
an jedem Mittag vom 24 April 





I. 


IL 


IIL 


IV. 


Datum. 


Länge 
in 
cm. 


a5 

3 


Länge 
in 
cm. 




Länge 
in 
cm. 




Länge 
in 
cm. 


Ü 

3 

tsi 


April 


24 


34.9 




37.2 




45.7 




27.2 






25 


38.1 


3.2 


43.1 


5.9 


51.3 


5.6 


31.0 


3.8 




26 


44.2 


6.1 


50.6 


7.5 


61.1 


9.8 


37.6 


6.6 




27 


53.4 


9.2 


62.4 


11.8 


75.0 


13.9 


46.1 


8.5 




28 


62.5 


9.1 


73.3 


10.9 


9].7 


16.7 


56.6 


10.5 




29 


74.1 


11.6 


87.4 


14.1 


114.4 


22.7 


72.9 


16.3 




30 


86.8 


12.7 


101.3 


13.9 


135.3 


20.9 


86.8 


13.9 


Mai 


1 


106.9 


20.1 


123.9 


22.6 


164.0 


28.7 


108.8 


22.0 




2 


137.9 


31.0 


158.3 


34.4 


205.8 


41.8 


141.9 


33.1 




3 


175.9 


38.0 


202.4 


44.1 


253.3 


47.5 


180.1 


38.2 




4 


199.1 


23.2 


230.7 


28.3 


283.5 


30.2 


207.7 


27.6 




5 





— 


277.1 


46.4 


336.4 


52.9 


263.2 


55.5 




6 


300.5 


50.7 


3o3.7 


56.6 


39S.6 


62.2 


319.0 


55.8 




7 


325.5 


25.0 


363.7 


30.0 


418.0 


19.4 


346.5 


27.5 




8 


338.6 


13.1 


375.6 


11.9 


447.6 


29.6 


361.2 


14.7 




9 


382.5 


43.9 


417.6 


42.0 


500.6 


53.0 


409.4 


48.2 




10 


423.9 


41.4 


464.5 


47.0 


546.9 


46.3 


452.6 


43.2 




11 


475.3 


51.4 


515.6 


51.1 


605.0 


58.1 


514.1 


61.5 




12 


551.0 


75.7 





— 





— 


597.9 


83.8 




13 


580.1 


29.1 


619.2 


5L8 


721.4 


58.2 


636.2 


38.3 




14 


631.6 


51.5 


661.9 


42.7 


786.4 


65.0 


664.6 


28.4 




15 


719.6 


88.0 


744.7 


82.8 


846.1 


59.7 


710.7 


46.1 

1 



NB.—* Zuwachs ist Mittel von 2 Tagen. 

^'^ Nach Beobachtungen des hiesigen meteorologischen Observatoriums. 
Anin. 1. Also bei diesen Messungen stieg der maximale Zuwachs nicht selten über 80 cm pro 
.\nm. 2. Das bisher bekannte stärkste Wachstum der Bambushalme beträgt 91.3 cm pro 24 
Anm. 3. Auf die hier beobachteten auffälligen Wachstumsschwankungen und andere interes. 

1) Die älteren Angaben über das Wachstum der Bambuspflanzen findet man bei Kraus 



WACHSTUMSGESCHICHTE D. BAMBUSGEWAECHSE. 



457 



Um die erstaunlich grosse Schnelligkeit des Wachstums von 
Bambnsschösslingen in dem 5ten Stadium zu demonstrieren^), 
führe ich hier einige von mir ausgeführte Messungen an Phyllo- 
stachys m it is an, 



von Phyäostachys-Halmeu, 
bis 15 Mai ausgeführt. 



V. 


VI. 




Länge 


50 


Länge 


i 




Mittlere 


Mittlere 


in 
cm. 


o 

1 
N 


in 
cm. 


s 


Wetterangaben. 


Tempera- 
tur.** 


relat. 
Humidi- 

tät.** 


106.6 








klar, leiser Wind 


14°.4 C 


50.2 


121.2 


14.6 


52.6 




klar- wenig trüb 


12°.5 


73.0 


144.8 


23.6 


76.3 


23.7 


klar 


15°.4 


74.3 


176.8 


32.0 


108.5 


32.2 


klar 


16°.4 


69.8 


208.3 


31.5 


140.4 


3L9 


ivetiig trüb 


15°.9 


71.3 


250.4 


42.1 


184.4 


44.0 


klar, leiser Wind 


18°.l 


43.5 


284.8 


34.4 


216.8 


32.4 


klar, leiser Wind 


14°.5 


64.9 


324.6 


39.8 


260.3 


43.5 


klar, windig 


16°.7 


73.0 


378.4 


53.8 


322.0 


61.7 


klar, windstill. 


18°.4 


66.0 


444.4 


G6.0 


388.9 


66.9 


Hegen 


17°.6 


80.0 


484.2 


39.8 


429.2 


40.3 


klar, leiser Wind 


l.'î°.8 


91.4 


551.7 


67.5 


403.9 


64.7 


ivenig trüb 


17°.5 


86.0 


633.9 


82.2 


566.7 


72.8 


Regen' 


17°.7 


82.6 


672.4 


38.5 


607.0 


40.3 


Regen 


11°.9 


93.3 


713.1 


40.7 


625.4 


18.4 


halbklar, v:indig 


. ir.6 


86.0 


755.5 


42.4 


675.4 


50.0 


halbklar, vjindstiü. 


15°.6 


76.1 


770.5 


15.0 


738.9 


63.5 


klar, leiser Wmd 
klar, leiser Wind 

Regen 

klar, leiser Wind. 

klar 


16°.6 
17°.8 
18°.3 
14°.6 
18°.5 


81.2 
65.0 
84.8 
89.6 
78.2 










klar 


20°.7 


69.4 



24 Stunden ! Kivière fand denselben bei Phyllostachys mitis in Algier 57 cm pro 24 Stunden. 

Stnnden.*(Vergl. Pfeffer, Pflanzenphysiologie. Bd. II, p. 83.) 

sante Fragen kann ich an dieser Stelle nicht weiter eingehen (Vergl. Kraus, I.e.). 

(Ann. d. Jard. Bot. de. Bait. vol. XII. p. 197.) zusammengestellt. 



458 K. SHIBATA : 

V. Verhalten der Baustoffe 
während der Entwicklung der Schösslinge. 

In diesem Kapitel will ich die Umwandlnngs- und Wander- 
ungsvorgäuge verschiedener Baustoffe während der Entwicklung 
der Bambusschösslino;e in wesentlichen Züii;en darzustellen ver- 
suchen. 

Die Resekvestoffe. 

Unter stickstofffreien Reservestoffen, die sich in Bambus- 
pflanzen vorfinden, kommt die Stärke in erster Linie in Betracht 

Im zeitigen Herbst, wo die Ablagerung der Reservestoffe 
schon stattgefunden hat, konnte ich sie in oberirdischen und 
unterirdischen Theilen aller untersuchten Arten in wechselnden 
Mengen auffinden. Die P/iyllostachys- Arten, welche mit einem 
umfangreichen, unterirdischen Rhizomsystem ausgerüstet sind, 
pflegen nur sehr kleine Mengen der Stärke in ihren Halmparen- 
chymzellen aufzuspeichern. Bei allen Arten wird die grösste 
Menge der Stärke in Rhizomen und Wurzeln deponiert. Die 
Parenchymzellen der Knoten sind stets äusserst stärkereich und 
die Blattscheiden einiger Arten speichern ebenfalls die Stärke im 
Parenchym auf. Die Siebiöhren sind dagegen höchst inhaltarm ; 
ich habe nur selten winzige Stärkekörner in denselben nach- 
gewiesen. Es waren aber kleine Mengen von Glykose und 
Rohrzucker stets vorhanden. Beachtenswerth ist die Gestalt der 
Stärkekörner ; bei Bamhusa 'palmata, B. Veiichii und B. jjani- 
culata sind sie aus zahlreichen kleinen Theilkörnern zusammen- 
gesetzt (Fig. 57). Solche polyadelphische Stärkekörner^) kommen 
nach Nägeli in Stammgebilden nur selten vor. 

1) A. Meyer, Untersuchungen über die Stärkekörner, p. 204. 



WACHSTUM GSESCHICHTE D. BAMBUSGEWAECHSE. 459 

Der red ucie rende Zucker ist ziemlich reichlich in Halmen 
und Rhizomen im Winterzustande nachzuweisen. Die winzigen 
Fetttropfchen sind oft im Halmparenchym von Phyllostachys 
mitis, Arundinaria Simoni, Arundinaria Hinchü u.s.w. ange- 
troffen, aber sie kommen jedenfalls als Reservestoffe kaum in 
Betracht. 

Um eine Vorstellung über die Mengenverhältnisse der auf- 
gespeicherten Stärke zu anderweitigen Bestandtheilen der 
Reservestoff'behälter zu gewinnen, habe ich einige Analysen der 
zweijährigen Rhizome von Phyllostachys mitis ausgeführt^). Es 
ergab folgendes : 

% Gehalt der 
Trockensubstanz. 

Stärke 24.01 

Rediicierender Zucker 0.95 

Nicht rediicierender Zucker 4.31 

KohproteinstofFe (N x 6.25) 5.41 

Fette 0.61 

Kohfaser 47.32 

Asche 8.74 

Unbestimmte Stoffe (Differenz).., 8.65 

100.00 

1) Das am 25 November gesammelte, kräftige Kliizomstück von Phyllostachys mitis, dessen 
Parenchym sich zuvor bei microscopischer Beobachtung als von Stärke strotzend erwies, wurde 
mittelst des Hobels abgcschaubt, schnell bei 70°-80° getrocknet, und zu einem feinen Schrot 
gemahlen. Von diesem lufttrockenen Khizonischrot wurde ein bestimmtes Quantum abge- 
wogen und zu jeder Bestimmung verwendet. 

Das Trockengewicht des Schrots wurde nacli weiterem 4 stündigen Trocknen bei 100° 
(zur Gewichtsconstanz) bestimmt. 

Die Stärke wurde mittelst der Erhitzung ira S o x h 1 e t'schen Autoclave verzuckert. 

Die löslichen Kohlehydrate wurden nach 5-G maligera Ausziehen mit kaltem 
Wasser binnen 24 Stunden erschöpft. Der nichtreducierende Zucker wurde nach 
Inversion mit verdünnter Schwefelsäure bestimmt. Alle Bestimmungen der Zucker wurden 
nach Meissl- Alli hn'scher Gewichtsmethode ausgeführt. 

Der Gesammts ticks t off wurde nach Kjel da hl und der Eiweisssticks off nach 
Stutzer bestimmt. 

Die Fasersubstanz wurde durch W e e n d e r'sches Verfahren bestimmt. 

Das Atherextract wurde ohne weitei-es als Oel angenommen. 



460 



K. SHIBATA 



Man sieht also, class die Starke wolil als Hauptreservestoff 
zu betrachten ist, dagegen sind die Protein Stoffe in ver- 
hältnissmässig geringer IMenge vorhanden. 

Nun schien es mir erwiinscht zu wissen, eine wie weite Strecke 
des Ehizoms zum Auswachsen eines Schösslings dienen sollte, so 
habe ich im Anfang Februar von einer Plantation von Pliyllostachys 
bambusoides eine Anzahl Rhizomstücke ausgegraben und die auf 
Knoten vorkommenden Schösslinge (im oten Stadium) aufgezählt. 
Es ergab folgendes Resultat : 



Zahl 
der Khizomstücke 


Gesummt aiiza hl 
der Interned ien 


Halme 


Rhizomzweige 


Schösslinge 


69 


632 


5 


39 


1.5 



Aus obigem berechnete ich das Zahlenverhältniss der Khizom- 
internodien zu einem Schössling, wie 42.1:1. 



Kohlehydrate. 



Die stickstofffreien Eeservestoffe in allen untersuchten Arten 
bestehen, wie schon erwähnt, hauptsächlich aus der Stärke. Dass 
der Stärkegehalt der Rhizome eine bemerkbare Verminderung 
während des Winters erleidet, wie es von Rosenberg^) für einige 
perennierende Gewächse dargethan wurde, konnte ich nicht in 
diesem Fall bestätigen, da ich grosse Anzahl von Rhizomen von 
Phyllostachys mitis, Phyllostachys bambusoides, Phyllostachys Kuma- 
sasa, Arundinaria Hindsii, Arundinaria Narihira, Arundinaria 
quadrangular is, Arundinaria, Jl/afs?fmu,rœ, Bambusa pahnata, 
Bambusa nana u.s.w\ im Winter (Anfang Januar — Ende Februar) 



1) Rosenberg, Die Stärke im Winter. Eot. Central M . Bd. LXVI, p. 337. 



WACHSTUMGSESCHICHTE D. BAMBUSGEWAECHSE. 461 

untersuchte und dabei keine merkliclie Differenz in Bezug auf 
Stärkegelialt von den im Herbst beobachteten Exemplaren auf- 
finden konnte. Audi die Wurzehi der untersuchten Arten 
enthielten in dieser Jahreszeit grosse Mengen von Stärke in 
ihrem Rinden- und Markparenchym, so z. B. bei den am 25. 
Januar gesammelten Exemplaren : 



Jr'htjllostacliys PhyUoslachy.< Aiundinario Arnndinaria 
rnitis bambusoides ' Hindsu Nanhira 



Kiudenpareiichym 4^) 4 

Markparenchym 3 2 



Gleiches gilt für die oberirdische Halme von Arundiiiana- 
und Bambiisa- Arten.-) Merkwürdigerweise nimmt die Menge des 
reducierenden Zuckers während des Winters unverkennbar zu. 
Sodann kann man leicht mehr oder minder l)edeutende Menden 
desselben im Hnlm- und Bhizomparencliym obengenannter Arten 
nachweisen.'^) 

Aber in Stadium IV, wo die unterirdischen Schösslinge ein 



1) Bequemliclikeitslialher luvbe ich zur Eezeichnung des Stärkegehaltes folgende Ziffern 
benutzt : 

— bei gänzlicher Abwesenheit von Stärke ; 

1— wenn ein Theil des Gewebes stärkefrei ist, wälireud der andere wenige Körnchen 

in den Zellen führt ; 
2 — wenn alle oder die meisten Zellen wenige Stärkcköruer enthalten ; 
3 — wenn ein Theil der stärkeführenden Zellen wenige stärkekorner enthält, während 

der andere recht viel Stärke führt; 
4 — wenn das Gewebe recht viel Stärke enthält ; 
5 — wenn alle oder die meisten Zellen strotzend gefüllt sind. 

2) Yergl. A. Fischer, Beiträge zur Physiologie der Holzgewächse. .Jahrb. f. wiss. Bot. 
Ed. XXII, p. 92, p. 112. 

o) Dieser Zucker geht aber grösstentheils schon im Anfang März wieder verloren, ohne 
dass dabei eine bemerkbare Stärkezunahme stattfand. Auch fielen die Versuche den Zucker 
in der abgeschnittenen Halmtheilen durch künstliche Erwärmung (im Treibhaus bei 17°-20°C.) 
zur Stärke überzuführen negativ aus. 



462 



K. SHIBATA 



rasches Wachstum begannen, ist die deutliche Stärkezunahme 
in mehreren Rhizominternodien in der Nähe von Knoten, an 
welchen der wachsende Schössling sitzt, zu beobachten. So 
z.B. bei Phyllostachys îriitis : 





Anfang Ende 
November December 


Anfang 
Februar 


Anfang 
]\Iärz 


Mitte 
April 


Rinden par en chy m 

C e n t r a 1 c y 1 i n d e r p a r e n - 
chy m 

M ark pa re neb y m 


4 
2 
3 


4-3 

2 
4 


3-4 

2 

3-4 


3-4 
2 

3-4 


5 

4-5 
5 



Diese Stärkezunahme mag jedoch darauf beruhen, dass 
die von ferneren Theilen des Rhizoms in Form von Zucker 
zu2:efülirten Kohlehvdrate hier in der Nähe des Schösslings 
transitorisch in Stärke umgewandelt werden. Dafür sprechen die 
Umstände, dass erstens in entfernteren Ehizomiuternodien keine 
entsprechende Stärkezunahme stattfand, und zweitens schon in 
dieser Zeit ein Blutungssaft, der eine wichtige Rolle beim 
Zuckertransport spielt, von jeder beliebigen Schnittfläche des 
Rhizoms hervorquillt. 

Die Stärkezunahme ist vor allem im verholzten Stieltheile 
des Schösslings ausgeprägt : 





Ende 


Anfang 


Anftmg 


Mitte 




December 


Februar 


März 


April 


E i n d e n p a r e n c h y m 


2-3 


2 


5-4 


5 


C e n t r a 1 y c li n d e r p a r e n - 
chy m 


3 


2 


4-3 


5 


Ha drom p are ncliy m 








0_, 






Gleichzeitig wurde die partielle Entleerung der Rhizomknoten, 
an welche die Schösslinge sitzen, beobachtet, obgleich die nächst 
folgenden Internodien, wie schon bemerkt, noch von Stärke 
erfüllt waren. 



WACHSTUMGSESCHICHTE D. BAMBUSGEWAECHSE. 



463 



Von jetzt ab wird die Starke im Rhizome nach und nach 
aufgelöst und sclion in dem Stadium V, wo die Schössliuge auf 
der Erde 4-6 Meter hoch wuchsen, verschwinden fast sämmtliche 
Stärkekörner vom Parenchym der benachbarten Rhizominter- 
nodien. So zum Beispiel bei Phyllodachys mitis : 
R h i z m i 11 1 e r n d i e n — 





16. April 


9. Mai 


19. Mai 


Hill denpareiichy m 

C e n t r a 1 c y 1 i n d e r p a r e n c h y m 

j\I a r k p a r e n c h y m 


•5 
5 

5 


1 

1-2 

4-3 




1 


Xodium — 




16. April 


9. Mai 


19. Mai 


Subepidermale aclerotische 
Schicht 

Eindenparenchym 

C e n t r a 1 c y 1 i n d e r p a r e n c h y m 


4 
4 
1-2 














Stiel des Schösslings — 










16. April 


9. Mai 


19. Mai 


Subepiderraale sclerotische 
Schicht 

Eindenparenchym 

C e n t r a 1 c y 1 i n d e r p a r e n c h y m 



5 

•5 




1-0 











Hier findet auch in der Wurzel eine entsprechende Stärkeent- 
leerung statt : 



16. April 


19. Mai 


Binden- fäusseres grosszelliges 
parencbymii„j,e,.es kleinzelliges 

M a r k p a r e nch y m 


5 
5 

.5-4 



1-2 





464 K. SHIBATA : 

Merkwürdigerweise konnte ich bei so raschem Auflösen der 
Stärke eine entsprechende Glykosebildung im Parencbj^n nicht 
beobachten; bei der Zuckerprobe nach Scliimper erhielt ich 
nur Spuren von Oxydulkörnern in Parenchymzellen. 

Die Kohlehydrate in wachsenden Schösslingen verhalten 
sich im Grossen und Ganzen analog mit denjenigen in von 
Sachs, H. de Vries u. A. untersuchten Pflanzen.^) Indess 
ist folgendes noch zu bemerken. 

Die Abwesenheit von Stärke im Urmeristem des Vegetations- 
punktes habe ich im allgemeinen constatiert. Die feinkörnige 
Stärke wird erst an der Stelle, wo die erste Differenzierung der 
Bündelanlage und des Grundparenchyms auftritt, nachweisbar 
und man kann abwechselnd stärkereicbe und stärkearme Zonen 
deutlich sehen. Nach unten tritt der Unterschied im Stärke- 
gehalt dieser abwechselnden Zonen immer schärfer hervor. Der 
reducierende Zucker tritt an der Spitze weiter unten als Stärke 
auf und zwar zuerst in dem Marke der internodialeu Zone, wo der 
erste Anfang der Zellstreckung sich durch Zerreissen von Gewebe 
kund thut. Selbst die fertig gestreckten, unteren Internodien des 
mehrere Meter hohen Schösslings bleiben noch lange Zeit von der 
Glykose erfüllt. Sobald die Streckung eines Internodiums voll- 
endet ist, verschwindet die Stärke aus dem Parenchym, abgesehen 
von einigen winzigen Körnchen in 2-3 Zellen bei Durchlassstellen 



1) Hier seien nur folgende erwähnt : 

Sachs, Physiologische Uutersuchiuig üb. d. Keimung von Schminkbohue. — Sachs, 
Keimungsgeschichte der Gräser. — De Vries, Wachstumsgeschichte der Zuckerrübe. — 
De Vries, Keimungsgeschichte der Kartofîëlknolleii.— Del mer, Vergl. Pliysiologie 
d. Keimungsprocess der Samen.— H oilman u. Über d. Stofiwanderung bei d. Keimung 
von Weizen- und Kleesamen. — A. F. A. C. Went, Chemisch-physiologische Unter- 
suchungen üb. d. Zuckerrohr. 

2) Vergl. Sachs, Üb. d. Stoffe welche d. Material z. Wachstum d. Zellhäute liefern. 
Jahrb. f. wiss. Bot. Ed. Ill, p. '207. 



WACHSTUMGSESCHICHTE D. BAMBUSGEWAECHSE. 465 

der Bündel. Da liierbei sämmtliclie Zellwände noch keine 
nennen swertlie Verdickung zeigen, so erfolgt die weitere Aus- 
bildung der Bündelelemente ohne Gegenwart der umgebenden 
Stärkescheide, in welcher ncicli Heine^) die nöthigen Baustoffe 
als Stärke deponiert werden sollen. Die besonders starke Zucker- 
ansammlung in einigen Parenchymschichten um die in Ausbildung 
beo-riffenen Bastbeleo-e herum vertritt hier die Stelle der fehlenden 
Stärkescheide und da lier mag sie als Zuckers eh ei de") bezeichnet 
werden. 

In der wachsenden Wurzel bemerkt man die kleinste Menge 
der winzigen Stärkekörner nur in der noch zartwandigen Endo- 
dermis. Hingegen ist in der Haube von der ersten Anlage die 
Stärke in ihren Zellen festgehalten.'') Der reducierende Zucker 
kommt in der ganzen Länge der Wurzel, ausser der 4-5 mm 
langen Strecke der Spitze und der Haube, reichlich vor. Erst in 
der ca. 40 cm lang gewachsenen Wurzel wird die Abnahme und 
zuletzt das Verschwinden vom Zucker an der AVurzelbasis 
bemerkbar. 

Wie schon erwähnt konnte ich in Rhizomen und Wurzeln, 
wo die Eeservestärke in Auflösung begriffen war, gewöhnlich nur 
eine Spur von Glykose auffinden. Analoge Fälle sind bereits 
bekannt. So z.B. 2:elan2;te es Sachs nicht, in Cotvledonen der 
keimenden Phaseolus- Samen, in Schildclien von Triticinn und 
Zea und auch in Funiculus verschiedener Samen die Glykose 
nachzuweisen^), obgleich hier das Vorhandensein der gelösten 



1) Heine, Die pliysiologische Eedontiing der sogenannten vStärkescheide. Landw. 
Vcrsuchs-St. 1888. p. 115. 

2)H. de Vries li;it frülier den Ausdruck im analogen Sinne mit „ Ijeitselieide '' 
S c h i m p e r's angewandt. 

3)Vergl. Sachs, Jahrb. f. wiss. Eot. Ed. Ill, p. 203. 

4) Sachs, Über die Stoffe, welche das Material zum Waclistum der ZelUiüute liefern. 
Jahrb. f. wiss. Eotauik. Ed. Ill, p. 248- 



466 K. SHIBATA : 

Kohlehydrate von vornherein erwartet werden musste. In 
gewissen Fällen dieser Art ist es nicht unwahrscheinlich, dass 
die Glykose durch andere lösliche Kohlehydrate ersetzt wird/) 

In treibenden Rliizomen von Phyllostacliys mitis habe ich 
nun den Kohrzucker in stärkehaltigen Parenchymzellen mittelst 
der Invertin-Methode nachgewiesen. Ferner in Rhizomen, von 
welchen fast alle Stärkekörner schon verschwunden waren 
(Stadium V), beobachtete ich noch erhebliche Mengen Rohrzucker. 
Uebrigens habe ich in folgenden Arten den Rohrzucker in Rhizo- 
men während des Austreibens der Schösslinge beobachtet : Phyllo- 
stacliys bambusoides, Phyllostacliys iniberula und Arundinaria 
japonica; und in folgenden im Halmparenchym w^ährend des 
Austreibens der Zweigknospen : Bambusa palmata und Arundi- 
naria japonica. Ferner erhielt ich die Rohrzucker-Reaction in 
Halmen von Phyllostachys lunnasasa, Arundinaria, Simoni, und 
Arundinaria Hindsii var. yraminea. In den Wurzeln von 
Phyllostachys mitis, deren grosskörnige Reservestärke in Auflösung 
begriffen war, konnte ich ebenfalls Rohrzucker nachweisen. 
Gleiches gilt für verholzte Stieltheile der Schösslinge. In allen 
diesen Fällen kommt Rohrzucker hauptsächlich im Parenchym 
und viel w^einiger in Siebröhren vor. Nnn liegt mir der Gedanke 
nahe, dass in diesem Falle die Kohlehydrate hauptsächlich in 
Form des Rohrzuckers von Zelle zu Zelle wandern"). 

Vom Rohrzucker ist noch zu erwähnen, dass ich ihn im 



1) So z. B. wurde für Graniinecn-.Scutelhim das Vorliandensein vom Rohrzucker anstatt 
Glykose von Grüss auf microcliemiscliem Wege so\vie_auf ex[)erimentelle Weise sicherge- 
stellt. (Vergl. Ber. d. D.'B. G. Bd. XVI, p/lT.)- Puriewitsch (Jalirb. f. wiss Bot. Bd. 
XXXI, p. 53.) hat bei der ersten Periode der Entleerung der Eeservestärke das Auftreten 
nichtreducierenden Zuckers beobachtet. Vergl. Leclare du Sablon, Recherche sur les 
Reserve Hydrocarbones des Bulbes et des Tubercules. Rev. gen. d. Bot. 18Ö9. 

2) Vergl. E. Schulze, Ueber die Verbreitung des Rohrzuckers in den Pflanzen und 
über feine physiologische Rolle. Zeitschrift f. physiol. Chemie. Bd. XX, p. 552. 



WACHSTUMSGESCHICHTE D. BAMBUSGEWAECHSÈ, 467 

jungen Gewebe unterhalb des Urmeristems, wo schon eine 
Differenzieruno; in nodiale und internodiale Zonen statto-efunden 
hat, manchmal, wenn auch nicht immer, durch Inverti n- 
Methode nachgewiesen habe^). 



EiwEiss UND Amidoverbindungen. 

Im Jahre 1872 hat Pfeffer-) zuerst die hohe Bedeutung des 
Asparagins in der Translocation und Bildung von Eiweiss beim 
Keimen von Lupiniis luteus und einigen anderen Papilionaceen 
auf microchemischem Wege nachgewiesen. Er hat nämlich con- 
statiert, dass das Asparagin als Auflösungsproduct des Reserve- 
proteins in Cotyledonen entsteht und dann wachsenden Theilen 
zugeführt wird, und ferner, dass wenn Kohlehydrate bei der 
Assimilation sich vermehren, das gebildete Asparagin zum Eiweiss 
regeneriert wird^). Die letzterwähnte Thatsache hat er später 
durch die Kulturversuche im Dunkeln und in kohlensäurefreier 
Luft weiter begründet.^) Seit diesen zum ersten Male exact 
ausgeführten Arbeiten Pfeffer 's wurde die Frage nach Eiweiss- 
umsetzungen mit immer wachsender Eifrigkeit seitens der 
Botaniker und Chemiker verfolgt, was zu zahlreichen Arbeiten 
Veranlassung gab. Schulze und seinen Schülern verdanken 
wir besonders eine Beihe der Versuche über jene Amidover- 



1) Ich habe im Gewebe dieser Kegion eiae schöne rosarothe Färbung mit conc. 
Schwefelsäure erzielt. Diese Reaction deutet darauf hin, dass hier ein lösliches Kohlehydrat 
neben Eiweiss vorhanden ist. Vergl. Frankfurt, Zur Kenntniss der chemischen Zusam- 
mensetzung des ruhenden Keimes von Triticum tu^gare. Landw. Versuchs-St. 1896. p. 4G1. 

2) Pfeffer, Untersuchungen über die Proteinkürncr und die Bedeutung des Asparagins. 
Jahrb. f. wis>. Bot. Bd. VIII, p. 429. 

3) Pfeffer, I.e. p. 558. 

4) Pfeffer, tjber die Beziehung de^ Lichtes zur Regeneration von Eiweissstoffen aus dem 
beim Keimungs process gebildeten Asparagin. Monatsber. d. Acad. d. Wiss. z. Berlin. Dec. 
1873. 



468 K. SHIBATA : 

bindimgen und Hexonbasen, wie Glutamin, Leucin, Tyrosin, 
Phenylalanin, Arginin u.s.w., die neben Asparagin beim 
Eiweissumsatz auftreten. Namentlich hat Schulze schon in 
einer im Jahre 1878 j)ublicierten Arbeit die Ansicht geäussert, 
dass in Lupinenkeimlingen andere Nichteiweiss- Verbindungen, 
die neben Asparagin auftreten, auch zur Regeneration des 
Eiweisses dienen müssen^). In demselben Jahre hat Borodin 
eine allgemeine Verbreitung von Asparagin im Pflanzen- 
reiche festgestellt, dabei sprach er aus : ,, Sobald irgend ein 
lebenskräftio-er Theil iro;end einer Pflanze arm an Stickstoff- 
freien Substanzen wird, sieht man in ihm Asparagin als 
Zersetzungsproduct des Eiweisses auftreten und sich mit der 
Zeit immer mehr anhäufen. "■) Demnächst fand Kellner^) in 
jungen Theilen der Gräser eine bedeutende Menge Amide, und 
äusserte zuerst die Ansicht, dass die Amide durcli Synthese aus 
anorganischen S tickstoffverbin düngen entstehen. Auch Horn- 
berger^) meinte, dass die Amide, die in Maiskeimpflanzen auf- 
treten, synthetische Producte seien. Suzuki"') hat angegeben, dass 
er bei Einführung von anorganischen Salzen wie Ammoniumuitrat 
und Natriumnitrat in verschiedenen Pflanzen eine Asparagin- 
bildung bewerkstelligen konnte. Emmerling'^^) hat auch Amido- 
säuren als synthetische Producte angesehen, aber es fehlt an 
einem experimentellen Beweis. So kann die Bildung des Aspara- 

1)E. Schulze, Über Zersetzung und Neubildung der Eiweissstoffe bei der Keimung von 
gelber Lupine. (Jahresber. f. Agr. Chem. 1878. p. 211). 

2) Borodin, Über die physiologiäche Rolle und die Verbreitung des Asparagins in 
Pflanzenreich. Bot. Zeit. 1878. p. 826. 

3) Kellner, Landw. Jahrbücher. Bd. VIII. Suppl. 1879. 

4) Hornberger, Chemische Untersuchung über das "Waclistum der Maispflanze. Landw. 
Jahrb. 1882. 

5) Suzuki, On the Formation of Asparagin in Plants under different Conditions. Bull, 
of the College of Agriculture. Bd. II, p. 409. 

6) Eramerling; Studien über Eiweissbildung in der Pflanze. Landw. Versuchsst. 1887. 
p. 7. 



WACHSTÜMSGESCHICHTE D. BAMBUSGEWAECHSE. 469 

gins und der anderen Amidokörper entweder durch Zerfall des 
Eiweisses oder durch geeignete Synthese erfolgen. Ob diese 
oder jene geschieht rauss von Fall zu Fall bestimmt Averden. 
Jedenfalls ist es seit Pfeö'er's bahnbrechender Untersuchung 
klar, dass die Amide und Amidosäuren, deren Entstehungen in 
verschiedenen Fällen verschieden sein können, nachher für 
Ei Weissregeneration verbraucht werden. Gegenwärtig ist es aber 
noch nicht sicher ob verschiedene Amide und Amidosäuren ganz 
gleichwerthig für Eiweissregeneration dienen. Zwar hat Hans- 
teen^) in seiner interessanten Arbeit gezeigt, dass bei Lemna 
minor verschiedene, künstlich eingeführte Amide und Amido- 
säuren je nach der Qualität der disponiblen Kohlehydrate sich 
für Eiweissbildung verschieden verhalten. So liegt der Gedanke 
nahe, dass die in bestimmten Keimpflanzen auftretenden Amido- 
körper auch ungleichen Werth für Eiweissbildung besitzen. 
Früher war Schulze-) der Meinung, dass das Asparagin schwerer 
verwendbar als andere Amidokörper ist und daher in Keim- 
pflanzen zur Anhäufung kommt. Aber Loew") behauptete, dass 
das Asparagin dem Ei weiss näher steht als andere Amidokörper, 
und vermuthete auch, dass die letzteren w^eiter zerfallen unter 
Bildung von Formaldehyd und Ammoniak, aus denen durch 
synthetische Processe Asparagin entsteht. Erst neulich ist 
Schulze^) zu einer ähnlichen Vorstellung gelangt. Er spricht 
die Ansicht aus, dass das Asparagin (und auch Glutamin) in den 



l)PIansteeu, Beiträge zur Keuntniss der Eiweissbildung und die Bedingung der Kea- 
lisirung. Ber. d. D. B. G. Bd. XIV, p. 362. 

2) Schulze, Über den Eiweissumsatz im Pflanzenorganismus. 1880. p. 30. 

3)0. Loew, The Energy of living Protoplasm. Bulletin of the College of Agriculture 
Bd. II, p. 64. 

4) Schulze, über den Umsatz der Eiweissstofle in den lebenden Ptlauzeu. Zeit. f. 
physiol. Chemie. Bd. XXIV, p. 60. 

Schulze, IJber die Bildungsweise des Asparaglns iu den Püanzen. Landw. Jahrb. 
1898. p. 509 ; p. 513. 



470 K. SHIBATxV : 

Keimpflanzen zum grossen Theil durch Umwandlung der Amido- 
säureu, die als directe Eiweisszersetzungsproducte betraclitet 
werden können, entstehen, und dass die Amidosäuren einmal zu 
leicht verwendbaren Amiden^) übergeführt werden, bevor sie sich 
in Eiweiss verwandeln.") Bei dieser Sachlage ist es wünschens- 
^verth im conereten Falle die Localisation und das Verhalten 
von Amiden und Amidosäuren zu verfolgen und damit einiger- 
massen Aufschlüsse über die Beziehung zur Eiweissregenera- 
tion der beiden verschiedenen Stoffe zu gewinnen. 

Bei dem vorliegenden Falle der Entwicklung der Bambus- 
schösslinge kommen Ty rosin und A spa rag in reichlich vor. 
Die genannten Vertreter von beiden Stoffgruppen sind glück- 
licherweise leicht auf microchemischem Wege bestimmbar. 

Hier lasse ich ältere Angaben über das Vorkommen des 
Tyrosins vorangehen. Gorup-Besauetz^) fand es zuerst im 
Wickenkeimlinge. Schulze und Barbie ri^) fanden es in etwas 
grösserer Menge in Kürbiskeimlingen. Auch in Lupinenkeim- 
liugen scheint es nicht zu fehlen, da Beiz un g') aus Extract der 



1) Eine entgegengesetzte Meinung, dass das Asparagin ein für Eiweissregeneraiion wenig 
geeignetes Material sei, wurde neuerdings wieder von Prianisclin iko w (Landw. Versuclis- 
St. 1899. Bd. LH, p. 347 ff.) vertreten. 

2) Unter neueren Publicationen über Eiweisssyn these, die nach Vollendung meines 
Manuscriptes in meine Hand gelangten, seien nur folgende zu erwähnen : 

Prianischnikow, Eiweisszerfall und Athmung in ihren gegenseitigen Verhältnissen. 

Landw. Versuchs-St. Bd. LH, p. 137. 
Hansteen, Über Eiweisssynthese in grünen Phanerogamen. Jahrb. f. wiss. Bot. 

Bd. XXXHI, p. 417. 
Prianischnikow, Die Eückbildvmg der Eiweissstofle aus deren Zerfallsprcduclen. 

Landw. Versuchsst. 1899. p. 347. 
Schulze, Über Eiweisszerfall und Eiweissbildung in der Pflanze. Per. d. B. G. 1900. 

Heft. 2. p. 36. 
Emmerling, Studien über die Eiweissbildung in der Pflanze. Landw. Versuchs-St. 

Bd. LIV, p. 215. 

3) Gorup-Besanetz, Ber. d. D. C. G. VIT, p. 146; ]>. 5G9. 

4) Landw. Jahrb. Bd. VH, p. 431. 

5) Beizung, Kecherche sur 1. Germination etc Ann. d. Sc. nat. Bot. Scr. VlI, T. 15. 
p. 231. 



WACHSTÜMSGESCHICHTE D. BAMBUSGEWAECHSE. 471 

Keimlinge von Lypinus luteus Tyrosinkrystalle isolieren konnte, 
obwohl ihm der mierochemische Nachweis des Ty rosins nicht 
gelang. Ferner fand es Schulze^) in Cotyledonen keimender 
Lupinusàv teil, etiolierten Keimlingen der Lupinus angudifolius, 
Endosperm von Ricinus communis und etiolierten Pflanzen von 
Tropaeolum m.ajus. In allen diesen Fällen ist die Menge des 
gefundenen Ty rosins immer sehr gering, so dass man auf micro- 
chemische Verfokuno- desselben verzichten muss. Schulze 
bemerkte, dass der Grund des geringen Vorkommens von Tyrosin 
darin liegt, dass es eine viel regere und schnell verlanfende Um- 
wandlung erleidet.-) In unterirdischen Pflanzentheilen ist Tyrosin 
öfters auf chemischem Weoe oefunden. Schulze und Barbieri 
fanden Tyrosin neben Leucin in den Kartofifelknollen und in der 
Wurzel von Bela vulgaris.^) Auch Planta^) fand es in den 
Knollen von Stachys tuherifera. In der botanischen Litteratur 
finden wir nur vereinzelte Angaben. PrantP) bat Krystalle, die 
wie Tyrosin reagierten, aus in Alcohol aufbewahrten Stengeln von 
Dahlia variabilis erhalten. Borodin*^) fand in Blättern der 
etiolierten Kartoffel, die mit absolutem Alcohol behandelt wurden, 
Tyrosinkrystalle. Ferner fand er dergleichen in Vida saliva, 
Tropaeolum majus etc. Aber es ist hier zu bemerken, dass diese 
Befunde ausschliesslich von abgeschnittenen und in Wasser weiter 
cultivierten Zweigen herrührten und gleichzeitige chemische 



1) Schulze, Üb. d. Umsatz d, Eiweissstofie in d. leb. Pflanze. Zeit. f. physiol. Chemie. 
Bd. XXIV, p. öS. 

2) Schulze, I.e. p. 50. 

3) Vergl. Schulze, Über den Eiweissumsatz im Pflauzcnorganismus. ISSO. p. 24. 

4) Planta, Über die Zusammensetzung der Knollen von Stachijs (uberiftra. Landw. 
Versuchs-St. Bd. 35, p. 473. 

Vergl. ferner Schulze, Zeits. f. physiol. Chemie. Bd. XXIV, p. 85. 
5)Prantl, Das Inulin. 1870. p. 61. 

6) Borodin, Über die physiologische Kolle und die Verbreitung des Asparagins. But. 
Zeit. 1878. p. 819. 



472 K. SHIBATA : 

Belege fehlten. Erst später hat er^) einmal in normalen, jungen 
Dahlia-JMiitteYn Tyrosin aufgefunden. Noch später hat Leitgeb") 
den Gehalt der Dahlia-KnoWen an Asparagin und Tyrosin 
constatiert. 

Bevor ich zur Besprechung meiner Beobachtungen fortsch- 
reite, will ich hier die Ergebnisse von chemischen Untersuchun- 
gen Kozai's") kurz erwähnen. Er hat die Analyse des Schöss- 
lings (Stadium IV) von Phylloslachys mitis ausgeführt ; sie ergab 
folgendes : 

?i Gehalt der 
Trockensubstanz. 

E h p r ü t e i n S 1 ff e 25.12 

Fette 2.49 

Kohfaser 11.60 

Stärke 3.33 

aiykose 8.15 

Andere N-frcie ext. Stoffe 30.49 

Asche 9.22 

Unbestimmbare StotFe 9.G0 

100.00 

Für die Yertheilung des Stickstoffs auf Proteinstoffe und 
nichtproteinartige Verbindungen ergaben sich folgende Zahlen : 

N in Proteinstoöen 1.22*^^ der Trockensubstanz. 

N in nichtproteinartigen Stoffen ...2.82i^ „ „ 

Gesammtstickstoff" 4.04^^ „ ,, 

So sieht man, dass die Schüssliuge grosse Mengen von 
stickstoffhaltigen Substanzen enthalten, im auffallenden Gegen- 

1) Boro din, Über einige bei Bearbeitung von Pflanzenschnitte mit Alcohol entstellende 
Niederschlag. Bot. Zeit. 1882. p. 589. 

2) Leitgeb, Der Gehalt der DahliaknoUen an Asparagin und Tyrosin. Mittheil. a. d. 
bot. Inst. z. Graz. 1888. p. 222. 

3)Kozai, Ou the nitrogenous non-albuminous Constituents of Bamboo shoots. Bulletin 
of the College of Agriculture. Vol. I. No. 7. 



WACHSTUMSGESCHICHTE D. BAMBUSGEWAECHSE. 473 

satz zum Eliizora, und insbesondere kommen die uicliteiweissar- 
tigen Verbindungen in überwiegender Quantität vor. Kozai hat 
nach dem Schulze'schen Quecksilbernitratverfabren die seiden- 
glänzenden Nadelkrystalle aus dem Wasserauszug von Schösslingen 
erhalten, welche mit Sicherheit mit Tyrosin identificiert wurden. 
Ferner hat er auch das Asparagin isoliert und durch verschie- 
dene Reactionen und Bestimmung der Stickstoffzahl sicher nach- 
gewiesen. 

Bei meinen Studien wurden die obengenannten Substanzen, 
das Asparagin und das Tyrosin in ihrem Verhalten näher verfolgt. 
Beide sind nach Borodin 'scher Methode reichlich und sicher 
nachweisbar, dabei scheint der vorhandene Zucker kein Hinder- 
niss zur Krystallisation darzubieten. 

Zunächst will ich das Verhalten von Asparagin und Tyrosin 
bei der Entwicklung der Schösslinge von Phyllostachys mitis kurz 
angeben. 

Ich konnte weder Tyrosin noch Asparagin im urmeristema- 
tischen Gewebe der ganz jungen Knospen (Stadium I) finden, 
während in deren basalen, von Bündelanlagen durchsetzten Theilen 
Tyrosin schon regelmässig vorkommt. Nun die Schösslinge 
nehmen sehr langsam an Grösse zu und ihre Stieltheile werden, 
wie schon erwähnt, allmählig verholzt. Ich beobachtete, dass das 
Tyrosin mit der Zeit im Schösslingskörper erscheint und seine 
Menge immer grösser wurde, zugleich auch das Asparagin in 
nachstehender Menge. AVenn man einen 4-") cm langen Schössling 
in diesem Stadium (Stadium II) untersucht, so sieht man 
folgendes : Der Vegetationspunkt bleibt frei von Amidosubstanzen. 
Erst 2-3 mm unten, wo die Bündelanlagen schon differenziert 
w^aren, erscheint die erste Spur von Tyrosin im parenchymatis- 
chen Gewebe. Asparagin tritt noch w^eiter unten ein, ayo Zucker 



474 K. SHIBATA : 

in grÖ3serer Menge vorkommt (etwa in der Mitte von der ganzen 
Länge des Schösslings), daneben viel Tyrosiii. Zuletzt fand ich 
im Stieltheile keine Amide mehr. So coincidiert Asparagin in 
seiner Localization fast mit reducierendem Zucker. Pfeffer^) 
bemerkte schon derartiges Zusammentreffen von Traubenzucker 
und Asparagin in der ersten Periode der Keimung von Liqyinus 
luteus. In diesem und auch im folgenden Stadium konnte ich 
weder Tyrosin noch Asparagin im Phizom nachweisen. 

Das oben definierte Stadium III wird im Laufe des Sommers 
erreicht. Während dieser Zeit nimmt die absolute Menge des 
Tyrosins sowie des Asparagins immer mehr zu, so dass die 
Krystalle des Tyrosins und des Asparagins unter dem Microskop 
in grösserer Menge und viel leichter gefunden werden ; die Be- 
handlung der Gewebe (die aber eiweissarm sind) mit Mil Ion 's 
Reagens bringt überall eine tiefere Färbung als in den vorigen 
Stadien. Die Vertheilung des Tyrosins und des Asparagins 
stimmt im AVeseutlichen mit der des vorigen Stadiums überein. 
Dabei ist noch zu bemerken, dass das Tyrosin weniger in Nodien 
als in Internodien vorkommt. In diesem Zustande überwin- 
tern die Schösslinge ohne bemerkbare Veränderung bis Ende 
Februar. Von jetzt ab erwacht ein regerer Process im Schöss- 
linge, und von Anfang — Mitte April wächst es schon zu einer 
beträchtlichen Grösse unter der Erde. Die Schösslinge in diesem 
Stadium (Stadium IV) werden auf dem Markt als Gemüse feil 
geboten. Die oben angegebene analytische Bestimmung Kozai's 
rührt auch von einem solchen Schösslinge her. In diesem Stadium 
bemerkte ich folgende Vertheilung : Der Vegetationspuukt ist 
frei von Amidokörpern, dagegen reich an Eiweiss, aber in jungen 



1) Pfeffer, tJber die Proteinkörne r und die Bedeutung des Asparagins. Jahrb. f. wiss. 
Bot. Bd. VIII, p. 539. 



WACHSTUMSGESCHICHTE D. EAMBUSGEWAECHSE. 475 

Scheideblättern, die 7a\ dieser Eegion gehören, lassen sich stets 
kleine Mengen Tyrosins nachweisen, daher muss man bei Fest- 
stellung der Abwesenheit der Amidosubstanzen in der Spitze sie 
thunlichst von Scheideblättern befreien. Das Eiweiss, welches 
Biuretreaction giebt, ist in dieser Region besonders reichlich in 
Procam bialsträngen nachweisbar. Nach Sachs^) wnrd das Eiweiss 
durch diese Gewebe dem Urmerislem zugeführt, und da ich hier in 
der Spitze keine Amide auffinden konnte, so kann die Wanderung 
des Eiweisses wohl in dem Sachs 'sehen Sinne geschehen. Das 
in dieser Weise von unten zugeführte Eiweiss befindet sich unter- 
halb des Urmeristems räumlich getrennt von Stärke in regel- 
mässig abwechselnden Zonen von je ca. 0.15 ram Dicke. Diese 
Zonen deuten schon zukünftige Internodien und Nodien an, und 
befinden sich in den ersteren Eiweiss und in den letzteren Stärke. 
Erst 4 mm unter dem Vegetationspunkt tritt die erste Spur von 
Tyrosin auf und nach unten nimmt es immer in den parenchy- 
matischen Zellen an Menge zu. Zugleich ist die Abnahme des 
Eiweisses in parenchymatischen Zellen leicht constatierbar. Das 
Asparagin kommt noch weiter unten (ca. 1-1.5 cm unter dem 
Vegetationspunkt) fast gleichzeitig mit reducierendem Zucker zum 
Vorschein. Da dicht unter dieser Region die erste Zerreissung 
im Markgewebe, die den ersten Anfang der Markhöhle andeutet, 
stattfindet, so soll hier die Zellstreckung erst recht ausgiebig 
geworden sein. Nach unten nimmt die Menge des Tyrosins und 
des Asparagins stetig zu, und dabei übertrifft die Menge des Tyro- 
sins bedeutend die des Asparagins. Am reichlichsten findet man 



1) Sachs, Über die Leitung der plastischen Stoffe durch verschiedene Gewebeformen. 
Flora. 1863. 

Es ist bekannt, dass das Eiweiss unter Umständen die Cellulosemembran hindurch 
diosmiren kann. Yergl. Puriewitsch, Physiol. Unters, üb. Entleerung der Eeserve- 
stoffbehälter. Jahrb. f. wiss. Bot. XXXT, p. 68; Pfeffer, Pflanzen physiologic. Bd. I, p. 613. 



476 K, SHIBATA : 

Tyrosin an Stellen, wo die Wurzelanlagen zur Bildung kommen,^) 
so dass Tyrosin beim Schneiden des Gewebes mit dem Messer 
sofort im Zelleinneren zu Krystallen erstarrt.^) In den Bastzell- 
anlagen der Gefässbündel, welclie noch keine Wandverdickung 
zeigen, kommt das Tyrosin bedeutend reicliliclier als im Parencliym 
vor. Die Ueberreste des Markparenchyms enthalten nur sehr 
wenig Tyrosin. Das Mill on 's Reagens bewirkt stark blutrothe 
Färbung des Zellsaftes, entsprechend dem hohen Gehalt an 
Tyrosin. Jedenfalls hat die absolute ]\Ienge des Tyrosins im 
Vergleich mit den vorigen Stadien bedeutend zugenommen. In 
dieser Region dagegen konnte ich das Asparagin nur mit Sch- 
wierigkeit auffinden. Es sei noch hervorzuheben, dass das 
Asparagin in der Regel in Knoten und Diaphragmen sich nicht 
befindet, dagegen fehlt es hier an Tyrosin nicht. 

In den untersten Internodien, wo die Verholzung der Bast- 
elemente schon eingetreten ist, verliert sich auch das Tyrosin. 

Im Laufe des Aprils durchbrechen die Schösslinge einer nach 
dem andern die Erde und wachsen ungemein rasch in die 
Länge. Es ist nicht zu bewundern, dass in so schnell wachsenden 
Pfianzentheilen ein ausgiebiger Eiweissumsatz vor sich geht. 
Tyrosin und Asparagin sind sehr reichlich in den oberen wach- 
senden Internodien vorhanden, mit gleicher Vertheilungsweise 
wie im vorigen Stadium, d.h. Tyrosin tritt ca. 2 cm unter dem 
Vegetationspunkt auf und Asparagin ca. 4 cm unter demselben 
gleichzeitig mit reducierendem Zucker. Aber sehr interessant ist 
die Vertheilungsweise in halberwachsenen Internodien. Nämlich 
in der unteren w^eichen Wachsthumszone eines solchen Interno- 
diums befindet sich das Asparagin ziemlich viel neben reichlichem 



1) ca. lote lutei-nodium von unten. 

2) Siehe unten p. 482. 



WACHSTUMSGESCHICHTE D. BAMBUSGEWAECHSE. 477 

Tyrosin. Hingegen der obere, schon erwacliseue Theil desselben 
Internodiums enthält kein Asparagin, aber Tyrosin in einer 
nahezu gleich grossen Menge wie im unteren weichen Gewebe. 
Hier werden Tyrosinkrystalle in jungen Bastzellen, die eben die 
erste Verdickung begonnen haben, reichlich ausgeschieden, aber es 
befindet sich viel weniger in Parenchymzellen. Ferner enthalten 
die Hadrora- und Leptomelemente niemals Tyrosin. Die Xodien 
und Diaphragmen enthalten weniger Tyrosin als in Internodien 
und gewöhnlich kein Asparagin. In den eben im Wachstum 
vollendeten Internodien kommt Tyrosin noch in fast gleich 
grosser Menge vor, aber Asparagin nicht mehr. In weiter unten 
liegenden älteren Internodien und Podien verschwindet allmählig 
auch das Tyrosin, und mehrere ganz erwachsene und in Verhol- 
zung begriffene Internodien auf der Erdoberfläche sind durch- 
gehends von Amidokörpern frei, obwohl sie noch reichlich die 
Glykose, wie schon bemerkt, im Parenchym aufspeichern. 

Von den Scheideblättern habe ich hier nur zu erwähnen, 
dass sich in der basalen, weichen Wachsturaszone jedes Blattes 
ziemlich viel Asparagin neben reichlichem Tyrosin befindet, und 
das erstere verliert sich schon an der Uebergangsstelle zu harten 
Theilen, während Tyrosin noch w^eiter oben im Parenchym der 
erwachsenen Spreitentheile reichlich vorkommt. So bemerkt 
man hier ein ganz ähnliches Verhältniss wie im Halm. 

In einige mm hoher Wurzelanlage, sowohl in Periblem wie 
in Plerom, lässt sich fast kein Tyrosin nachweisen, obwohl dicht 
darunter liegendes Knotengewebe an demselben reich ist. In 
verschieden laugen wachsenden Wurzeln ist die Spitze stets 
tyrosinfrei, und erst 1.5-2 cm unten ist eine Spur nachweisbar. 
Allerdings kommt Tvrosiu nur in sehr kleiner Menge im 
Wurzelparenchym vor, so dass die microchemische Nachweisung 



478 K. SHIBATA : 

immer schwierig ausführbar ist. Noch spärlicher kommt 
Asparagin in wachsender Kegion vor. Diese Umstände können 
zum Theil dadurch erklärt werden, dass die Wurzeln nur lang- 
sam wachsen und demgemäss hier der ausgiebige Eiweissumsatz 
nicht stattfindet. 

Die mit dem oben angegebenen ganz übereinstimmende 
Vertheilungsweise des Asparagins und des Tyrosins habe ich auch 
in Schösslingen folgender Arten constatiert : 

Phyllostachys bapibusoldes, 

Phyllodachys puberula, 

Bamhu^ia fcdmata. 
In den von mir in dieser Beziehung untersuchten Arundi- 
naria- Arien, nämlich : 

Ärundinaria japonica, 

Arundincv) 'ia q u adranyidaris, 

Ärundinaria MatsumurWy 

A riindinaria Hindsii, 
zeigte Asparagin auch das gleiche Verhalten, während ich 
Ty rosin nur schwierig auffinden konnte, in auffallendem Gegen- 
satz zu Fhyllosiachys- Avteu. Wie dies zu Stande kommt ist mir 
unbekannt. 

Ferner ist hier zu bemerken, dass ich in Hhizomen von 
Bamhusa pahnata Asparagin in geringer Menge nachweisen konnte, 
während es mir bei Phyllostachys- Äxten nicht gelang. 

Aus dem oben erörterten Befunde lasse ich folgende vier 
Sätze gelten, nämlich : 

1. Ty rosin übertrifft Asparagin in Menge. 

2. Tyrosin tritt in der Nähe vom Vegetationspunkt auf, 
dagegen kommt Asparagin noch weiter unten gleichzeitig 
mit Glykose zum Vorschein. 



WACHSTÜMSGESCHICHTE D. BAMBüSGEWAECHSE. 479 

3. Aspa ragin verschwindet aus Noclien und Internodien 
sobald ihre Streckung aufhört, während Tyrosin noch 
lange Zeit in denselben erhalten bleibt. 

4. Tyrosin verschwindet zuletzt aus ganz erwachsenen und 
in Zell\Yandverdickung begriffenen Internodien und 
Nodieu. 

Nun müssen die einmal vorhandenen und allmählig ver- 
schwindenden Amidokörper, wie schon bekannt, zur Eiweissre- 
geueration, sei es direct oder indirect, verwendet werden, weil 
sonst weitere Zersetzungs- oder Oxydationsproducte, wie Am- 
moniak, iSî^itrate us.w. in den Schösslingsgeweben angehäuft werden 
müssen, was durchaus nicht der Fall ist. Zwar habe ich weder 
Ammoniak noch Nitrat mit Nessle r 's Reagens resp. Dipheny- 
lamin-Schwefelsäure in den betreffenden Geweben nachgewiesen. 

Bei der Betheiligung an diesem Eiweissregenerationsprocess 
scheint, wie aus den oben erwähnten Thatsachen ersichtlich ist, 
Asparagin viel leichter verwendbar zu sein und so kommt es 
nur an Stellen, wo regere Eiweissbildungsprocesse stattfinden, 
vor. Auch in diesem Sinne lassen sich die folgende Beobach- 
tungen erklären. In verkümmerten Schösslingen von Phyllostachys 
iiiitis, die täglich nur einige mm wachsen, während nebenbei 
stehende kräftige Exemplare täglichen Zuwachs von mehr als 
70 cm zeigten, konnte ich in verschiedenen Internodien Aspara- 
gin niemals auffinden, dagegen kam Tyrosin dort reichlich vor. 
Ganz ähnlich verhalten sich die Rhizomspitzen von Phyllostachys 
mitis und Phyllostachys bamhusoides, die im Spätherbst (October) 
untersucht wurden, wobei sie äusserst langsam wachsen. In 
verschiedenen Internodien derselben konnte ich trotz vielfacher 
Bemühungen kein Asparagin mit Sicherheit nachweisen, während 
Tyrosin dort reichlich vorkommt. Dass das Vorkommen von 



480 K. SHIBATA : 

Asparagin stets mit der lebhaften Stoffbildung bei schneller 
Streckung verbunden ist, ergiebt sich auch aus folgender Bemer- 
kung Pfeffers^): ,, War früher die Wurzel das lebhaftest wach- 
sende Organ des Keimpflänzchens, so ist dieses jetzt das Stämm- 
chen geworden und dementsprechend wendet sich jetzt der 
Hauptstrom von Glykose und Asparagin in dieses." 

Hingegen verhält sich das Ty rosin viel träger in dieser 
Beziehung, so dass es in schon erwachsenen Thoilen lange Zeit 
zurückbleibt. Beim Verschwinden des Tyrosins aus ganz er- 
wachsenem Internodium wird ein Theil in loco verbraucht, aber 
ein anderer Theil wird vielleicht den oberen wachsenden 
Internodien zugeführt und dabei müssen die jungen Bastelemente 
als Leitungsbalmen benutzt werden, wie besonders reichlicher 
Gehalt an Tyrosiu es vermuthen lässt. 

In jeder Hinsicht sind Asparagin und Tyrosin nicht von 
gleichem Werthe. Das erste ist ausgezeichneter Eiweissbaustoff. 
während das zweite es nur bis zu einem gewissen Grade ist, 
Soweit es leichte Verwendbarkeit des Asparagins anbetrifft, steht 
mein Ergebniss mit Hansteen') im Einklang. 

Wie entstehen das Asparagin und das Tyrosin ? 

Aus der Localisation ergiebt es sich schon, dass das Tyrosin 
nur bei der Zersetzung des schon vorhandenen Eiweisses ent- 
steht und nicht durch synthetischen Process. In den jungen 
Geweben unterhalb des Urmeristems, wo noch kein reducierender 
Zucker vorhanden ist, tritt es schon auf und vermehrt sich nach 
unten, in dem Masse wie das Eiweiss in den Zellen abnimmt. 
Andererseits konnte ich in fungierenden Wurzeln und Bhizomen, 



1) Pfeffer, Untersuchungen über die Proteinkörner und die Bedeutung des Asparagins. 
Jahrb. f. wiss. Bot. Bd. VIII, p. 548. 

2)Hansteen, Über Eiweisssynthese in grünen Phaneroganien. Jahrb. f. wiss. Bot. 
Bd. XXXIII, p. 449, p. 485. 



WACHSTUMSGESCHICHTE D. BAMBUSGEWAECHSE. 481 

die wohl als Bildungsstätte der organischen Stickstoffverbindungen 
betrachtet werden dürfen, in allen untersuchten Fällen niemals 
Tyrosin nachweisen. Ferner giebt der Blutungssaft, der beim 
Stoffabfuhr vom Rhizome eine wichtige Rolle spielt, keine Tyro- 
sinreaction. Dafür sprechen noch folgende Versuche, die ich 
wiederholt ausgeführt habe. Wenn man irgend eine abgeschnit- 
tene Rhizom spitze oder einen Schössling von Fhyllodachys mitis, 
FhyUoslachys hambusoides oder Phyllostachys puberula in destilliertes 
Wasser stellt und am Licht oder im Dunkeln verweilen lässt, so 
sieht man nach 2-o Tagen bedeutende Zunahme von Tyrosin in 
beliebigen Internodien. Da in diesem Falle vorher weder Nitrat 
noch Ammoniak in den Zellen nachweisbar war, so ist die 
nachträgliche synthetische Bildung von Tyrosin wohl ausgeschlos- 
sen, und die beobachtete Tyrosinzunahme muss allerdings auf 
die Ei Weisszersetzung zurückgeführt werden. 

Mit dem Asparagin ist die Sache schwieriger zu entscheiden. 
Obw^ohl ich in oben erwähnten Versuchen die gleichzeitige 
Asparaginbildnng gewöhnlich nicht beobachten konnte, ist es 
natürlich nicht ausgeschlossen, dass bei der Eiweisszersetzung 
hierbei entstandenes Asparagin schnell zur Eiweissregeneration 
verbraucht wurde und demgemäss nicht in gleichem Masse wie 
Tyrosin zur Anhäufung kam. Andererseits mag die oben erwähnte 
Localisation des Asparagins in kräftig wachsenden Theilen dadurch 
zu Stande gekommen sein, dass das im Rhizome fortwährend 
gebildete Asparagin mit dem Blutungssaft den wachsenden 
Schösslingen zugeführt und in den betreffenden Theilen ange- 
häuft wurde^). Gegenwärtig haben wir drei Möglichkeiten in 
Bezug auf Asparaginbildung : erstens durch directe Eiweiss- 



1) Natürlich muss das Material der in wachsenden Schösslingen auf eine so erliebliche 
Weise umgesetzten Eiweissstoffe von Ehizomen entstammen. 



482 K. SHIBATA : 

Zersetzung/) zweitens durch Synthese aus Ammoniak^) und 
drittens durch Umwandking von Amidosäuren etc.^) Ob die eine 
oder andere von diesen Möglich keiten in unserem Falle zutrifft 
muss vorläufig unentschieden bleiben. 

Nun gehe ich zur Besprechung der interessanten Löslichkeils- 
verhältnisse des Tyrosins über. Bei der Untersuchung der Schöss- 
linge von Phyllostachys mitis im; IV und V Stadium habe ich 
gefunden, dass alle jungen, noch mit plasmatischem Wandbeleg 
versehenen Bastelemente und oft auch parenchymatische Zellen 
mit schönen Tyrosin-Nadelbüscheln erfüllt sind (Fig. 59 u. 60). 
Dieser Umstand Hess mich zuerst vermuthen, dass das Tyrosin 
schon in den lebenden Zellen in Krystallform vorkommt. Aber 
nach genaueren Untersuchungen lässt es sich bald feststellen, dass 
das Tyrosin in den intacten lebenden Zellen ganz gelöst im Zellsaft 
vorkommt und nur erst in den beim Schneiden geöffneten Zellen 
zu Krystallen erstarit. Man kann diese Thatsache mit aller 
Bestimmtheit in folgender Weise beweisen : ein 3-4-zelllagendicker 
Längsschnitt des tyrosinhaltigen Internodialgewebes wird zuerst 
durch Heruraschwenken in Wasser von den an den Schnittflächen 
anhaftenden Tyrosinkrystallen befreit und dann unter dem 
Microskop mit feiner Nadel zerzupft, so sieht man bald, dass 
in vorher klarem Zellsaft der verletzten Bastelemente und 
Parenchymzellen eine Krystallbildung stattfindet, welche nach 
wenigen Secunden sich als Tyrosin deutlich erkennen lässt. So 
wird hier das äusserst schwerlösliche Tyrosin'') in so hohem 
Masse im Zellsaft in Lösung gehalten, dass es sich schon nach 



1) Pfeffer, Pflanzenpliysiologie. Ed. I, jd. 4G4. 

"2)0. Loew, Die chemische Energie der lebenden Zelle, p. 77; p. 78. 

3) E. Schulze, Üb. d. Umsatz d. Ei weissstoffe in der lebenden Pflanzen. Zeit. f. physiol. 
Chemie. Bd. XXIV, p. 63. 

4) 1 Thcil Tyrosin ist löslich in 1900 Theil Wasser bei IG^C. 



WACHSTÜMSGESCHICHTE D. BAMBUSGEWAECHSE. 483 

blosser mechanischer Verletzung der Protoplasten als Krystalle 
abscheiden lässt. Die Tödtung des Gewebes durch Chloroform- 
dampP, Osmiumsäuredampf sowie Erhitzung bewirkt ebenfalls die 
Abscheidung der Tyrosinkrystalle. Ferner kommt in den duich 
ofo KNOo-Lösung sehr stark plasmolysierten Bastzellen Ty rosin 
nach lauger Zeit nicht als Krystalle zum Vorschein^), aber nach 
Verletzung der plasmolysierten Zellen treten die Krystalle bald 
hervor. Wie solch eine hohe Löslichheit des Tyrosins zu Stande 
kommt ist schwer zu beantworten'). Allerdings muss hier der 
Einfluss von den Saftraum umgebenden, lebenden Protoplasten in 
erster Linie in Betracht kommen. Die Ausscheidung von Tyro- 
sinkrystallen beim Schneiden des Gewebes lässt sich nicht in 
Schösslingen von I, II und III Stadien zeigen. Ebenso verhalten 
sich die tyrosinarmen Schösslinge von Arundinaria- und Bamhusa- 
arten. Die Schösslinge von Phyllostachys 'puherula und Phyllo- 
stachys bambusoides zeigen ganz gleiches Verhalten wie im oben 
dargestellten Fall von Phyllostachys mitis. 

Ausserdem scheint sich ein Theil des Tyrosins auch in 
die Zellwände einzulagern, da die Zellwände der jugendlichen 
Basteleraente und später auch des Parenchyms immer stärker 
roth durch Mill on 's Reagens gefärbt werden, in dem Masse, 
dass Tyrosin in den Zellen selbst abnimmt. Bekanntlich haben 
früher Correns") und Fischer^) die Vermuthung ausgesprochen, 
dass die sogenannte Eiweissreaction der Zellw^ände vetschiedener 
Pflanzen von in dieselben eingelagertem Tyrosin herrühre. Neuer- 



1) Vergl. Pfeffer, Pflanzenpliysiologie Bd. I, p. 465. 

2) Jedenfalls ist die Ansicht Belzung's (Recherche chimique snr 1. Germination etc. 
p. 219), dass das im Zellsaft gelöste Eiweiss die Krystallisation von Asparagin, Leucin etc 
verhindern soll, ganz unzutreffend, denn in unserem Falle wird die Krystallisation schon 
durch blosse mechaniiche Yerlelzung des Protoplastes im Zellsaft eingeleitet . 

3) Correns, Über die vegetabilische Zellmembran. Jahrb. f. wiss. Bot. Bd. XXVI, p. 616. 
4j Fischer, Zur Eiweissreaction der Zellmembran. Ber. d. D. Bot. Gesells. Bd. V, p. 429. 



484 K. SHIBATA : 

dings vertritt aber Czapek') die Ansicht, dass es sich um die 

Reaction eines phenolartigen Körpers handelt, welchen er dem 

von ihm in Mooszellwänden aufgefundenen Sphagnol als nahe 
verwandt betrachtet. 

Gerbstoffe uxd Fette. 

Gerbstoffe und Fette spielen bei Entwicklung der Bambus- 
schösslinge nur eine untergeordnete Eolle. 

Bei den von mir untersuchten Phyllostachys- und Bamhusa- 
Arten sind Gerbstoffe in verschiedenen Theilen der Schösslinge 
überhaupt nicht in nachweisbarer Menge vorhanden.") In dieser 
Hinsicht bietet Arundinaria quadrangular is ein abweichendes 
Verhalten dar. Ein ca. 100 cm langer Schössling zeigte folgende 
Vertheilung der Gerbstoffe : Das Urmeristem des Vegetations- 
punktes ist frei davon, und dann treten sie plötzlich in reich- 
licher Menge im Niveau der 3ten Scheideblattanlage auf, zugleich 
mit der ersten nachweisbaren Stärke. Gerbstoffe kommen in 
einigen nachfolgenden Internodien in gleicli reichlicher Menge 
vor und dann nehmen sie nach unten ab. Dabei reagieren am 
stärksten die Bindenzellen und die peripherischen Centralcylinder- 
parenchymzellen. Selbst in erwachsenen Internodien und Nodien 
in der Nähe der Erdoberfläche ist eine schwache Beaction be- 
merkbar. Die Wurzelhaube enthält auch kleine Mengen der 
eisenbläuenden Gerbstoffe. In der dritten oder vierten Scheide- 
blattanlage am Vegetationspunkt treten die Gerbstoffe auf und 
nehmen nach unten zu. So sieht man hier eine analoge Verthei- 
lungsweise wie in den bisher bekannten Fällen, z. B. bei Vicia, 



1) Czapek, Zur Clieniie der Zellmembran bei den Laub- und Lebermoosen. Flora. 1899- 
Bd. 86, H. 4. 

2) Abgesehen von kleinen Mengen in Wurzelhaul)en, Sclieideblilttern etc. 



WACHSTUMSGESCHICHTE D. BAMBUSGEWAECHSE. 485 

Helianthuê) , Zuckerrohr-) u.s.w. Im vorliegendeu Falle scheinen 
die autochthonen^) Gerbstoffe sich aplastisch zu verhalten, weil 
hier Zucker, Stärke u.a. auf ganz gleiche Weise vorkommen wie 
bei anderen gerbstofffreie u Arten. 

Auch kommt den Fetten höchstens nur eine locale Bedeut- 
ung zu als Baumaterial der verkorkenden oder verholzenden Zell- 
wände, so zum Beispiel verschwinden die kleinen Fetttropfen in 
einer subepidermalen Zellschicht der Wurzel schon bei eintreten- 
der Verkorkung der Zellwände. In jeder älteren Halmparen- 
chyrazelle von Arundinaria japonica, A. Ilinchii, Bamhusa 
floribunda etc. befindet sich je ein ökirtiger gelber Tropfen, meist 
in Verbindung mit Calciumoxalatdrusen. Diese kugeligen 
Gebilde unterscheiden sich von echten Fetttropfen dadurch, dass 
sie niemals mit Alkannatinktur sich färben. In vielen Punkten 
stimmen sie mit dem zuerst von Monteverde^) in Gramineen 
aufgefundenen ,, Harzkörper" überein. 

Mineralstoffe. 

Ich habe in verschiedenen Jahreszeiten die Vertheilung der 
Mineralstoffe in den Beservestoffbehältern und den Scliösslingen 
verfolgt. Als Untersuchungsmaterial dienten mir hauptsächlich 
Phyllostachys mitis und Phyllostachys bambusoides. 

Ich konnte in den Rhizomen, die schon beträchtliche 
Quantität der Stärke aufgespeichert haben, die Mineralstoffe 
leicht auffinden. Sie zeigten folgende Vertheilung sowohl in 
Internodieu als in Nodien : 



1) Vei-gl. Kutscher, Über die Verwendnng der Gerbsäure im Stoifwechsel der Pflan- 
zen. Flora. 1883, p. 33. 

2) Went, Chemisch-physiologische Untersuchungen über das Zuckerrohr. Jahrb. f. wiss. 
Bot. Bd. XXXr, p. 297. 

3) Kraus, Grundlinien zu einer Physiologie des Gerbstoffes, p. öS. 

4) Monteverde, Über Ablagerung von Calcium- und ^Magnesiumosalat in der Pflanze. 
Bot. Centralb. 1890. Bd. XLIII, p. 327. 



486 Iv. SHIBATA : 

Pliosplior reiclilich im Parenchym des Centralcylinders ; 
nur wenig in Siebröliren. 

Magnesium reiclilich vorzugsweise in Siebröliren. 

Kalium reiclilich im Parenchym. 

Calcium nicht nachweisbar in Schnitten ; nur kleine 
Mengen in Aschen. 

Schwefel nur spurweise in Schnitten.^) 

Chlor ziemlich reichlich im Parenchym ; aber fast keins in 
Siebröhren und anderen Elementen der Gefässbündel. 

In den AVurzeln können die obengenannten Stoffe auf über- 
einstimmende Weise aufgefunden werden. Die Nitrate^) sind 
in Phizomen nur sehr wenig vorhanden, aber zeitweilig etwas mehr 
in der Wurzelrinde. Sie kommen hier also überhaupt nicht zur 
nennenswerthen Aufspeicherung. Leitgeb^) hat früher gezeigt, 
dass der Phosphor in DahlictkiioWen hauptsächlich als Calcium- 
phosphat vorkommt. Hier ist aber dies nicht der Fall, da ich 
keine lösliche Ca- Verbindung in den Zellen auffinden konnte. 
Die Siebröliren der Rhizome und Wurzeln enthalten, wie schon 
erwähnt, nur geringe Mengen des Phosphors, dagegen reichlich 
Magnesia. Daher scheint Magnesium theils als Phosphat, Iheils 
als lockere oro-anische Verbinduno; in Siebröhren vorzukom- 
inen. Schimper^) und Zacharias') haben auch im Siebröhr- 
ensafte von Cucurbita, Wista7'ia, ÄristolocMa und Menispermum 
reichliche Mengen des Magnesiums aufgefunden. 



1) Schi m per (Zur Frage der Assimilation der Mineralsalze. Flora, 1890. p. 223j hat 
auch in den meisten Illiizomen keine Sulfatreaction erhalten und es dem Vorhandensein von 
Krystallisation verhindernden Substanzen in Zellen zugeschrieben. 

2) Vergl. Mol isch, Über microchem. Nachweis von Nitraten. Ber. d. D. Bot. G. Ed. I, 
p. 154. 

3) Leitgeb, Über die durch Alcoliol in -DaW/a-Knollen hervorgerufenen Krystalle. Bot. 
Zeit. 1887. p. 29. 

4) S eil imp er, Zur Fragi- der Assimilation der Mineralsalze- Flora, 1890, p. 228. 

5) Zac liar ias, Über d. Inhalt d. Siebröhren von Cucurbita. Bot. Zeit. 1884. p. 71. 



WACHSÏUMGSESOHICHTE D. BAMBUSGEWAECHSE. 487 

Schritt für Schritt mit der Entleerung der Kohlehydrate 
verschwinden auch die Mineralstofife aus dem Rhizomparenchym. 

Die Nitrate sind schon im IV Stadium in Rhizomen und 
Wurzeln nicht mehr nachweisbar. Ebensowenig konnte ich im 
Blutungssaft die Nitratreaction erhalten. Wahrscheinlich werden 
die Nitrate hierbei fortwährend zu organischen Verbindungen 
verarbeitet.^) 

An den unteren, mit zahlreichen jungen Wurzeln besetzten 
Theilen der Schösslinge besitzen Phosphor und Kalium eine 
andere Vertheiluugsweise als im Ehizome ; sie kommen nun 
reichlicher in den Bündeln vor. Das Magnesium befindet 
sich hier auch in Siebröhren bevorzugt. Ich konnte niemals 
die Nitrate im Schösslingskörper auffinden.-) Da sie auch im 
Bhizome fehlen, so ist es höchst wahrscheinlich, dass überhaupt 
nur w^enig Nitrat als solches in den Schössling eingeführt wird. 

Von dieser Region nimmt die direct nachweisbare Menge 
von Phosphor, Magnesium, Kalium und Schwefel oben nach dem 
Vegetationspunkte hin immer mehr zu, wie man sich durch die 
Musterung succ.essiver Querschnitte überzeugen kann. Phosphor 
kommt anscheinend am reichlichsten in 2-3 cm Entfernung vom 
Vegetationspunkt vor, und von hier nach oben nimmt die direct 
nachweisbare Menge desselben wieder ab. Merkwürdigerweise 
kommt Phosphor fast ausschliesslich in den eiweissreichen Pro- 
cambialsträngen vor.^) Im Urmeristem lassen sich die anorgani- 



1) Die feinen Xebenwurzeln geben stets mehr oder minder starke Nitratreaction. Ob sich 
in irgend einer "Weise der hier befindliche Pilzsymbiont an der Stickstoöassimilatiou 
betheiligt, muss zur Zeit dahingestellt bleiben. 

2) Übrigens ist es klar, dass die Nitrate sich nicht Lei so lebhafter Eiweisszersetzung 
bilden. (Vergl. Schulze, Über d. Vorkommen von Nitraten in Keimpflanzen. Zeit. f. 
physiol. Chemie. Bd. XXIf, p. 83). 

3) Ich habe auch die Lil ienfeld'sche Methode für Erkennung der Localisation des 
Phosphors mit Erfolg benutzt, (vergl. Strasburger, Das botanische Practicum. III. Aufl. 
p. 144). 



488 K. SHIBATA : 

sehen Phosphorverbindungen nicht mehr nachweisen, sondern 
grosse Mengen organischer Verbindungen^). Magnesium tritt 
ebenfalls in der Nähe der Spitze fast ausschliesslich in den 
Bündelanlagen auf, aber kleine Mengen sind auch in den 
eiweisshaltigeu Internodialzonen vorhanden. Beachtet man die 
bevorzugten Vorkommnisse des Magnesiums in Siebröhren, 
Bündelanlagen, Internodialzonen und auch im Urmeristem, so 
darf man wohl annehmen, dass es irgend eine wichtige Rolle 
bei Eiweissumsatz oder Eiweisswanderung spielt.") Kalium lässt 
sich in der Asche des Vegetationspunktes reichlich nachweisen. 
Hingegen ist die Schwefelsäure nicht direct im Vegetations- 
punkt nachweisbar, obgleich sich schon ca. 2 cm weiter unten 
eine ziemlich starke Reaction zeigt. Das in minimaler Menge 
zugeführte Calcium wird in einiger Entfernung vom Vegetations- 
punkte als Kalkoxalat niedergeschlagen. Chlor lässt sich ziemlich 
viel in eiweissreichen Internodialzonen nachweisen. Im Urme- 
ristem scheint es jedoch gänzlich zu fehlen. 

Wenn die Schösslinge über die Erde emporwachsen und die 
Internodien nach einander ihre definitive Länge erreichen, so 
nimmt die nachweisbare Menge der Mineralstoffe in denselben 
stetig ab. In unterirdischen Internodien tritt nun mehr oder 
minder starke Nitratreaction ein, da die erwachsenen Wurzeln an 
dieser Region schon ihre Thätigkeit entfalteten und begannen 
die Bodensalze aufzunehmen. 

In der ersten Entwicklungsphase der Wurzel ist eine starke 
Ansammluns; der Miueralstoffe im meristematischen Gewebe leicht 
constatierbar. In etwas länger erstreckten Wurzeln findet eine 
ähnliche Ansammlung in der Spitze statt. Phosphor und Mag- 

l)Vergl. Schimper, I.e. p. 224. 

2) Vergl. Hornberger, Cheiuiiclie Untersuchungen über das Wachstum der Maisjiuanze. 
Landw. Jahrb. 1882. p. 278. 



WACHSTÜMSGESCHICHTE D. BAMBUSGEW AECHSE. 489 

nesium kommen, wie in Scbösslingen, liauptsächlicli in jungen 
procambialen Elementen vor, die vielleicht ihre Wanderbahn 
herstellen. Im Urmeristem fehlen nachweisbare Mengen von 
Schwefel und Chlor. Die mehr als 40 cm lang gewachsenen 
AVurzeln zeigen ziemlich starke Nitratreaction an der Rinde, 
welche von von aussen aufgenommenen Salzen herrührt. 

In den jüngsten Scheideblattanlagen in der Umgebung des 
Vegetationspunktes lässt sich sehr früh eine Ansammlung von 
Phosphor und Magnesium beobachten. Uebrigens bedarf dies 
keiner Besprechung mehr. 

VII. lieber die Entleerung der Reservestoffe. 

Die Versuche von Hansteen^) und Purie witsch") haben 
die Thatsache festgestellt, dass bei Endospermen, Samenlappen, 
Knollen und Khizonien die Entleerung der deponierten Reserve- 
stoffe mehr oder minder selbstthätig stattfinden kann. Nun schien 
es mir geboten zu bestimmen, erstens inwieweit die Entleerung des 
Rhizoras unabhängig von wachsenden Scbösslingen vor sich gehen 
kann, und zweitens in welchem Grade die Entwickluno- der 
Schösslinge durch die totale oder partielle Separierung vom 
Rhizomsystem beeinflusst wird. Zu diesem Zweck habe ich 
Mitte April eine Anzahl kräftig wachsender, unterirdischer 
Schösslinge aufgesucht und verschieden tiefe Einschnitte in ihre 
Stieltheile und benachbarte Rhizominternodien gemacht. Die 
betreffenden Rhizomtheile waren in diesem Stadium mit Stärke 
strotzend erfüllt, wie ich durch die Musterung zahlreicher 
Exemplare überzeugt war ; es zeigte sich folgendes : 

l)Hansteen, tJber die Ursache der Entleerung der Eeservestofïë aus Samen. Flora. 
1894 Bd. 79, p. 419. 

2) Pur ie witsch, Physiologische Untersuchungen über die Entleerung der Reserve- 
stoffbehälter. Jahrb. f. wiss. Bot. Bd. XXXI, p. 1. 



490 



K. SHIBATA : 





Starke Gllykose 


Eohrzncker 


Rindenparenchym 


strotzend erfüllt (5) 


1 
sehr wenig ziemlich viel 


Centralcylinderparen- 
c ]i y ni 


strotzend erfüll t- 
reclit viel (5-4) 


desgl. 


desgl. 


Markparenchym 


strotzend erfüllt (5) 


desgl. desgl. 



In der folgenden Tabelle stelle ich die Ergebnisse einiger 
Versuche zusammen : 




WACHSTÜMSGESCHICBTE D. BAMBUSGEWAECHSE. 491 

Alle oben angeführten Versuche ergaben übereinstimmend, 
dass die Entleerung, zumal die Stärkeauflösung in bestimmten 
E.hizompartien unabhängig von Schösslingen vor sich gehen kann. 
Puriewitsch hat bei den Versuchen mit den Rhizomen von 
Curcuma und Rudhechia gezeigt, dass in solchen Rhizomen eine 
partielle Entleerung selbstthätig stattfand, wenn die continuierliche 
Ableitung der Lösungsproducte mittelst Gyps besorgt wurde^). 
Nun in meinen Versuchen war die Bedingung für derartige Stärke- 
entleerung besonders günstig. Die zahlreichen kräftigen AVurzeln 
an Khizomknoten erzeugten zur Zeit einen ansehnlichen Blutungs- 
druck und der zuckerhaltige Blutungssaft wurde immer fort von 
den Schnittflächen der Rhizome ausgeschieden. Damit wurde 
eine fortwährende Wegführnng der Lösungsproducte erzielt, 
welche eine so volkommene Entleerung herbeiführte. 

Ferner ist es aus obigen Versuchen ersichtlich, dass die 
Entwickelung der Schösslinge durch jeden operativen Eingriff in 
benachbarte Rhizominternodien — d.h. durch jede Herabsetzung 
des Blutungsdrucks — bald sistiert wird. 

Macht man in der Nacht oder frühmorgens ein Bohrloch in 
ein beliebiges Internodium des kräftig wachsenden Schösslings, so 
quillt bald ein zuckerhaltiger klarer Saft hervor. Am 19. Mai 
wurde der Blutungssaft von einem mittleren Internodium eines 
ca. 1.5 Meter hohen Schösslings von Phyllostachys imherula 
gesammelt. In 100 ccm dieser Flüssigkeit fand ich 0.289 gr 
Glykose. Ausserdem enthält der Blutungssaft eine kleine Menge 
der Amide, da er nach Beseitigung des Eiweisses") eine starke 
Trübung beim Zusatz von Quecksilberoxyd nitrat giebt. 



1) Puriewitsch, I.e. p. 28. 

2) Nach der S t ü t z e r 'sclien Methode. 



492 K. SHIBATA : 

Schröter^) schrieb: ,, Später, in den ausgewachsenen Glie- 
dern, findet sich oft ein khires Wasser, das in manchen trockenen 
Gegenden den Reisenden ein höchst willkommener Fund ist." 
Derartige mit Wasser erfüllte Markhöhlen habe ich manchmal bei 
jungen Halmen von Phyllostachys mitis gefunden. Das Wasser 
ist nichts Anderes als der Blutungssaft, der sich von radialen 
Rissen an der Peripherie des Diaphragms ausgeschieden hat. In 
einem Falle betrug der Glykosegehalt der Flüssigkeit, die sich in der 
unteren Internodialhöhle von Phyllostachys mitis befand, 0.2G9?ö. 
Wenn man frühmorgens einen Bambusbusch besucht, so wird 
man einen förmlichen Regen von Wassertropfen aus den Scheide- 
blattspitzen der wachsenden Schösslinge bekommen.'") Dies in 
bekannter Weise von Blattspitzen ausgeschiedene Wasser enthält 
auch Glykose neben einer Spur von Amiden. Eine Zuckerbes- 
timmung der am 28. April gesammelten Flüssigkeit ergab 
0.09589^ Glykose. 

Alle diese Thatsachen weisen darauf hin, dass eine erheb- 
liche Menge der Kohlehydrate und vielleicht auch Amide mit 
dem Blutungssaft den Schösslingen zugeführt werden. Dadurch 
werden die Schösslinge mit den Baustoffen genügend rasch ver- 
sorgt.^) Die oben erörterten Bauverhältnisse der Stieltheile lassen 
sich auch nicht anders denken, als dass hier der ausgiebige 
Stofftransport nur durch die Bündel und zwar durch die wohl 
ausgebildeten Gefässe geschehen kann. 



1) Schröter, Der Bambus und seine Bedeutung als Nutzpflanze, p. 14; Colin, tJber 
Tabaschir. p. 375. 

2) Molisch, Über das Bluten tropischer HulzgewJiclise. Ann. d. Jard. Bot. Buit. 
1898. Suppl. II, p. 23. 

3) Vergleiche hierzu: Strasburger, lîau und Verrichtungen der Leitungsbahnen. 
p. 877; Fischer, Beiträge zur Physiologie der Holzgewilchse. Jahrb. f. wiss Bot. Bd. XXII, 
p. 75 ; p. 150. 



WACHSTUMSGESCHICHTE D. BAMBUSGEWAECHSE. 493 

VII. Zusammenfassung. 

In Obigem habe ich die wesentlichen Züge der Wachs- 
tumsgeschichte der Bambusgewächse darzustellen versucht. Die 
Hauptresultate werden hier kurz in folgenden Worten zusam- 
mengefasst : 

1. Die Stärke wird in parenchymatischen Zellen der 
Khi z cm e, Halme und Wurzeln als Haupt reservestoff abgelagert. 
Die Verminderung derselben im Winter wurde nicht beobachtet, 
während zur Zeit des raschen Austreibens von Schösslingen eine 
unverkennbare Stärkezunahme (transitorisch) in benachbarten 
Rhizomtheilen constatiert wurde. 

2. Die Gl y kose dient als Baumaterial in wachsenden Theilen 
der Schössliuge und ist in schon fertig gestreckten Internodien 
derselben transitorisch reicblich aufgespeichert. 

3. Der Rohrzucker tritt als das Lösungsproduct der 
Stärke im Parenchym der Rhizome und Halme auf. 

4. In schnell wachsenden Schösslingen fand eine ausgiebige 
Eiweisszersetzung statt, dabei trat Tyrosin in bedeutender Menge 
auf. 

Tyrosin und Aspa ragin zeigen einen weitgehenden 
Unterschied in ihrem Verhalten. Tyrosin wird schwerer und 
langsamer für Eiweissregeneration verbraucht, so dass es in schon 
erwachsenen Theilen eine Zeit lang zurückbleibt. Hingegen ist 
Asparagin leicht und rasch dazu verwendet und kommt nur 
an Stellen vor, wo eine lebhafte Stoffbildung stattfindet. 

5. Gerbstoffe kommen nur in Schösslingen einzelner Arten 
vor, und Fette spielen hierbei keine wichtige Rolle sowohl als 
Wanderstoffe w^ie als Reservestoffe. 

6. Phosphor, Kalium, Magnesium und Chlor werden 



494 K. SHIBATA : 

in cleii Reservestoffbehiiltern aufgespeichert, dabei kommt Mag- 
nesium vorwiegend in Sieb röhren vor. Calcium und Schwefel 
sind gewöhnlich nicht direct nachweisbar. 

7. Die Mineralstoffe wandern bei rascher Entwicklung der 
Schösslinge schnell von den Ehizomen aus und werden in den 
wachsenden Theilen angesammelt. In der Spitze der Halme, 
Rhizome und Wurzeln befinden sich Phosphor und Magnesium 
in direct nachweisbarer Form fast ausschliesslich in Procambial- 
stränsen. Schwefel wird erst im wachsenden Theile der 
Schösslinge deutlich nachweisbar. 

8. Die vom Boden aufgenommenen Nitrate werden wahr- 
scheinlich schon in den Wurzeln und Rhizomen zu organischen 
Verbindungen verarbeitet. 

9. Die Auflösung der Stärke und die Entleerung der Lö- 
sungsproducte aus den Rhizomen können unabhängig von der 
Entwickelung der Schösslinge fortgehen. 

10. Der ausgiebige und schnelle Stofftransport nach wachsen- 
den Schösslingen von den Rhizomen kann in Wasserbahnen 
geschehen. Dafür sprechen vor allem die Blutungserscheinungen 
der Rhizome und Schösslinge und die Bau Verhältnisse der Schöss- 
lingsstiele. 

Botanisches Institut 
Juni 1890. Kaiserl. Universität 

zu Tokio. 



WACHSTUMSGESCHICHTE D. BAMBUSGEWAECHSE. 495 



VeRZEICHNISS der U]SrTERSüCHTEN ArTEN/) 

Phijllostacltiis mltis llivàère. (Nom. Jap, Bloso-cldhu.) 
Fhyllostacliys hamhusoides Sieb, et Zucc. (Nom. Jap. Ma-dahe.) 
Fhyllosiachys hamhusoides Sieb, et Zucc. var. aurea Makino. (Nom. Jap. 

Hotei-chihu.) 
Plryllostachys jntberula Munro. (Nom. Jap, Ha-chiku.) 
Phylloslacliys puherula Munro. var. nigra. (Nom. Jap. Kuro-chikic.) 
Phyllostacliys Kumasasa Munro. (Nom. Jap. Ohame-sasa.) 
Arundinaria japonica Sieb, et Zucc. (Nom. Jap. Ya-dake.) 
Arundinaria Simoni Rivière. (Nom. Jap. Me-dahe.) 
Arundinaria Blatsumurœ Hackel. (Nom. Jap. Kan-chiku.) 
Arundinaria quadrangidaris Makino. (Nom. Jap. Shikaku-dake.) 
Arundinaria Hindsii Munro. (Nom. Jap. Kansan-chiku.) 
Arundinaria Hindsii Munro. var. graminea Bean. (Nom. Jap. Taimin- 

cliiku.) 
Arundinaria Fortunei Eivière. (Nom. Jap. Chigo-sasa.) 
Arundinaria variabilis Makino. (Nom. Jap. Ne-sasa.) 
Arundinaria pygmcea Mitf. (Nom. Jap. Oroshima-cJäku .) 
Arundinaria. Narihira Makino. (Nom. Jap. Narihira-dake.) 
Arundinaria Tootsik Makino. (Nom. Jap. To-chiku.) 
Bamhusa horealis* Hackel. (Nom. Jap. Suzu-dake.) 
Bamhusa palmatd* Marliac. (Nom. Jap. Chimaki-sasa.) 
Bamhusa Veitchii* Carrière. (Nom. Jap. Kuma-sasa.) 
Bamhusa panicidata'^' Makino. (Nom. Jap. Nemagari-dake.) 
Bamhusa nipponicc(/^ Makino. (Nom. Jap. Miyako-sasa.) 
Bamhusa ramosa* Makino. (Nom. Jap. Azuma-sasa.) 



1) Die ausführliche Beschreibung der hier angeführten Arten findet man bei Makino, 
Bambusaceœ Japonicre (The Botanical Magazine, Vol. XIV, Nr. 156, p. 20 fF.), die beige- 
fügten japanischen Namen sollen zum Herausfinden der betreffenden Arten in der genannten 
Schrift dienen. 

Die mit * bezeichneten Arten gehören meiner Ansicht nach nicht eigentlich zu Bambusa, 
sondern sie würden vielleicht eine selbständige Gattung bilden. (Vergl. oben p. 446 und auch 
Makino, Ic. p. 20). 

An dieser Stelle spreche ich Herrn Makino für die von ihm gütigst vorgenommene Bestim- 
mung einiger Arten und auch Herren Asö und In ami für ihre freundliche Unterstützung 
bei einigen analytischen Arbeiten meinen besten Dank aus. 



496 K. SHIKATA : 

Bumhusa vulgaris Wendl. (Nom. Jap. Daîsan-chîhu.) 
Bambvsa nana Eoxb. (Nom. Jap. Howo-chihu.) 
Bamhusa nana var. normalis Makino. (Nom. Jap. Taiho-cJiikii.) 
Bamhnsa stenostachya Hackel. (Nom. Jap. Shi-chihu.) 
Dendrocalarnvs laiißorus, Mnnro. (Nom. Jap. Bla-chiku.) 



Inhalt. 



I. Einleitung 427 

II. Uutersuchungsmaterial und Methudisches 429 

III. Die Bauverhältnisse 433 

IV. Der Eotwicklungsvorgang der Schösslinge 453 

V. Verhalten der Baustoffe während der Entwicklung 

der Schösslinge 458 

VI. Ueber Entleerung der Eeservestoffe 489 

VII. Zusammenfassung 493 



Erklärung der Tafeln. 

Tafel XXII. 

Fig. 1. Zwei neben einander stehende Siebrörenglieder mit zahlreichen 
Siebtüpfeln {sJp) an den Seitenwänden, aus Rhizomknoten von 
PhyUostachijs mitis. spl Siebplatte, sip Siebtüpfel. Vergr. 360. 

Fig. 2. Querschnitt durch das Khizom von Bamhusa nipponica. R Rinde, 
Brg subcorticaler Bastring, cent Centralcylinderparenchym. Vergr. 30. 

Fig. 3. Querschnitt durch das Rhizom von Arnndina.ria japomca. B 
Bastbänder. Vergr. 30. 

Fig. 4. Die spindelförmige Anschwellung des Leptoms eines Knospenbündels 
bei der Ansatzstelle an der Rhizombündel. S Siebröhren, gl Geleit- 
zellen, G Gefâsse, P Parenchymzellen, c^j/carabiformartige Elemente 
Vergr. 70. 

Fig. .5. Querschnitt durch die Anschwellung. B Bastzellen, Pu. chf wie 
in Fig. 4. Vergr. 125. 

Fig. G. Theil der langgestreckten cambiformartigen Elemente aus dem 
mittleren Theile der Anschwellung, mit Querstreifen auf den Seiten- 
wänden. Vergr. 450. 

Fig. 7. Derselbe im Querschnitt. Vergr. 450. 

Fig. 8. Die Leptomanschvvellung in einem früheren Entwicklungsstadium. 
Längsschnitt durch den Knoten. S Siebröhren, B Bastzellen, chf 
cambiformartige Elemente. Vergr. 83. 

Fig. 9. Übergangsstelle der cambiformartigen Elemente zum normal gebauten 
Leptom, in einem jugendlichen Zustand. Sämmtliche Elemente mit 
auffallend grossen Zellkernen und reichlichem Plasmagehalt. S u. chf 
wie in Fig. 8. Vergi-. 450. 

Fig. 10. Dergleichen im fertigen Zustand, /^j Tüpfel, c6/ wie oben. Vergr. 
450. Figuren 4-10 beziehen sich auf Phyllostachys mitis. 

Fig. 11. Ein Gefässbündel aus einem inneren Teil des Rhizoms von Aruncli- 
luiria Hindsii. S Siebröhren, gl Geleitzellen, G Gefässe, d Durch- 
lassstelle. Vergr. 125. 

Fig. 12. Eine subepidermale sclerotische Parenchymschicht des Rhizoms 
von Bamhusa pcdmcda. Längsschnitt, ep E|)idermis, sei sclerotische 
Parenchymzellen, R Rindenzellen. Vergr. 360. 



Fig. 13. Querschnitt durch den Stieltheil. R Rinde, B BastLänder, hs 

Bastscheide des Mestombündels, mes Mestora. Vergr. 17. 
Fig. 14. Ein Mestorabündel im Stieltheile, mit volliîommen nmschhessender 

Bastscheide, t Tracheiden, P, hs, S, gl, u. G wie oben. Vergr. 125. 

Fig. 13-14. Phyllostachys mitis. 
Fig. 15. Querschnitt durch den dünnen Halmzweig von Arundinaria 

2Vjgmcea. ep u. Pt, wie oben. Vergr. 83. 
Fig. 16. Theil desgleichen von Arundinaria japonica. ep, B u. R wie 

oben. Vergr. 360. 
Fi"-. 17. Ein Halm-Bündel von Bambusa nana var. normalis, mit Paren- 

chymlamelle im innenseitigen Bastbeleg, ^jar. l Parenchymlamelle, 

S, G u. B wie oben. Vergr. 200. 
Fig. 18. Einige verschiedenartige Vorkommnisse des Parenchymgewebes 

im Bastbelege. Arundinaria Hindsii. B Bastbelege, par paren- 

chymatisches Gewebe. Vergr. 70. 
Fio-. 19. Ein Halmbündel von Bamhiisa nana. Die durch parenchymatische 

Zellen vom Mestom abgetrennte Masse des Bastbelegs bleibt nnverdickt. 

par. l, B wie oben. Vergr. 200. 
Fic. 20. Auftreten der Stärkekörner in neu differenzierter Parenchymlamelle. 

Vergr. 70. 
Fig. 21. Obiges im Längsschnitt. Vergr. 125. 

Fig. 22. Die durch successive Qucrtheilungen von Procambialzellen ent- 
standenen Parenchymzellen. pre Procambialzellen, /.; Kern, st Stärke- 

körner. Vergr. 360. Fis-uren 20-22. Arundinaria Hindsii. 



Jour. Sc. Coll. Vol. M PrWT 




Tafel XXIII. 

Fig. 23. Querschnitt durch die junge Wurzel von Eamhusa palmcda. i.U 
innere Rindenzellen, end Endodermis mit C asp a ry' sehen Streifen, 
per PericamLiuui, p.had peripherische Hadromstränge (primordiale 
Netztracheiden), ]).lep peripherische Leptomstränge. Vergr. 360. 

Fig. 24a. Peripherischer Theil der Wurzelrinde von Fhyllosiachys mais, 
ep Ptest der Epidermis, Jiypj stark verdickte (au den Aussenwänden) 
subepidermale Zellen, sei peripherische sclerotische Elemente, a.R 
äussere Rindenzellen. Vergr. 200. 

Fig. 246. Desgleichen im jugendliclien Zustand, ejj, liyp n. sei wie oben. 

Fig. 25. Starkverdickte Subepidermalzellen (Aussenscheide) von Fhyllo- 
staehys hamhusoides var. aurea. Vergr. 360. 

Fig. 26. Peripherischer Theil der Wurzelrinde von Bambusa vulgaris, ep 
Epidermis, Ityp unverdickt gebhebene Subepidermalzellen, sei peripheris- 
che Sclerenchymzellen, a.R äussere Rindenzellen. Vergr. 360. 

Fig. 27. Wurzelrinde von Bambitsa nana, ep, hyp, sel, a.Ru. LR wie oben. 
Vergr. 125. 

Fig. 28. Endodermis (end) und starkverdickte Pericambiumzellen (per) 
von Phyllostachys mit is. Verg. 360. 

Fig. 29. Längsschnitt durch die Endodermis von Phyllostachys mitis. Vergr. 
360. 

Fig. 30. Längsschnitt durch den perii)herischen Theil der Wurzelrinde von 
Arundinaria Maisuniurœ. Jiyp u. sei wie in Fig. 24. Vergr. 360. 

Fig. 31. Endodermis und angrenzende Rindenzellen von Bambusa steno- 
staehya. Längsschnitt, cel Zellstofiausvvüchse, end u. ^jer wie oben. 
Vergr. 360. 

Fig. 32. Dieselbon im Querschnitt, end, per u. cel wie oben. Vergr. 360. 

Fig. 33. C-förmig verdickte Eododermiszellen von Bambusa palmata. 
Vergr. 360. 

Fig. 34. Theil des Wurzelquerschnittes von Phyllostachys Kumasasa. ver 
Verstärkungsring, If Lufträume, end, i.R, per u. p.lep wie oben. 
Vergr. 360. 

Fig. 35. Verknüpfung der Le[)tomstränge durch dünnwandiges Verbindungs- 
gewebe. Phyllostachys bambusoides. ver.p. Verbindungsgewebe, 
nb.io Neben Wurzel, p.lepj, i.lep, G, end wie oben, (schematisiert) 
Vergr. 70. 



Fig. 36. Dergleichen bei Bambusa vulgaris. Vergr. 45. 

Fig. 37. Querschnitt durch die Hauptwurzel an der Ansatzstelle der Neben- 
wurzel. Arundinaria Matsumurœ. n.lep Leptomstrang der Neben- 
wurzel, leji Leptomstränge der Hauptwurzel, end, i.R, per wie oben. 
Vergr. 360. 

Fig. 38. Innerer Leptomstrang von Bamhusa vulgaris. S Siebröhre, ch 
Cambiformzellen, mz mechanische Zellen. Vergr. 360. 

Fig. 39. Verschmelzung des inneren Leptomstrangs mit dem peripherischen. 
i.lep innerer Leptomstrang, p.lep peripherischer Leptomstrang, S, 
ch wie oben. Vergr. 360. 

Fig. 40, Verschmelziüig zweier peripherischen Le[)tomstrange. Vergr. 360. 

Fig. 41. Directer Anschluss des Leptomstrangs an Hadromparenchym. G 
Gefäss, hj) Hadromparenchym (Gefässbelegzellen), lep u. mz wie oben. 
Vergr. 360. 

Fig. 42. Zusammentreffen zweier Hadromstränge. G, lip u. mz wie oben. 
Vergr. 360. 

Figuren 39-42. FhyUostaclujs hambusoides. 

Fig. 43. Verbiudungsgewebe zwischen inneren und peripherischen Leptom- 
strängen. i.lej), p.lep, mz u S wie oben. Vergr. 360. 

Fig. 44. Dasselbe von Baonhnsa vulgaris im Längsschnitt. S, mz, ver.p 
wie oben. Vergr. 360. 

Fig. 45. Querschnitt durch den Basaltheil der Nebenwurzel von Fhyllostachys 
mitis. lep Leptomstränge, mz mechanische Zellen. Vergr. 360. 

Fig. 46. 'J^heil der Einde der Nebenwurzel von Fhyllostachys puberula, mit 
endophy tischen M\"celläden. M Ptindenzellen, rnyc Pilzfäden, ves 
Vesiculen, kor gelbe körnige Substanz. Vergr. 360. 

Fig. 47. Querschnitt durch die Nebenwurzel von Fhyllostachys puberula. 
hyp Sube[ndermalzellen, sei peripherisclie sclerotische Zellen, B, myc, 
end, lep wie oben. Vergr. 200. 



Jnur. So. Coll. Vol.JfïTJïinm. 




/./* *■ »V..A.»... r«i 



Tafel XXIV. 

Fig. 48. Ein Scheideblattbündel mit starkem Bastbeleg (B) von Phyllo- 

stachys mitis. Vergr, S3. 
Fig. 49. Querschnitt durch das Scheideblatt von Arundinaria Matsumurœ 

VergT. 360. 
Fig. 50. Kleinere Scheide blattbündel von Bamhusa stenostachya ; Leptom 

(lep) ist stets von einschichtigen verholzten Elementen (h) umgeben. 

Vergr. 360. 
Fig. 51. Querschnitt durch das Scheideblatt von Bamhusa stenostachya. B 

Bastbelege, suhep.h subepidermal Bastplatte, P Parenchym, laes 

Mestom. Vergr. 83. 
Fig. 52. Desgleichen von Phyllosiachys mitis. If Lufträume, B, suhep. b, 

P, mes wie oben. Vergr. ca. 10. 
Fig. 53. Queranastomose des Scheideblattbündels von Arundinaria Eindsii. 

Vergr. 360. 
Fig. 54. Ligninauswüchse an Zellwänden. („Zwickel''). Vergr. 360. 
Fig. bD. Ein kleines Bündel im Laubblatt von Arundinaria Eindsii. ps 

Pareuchymscheide, hs Bastscheide, subep.h, Ead, Lep wie oben. 

Vergr. 360. 
Fi«'. 56. LäDcrsschnitt durch einen kleinen Nerv des Laubblattes von Bam- 

husa pahnata. ep, suhep. h, ps, bs wie oben. Bastscheideelemente 

sind reichlicher betüpfelt als subepidermale Bastelemente. Vergr. 360. 
Fig. 57a. Stärkekörner aus Fihizom von Phyllosiachys Kumasasa. Vergr 

830. 
Fig. blh. Polyadelphische Stärkekörner aus dem Halm von Bamhusa 

fjalmata. Vergr. 830. 
Fig. 58. Einige Tyrosinkrystalle, die ausserhalb des nach Boro din 'scher 

Methode behandelten Schnittes entstanden sind. Vergr. 360. 
Fi"-. 59. Beim Schneiden sofort in jungen Bastzellen auskiystallisierende 

Tyrosinkrystalle. B junge Bastzellen (mit plasmatischem Wand- 
belege), P Parenchymzellen, tyr Tyrosinkrystalle. Vergi*. 360. 
Fig. 60. Desgleichen in Längsschnittansicht. Vergr. 200. 
Fig. 61. Beim Einlegen vom Schnitt in Glycerin in das Zelllumen ausges- 
chiedene Tyrosinkrystalle. tyr Tyrosinkrystalle, si Stärkekörner, P. 

Parenchymzellen. Vergr. 360. 



Jour. Sc. Coll. Vol. XIII. PI. XXIV. 




K,Shibata del. 



PrinJ^-d ày Aoskièa hanAa Tùk^o Ja^-- 



Decomposition of Hydroxyamidosulphates by 
Copper Sulphate. 



By 



Edward Divers, M. D., D. Sc, F. R. S., Emeritus Prof., 

and 

Tamemasa Haga, D. Sc, F. C. S., 

Professor, Tokyo Imperial University. 



When copper sulphate is added to a sokitioii of a hydroxy- 
amidosulphate and tlie mixture heated, the acid of the salt is 
quickly decomposed into water, sulphur dioxide, sulphuric acid, 
amidosulphuric acid and nitrous oxide, with, possibly a little 
nitrogen. By itself, a heated solution of an alkali hydroxy- 
amidosulphate is in a state of very unstable equilibrium, generally 
hydrolysing into a solution of hydroxylamine acid sulphate, 
and always doing so in presence of a trace of acid, whilst in 
presence of even a trace of alkali it slowly passes into sul- 
phite and hyponitrite (this Journ, 3, 219). In the cold wâth 
alkali and copper salt, the hydroxyamidosulphate becomes 
oxidised at once to sulphite, sulphate, nitrous oxide, and water 
with reduction of the cupric hydroxide {op. cit., 225), and when 
heated with cupric chloride it reduces the latter to cuprous 



498 DIVEES AND HAGA : DECOMPOSITION OF 

chloride, becoming itself converted into sulphur dioxide, sulphate, 
nitrous oxide, and water. Mercuric nitrate oxidises hydroxy- 
amidosulphate more completely, but ferric chloride seems to act 
like copper sulphate, and liberates sulphur dioxide. 

An alkali hydroximidosulphate is also decomposed by copper 
sulphate, but not so easily, for it can be beated with it at 100° 
for a short time without change, and only decomposes (but 
then suddenly) some degrees above that temperature, yielding 
the products which a hydroxyamidosulphate gives, together with 
sulphuric acid from its hydrolysis into that salt. 

Although the presence of much sulphuric acid prevents the 
action of copper sulphate on a hydroxyamidosulphate, the acid 
in moderate excess has but little effect. 

Sodium hydroximidosulphate, if kept with care, decomposes 
only very slowly in a way which has hitherto been obscure 
(this Journ, 7, 45), but if considered in connection with the 
action of copper sulphate it may be regarded as essentially the 
same as that brought about by heating it in solution with that 
salt. For, the decomposed hydroximidosulphate contains, besides 
acid sulphate and hydroxyamidosulphate, both a little gas 
(nitrous oxide or nitrogen) shut up in its pores which escapes 
when the mass is dissolved in water, and also a little amido- 
sulphate, which can be separated from the other salts by pre- 
cipitation with mercuric nitrate (this Journ., 9, 242, also 229, 
230). 

The decomposition of hydroxyamidosulphates by copper 
sulphate is also in evident relation with the gradual decomposi- 
tion of impure hydroxylamine hydrochloride, particularly when 
ferric chloride is among the impurities, water, nitrous oxide and 



HYDROXYAMIDOSULPHATES BY COPPER SULPHATE. 499 

ammonia (in place of amidosiilphiiric acid) being the principal, 
if not the sole, products. 

There is a very marked difference in the proportions of the 
products of decomposition between a hydroxyamidosulphate and 
a hydroximidosulphate, but this seems to be owing merely to 
the fact that the temperature of the decomposition is different, 
for according as hydroxyamidosulphate is heated slowly or 
rapidly the proportions of the products of decomposition deviate 
from or approach those which obtain when a hydroximidosul^^hate 
is decomposed, this only taking place at a temperature above 100°. 

As little as one-tenth of an equivalent of copper sulphate 
has been found to suffice for the complete decomposition of an 
alkali hydroxyamidosulphate, the copper sulphate not being- 
consumed in the change it effects ; this allows of the decomposition 
being to a great extent carried out at the boiling temperature, 
when again the result approaches that observed where hydrox- 
imidosulphate is the salt decomposed. Even much less than the 
amount above named will effect an almost complete decomposition 
but that the quantity of the catalytic agent cannot be very 
greatly reduced seems to be due in part to the simple hydrolysis 
of some of the hydroxyamidosulphuric acid set free by the 
copper sulphate daring the prolonged heating here necessary, 

Since the cupric salt suffers no reduction, it will be seen 
that one part of the hydroxyamidosulphate becomes reduced to 
amidosulphate by yielding oxygen for the oxidation of the other 
part to water, sulphate and nitrous oxide. The following equa- 
tion shows that the hydroxyamidosulphate may change by 
cumulative resolution, half into a reduced product (amidosulphate), 
and half into oxidised products together equivalent to the non- 
existent dihydroxyamidosulphate : 



500 DIVERS AND HAGA : DECOMPOSITION OF 

(1). 2Cu(H2NSO,), = N20 + H20 + H,S04 + CuS04 + 
Cu(HoNS03)2 = Cu(H2NS05)2 + Cu(H2NS03)o. 
Such an equation expresses much of what happens in the de- 
composition of a hydroxyamidosiilphate at a lower temperature, 
but even in this case, and much more in the decomposition of 
a hydroximidosulphate by copper sulphate, where the tempera- 
ture is higher, a third molecule decomposes in another way. 
The result is that the free sulphuric acid shown in the above 
equation gets neutralised, and the third molecule of hydroxyami- 
dosulphuric acid yields neither sulphate nor amidosulphate, all 
its sulphur being eliminated as dioxide, its nitrogen as nitrous 
oxide, and its hydrogen as water thus reverting to sulphurous 
and hyponitrous acids, just as it does under the influence of an 
alkali (p 497) adding to equation (I) that of Cu(HoNSOj),=N,0+ 
2H,0 + 2SO. + CuO, we get (2), 3Cu(H,NSOJo=2N.>0 + 4H,0 + 
2S02+2CuSO, + Cu(tLNSO,)„ with products free from acid. 

It is possible to express the decomposition of hydroxy- 
amidosulphate differently, by making nitrogen one of the pro- 
ducts in place of nitrous oxide, thus : 

(3)3Cu(H2NSOJ.>=2No + 2HoO + 2HoSO,+ 2CuSO,+ Cu(H,^'S03),; 
(4) Cu(H2NS04).2 = K + 2H2O + SO2 + CUSO4. 

In (3) sulphur dioxide is not a product whilst in (4) it is. 
AVhether, however, nitrogen is formed, even in small quantity, 
is doubtful. Along with the nitrous oxide soluble in alcohol, we 
found a little insoluble gas— about 4 per cent, by volume of the 
wiiole gas, — but we are not prepared to assert that this was not 
due to air in spite of the precautions we took to expel all air 
from the apparatus by carbon dioxide before the decomposition. 
It will be seen from the equations that, with nitrous oxide as a 
product of the decomposition, the sulphur appearing as sulphate 



HYDROXYAMIDOSULPHATES BY COPPER SULPHATE. 501 

equals that as amidosulpliate, whereas, with nitrogen as a pro- 
duet, tlie sulphur as sulphate is double that as amidosulphate in 
(o), whilst in (4) there is none as amidosulphate. Now, in the 
observed decompositions of hydroxyamidosulphate the sulphur as 
sulphate has been found equal, on the average, to that as 
amidosulphate, a result showing that within the limits of accu- 
racy of the somewhat complex analytical work, no nitrogen is 
generated. 

Although, when using copper sulphate or copper hydroxy- 
amidosulphate, no change to cuprous salt is observable, the 
reduction of cupric chloride to cuprous chloride points clearly 
to the activity of the copper salt as a ' carrier of oxygen ' from 
one molecule of the hydroxyamidosulphate to another. 



Results and Method of the Quantitative Experiments. 

The results of the experiments are given, not in the order 
in which they were obtained but in that of the growth in 
quantity of the sulphur dioxide produced. 

In an experiment in which copper hydroxyamidosulphate 
was heated very slowly, so as to carry out the decomposition at 
as low a temperature as possible (boiling the solution only at the 
end in order to expel the last portions of sulphur dioxide), re- 
sults were obtained which agree sufficiently well with those 
calculated on the assumption that 3.7 per cent, of the salt gives 
all its sulphur as dioxide, its hydrogen as water, and its nitrogen 
as nitrous oxide, whilst the rest of the salt decomposes accord- 
ing to equation (1) : 



502 DIVEES AND HAGA : DECOMPOSITION OF 

Sulphur as dioxide ; as trioxide and amidosulphate. 

Found 3.5 96.2 

Calc 3.7 96.3 

An experiment with sodium hydroxyamidosulpliate and its 
equivalent of copper sulphate, gave results indicating that about 
0.3 per cent yielded all its sulphur as dioxide, the rest of the 
salt giving sulphur trioxide (sulphuric acid) and amidosulphate 
(equation 1) : 

Sulphur as dioxide ; as trioxide ; as amidosulphate ; as acidit}-. 

Found .5.5 46.0 48.0 21.6 

Calc 5.3 47.4 47.4 21.0 

Copper hydroxyamidosulpliate in four experiments gave re- 
sults agreeing nearly with the assumption that 13.2 per cent, of 
the salt gave sulphur dioxide, the rest decomposing according to 
equation (1) : 

Sulphur as dioxide ; as trioxide ; as amidosulphate : as acidity. 

Found 13.0 43.0 43.6 11.1 

„ 13.0 43.3 43.2 

„ 13.1 86.6 

„ 13.3 86.5 

Calc 13.2 43.4 43.4 15.1 

In anothar experiment copper hydroxyamidosulphate gave 
the following results, as against calculation for 15.4 per cent, to 
decompose so as to yield its sulphur as dioxide : 

Sulphur as dioxide ; as trioxide ; as amidosulphate ; as acidity. 

Found 15.1 42.9 41.6 10.7 

Calc 15.4 42.3 42.3 13.5 

In one more trial, copper hydroxyamidosulphate decomposed 
nearly as if 16.6 per cent, of it yielded all of its sulphur as dioxide : 



HYDROXYAMIDOSULPHATES BY COPPEPw SULPHATE. 503 

Sulphur as dioxide ; as trioxide and amidosiilphate ; as acidity. 

Found 16.3 83.3 9.3 

Calc 16.6 83.4 12.5 

A solution of potassium hydroximidosulphate heated with 
very little more than its equivalent of copper sulphate, gave 
results showing that 25 per cent, of the salt yielded all the 
sulphur of the hydroxyamidosulphate coming from it by hydro- 
lysis, as sulphur dioxide : 

Sulphur as dioxide ; as trioxide ; as amidosulphate ; as acidity. 

Found 2.5.2 37.3 37.0 5.5 

Calc 25.0 37.5 37.5 6.25 

A solution containing sodium hydroximidosulphate and 
copper sulphate decomposed in two experimenis, in such a way 
that about 28 per cent, of the hydroxyamidosulphate sulphur 
became dioxide : 

Sulphur as dioxide ; as trioxide ; as amidosulphate ; as acidity 

Found 27.6 36.9 35.0 4 

, 28.0 36.2 3.5.8 5.8 

Calc 28.0 36.0 36.0 4 

The numbers in the above table stand for parts per hundred 
of the sulphur of the total hydroxyamidosulphate decomposed, 
and not of the sulphur of the hydroximidosulphate even when such 
a salt has been that experimented with. The ' acidity ' sulphur 
is calculated as if the acidity is due to sulphuric acid, not 
amidosulphuric acid. The * trioxide ' sulphur is that of the 
sulphuric acid and copper sulphate yielded by the decomposition. 
The differences between the calculated quantities and those 
found must be largely attributed to imperfect estimation ; they 
cannot be due to error in theory^, because no other explanation 
of the change than that adopted is possible. In a copper-salt 



504 DIVEES AND HAGA : DEC03IP0SITI0N OF 

solution mixed with rancli barium sulphate, it was not easy to 
titrate acid with lacmoid pajoer as indicator. The separation of 
sulphate and amidosulphate is not a simple process, especially 
when much sulphate is present derived from sources other than 
the reaction to be dealt with. 

The salt employed in the experiments was either copper 
hydroxyamidosalphate, or sodium hydroxyamidosulphate with 
copper sulphate, or one of the alkali hydroximidosulphates with 
copper sulphate. 

1. A solution of the copper salt, containing onl}^ a very 
little copper sulphate was prepared from normal barium hydroxy- 
amidosulphate and copj^er sulphate, the barium salt (this Journ., 
3, 213, 216) liad to be prepared as wanted, because of the 
instability of the hydroxyamidosulphates. The strength of the 
solution was determined by a barium estimation (hydrolysis in 
sealed tube and weighing of the barium sulphate). Copper sul- 
phate in slight excess and carefully weighed was added to the 
weighed solution of the barium salt, and the copper hydroxy- 
amidosulphate at once used without filtering off the barium 
sulphate. 

2. Sodium hydroxyamidosulphate solution was prepared just 
before use by hydrolysing a centigram-molecule of the hydrox- 
imidosulphate by adding to its solution a minute and known 
quantity of sulphuric acid (this Journ., il, o) and to it was 
added after neutralisation with sodium hydroxide half a centigram- 
molecule of copper sulphate. 

3. Potassium or sodium hydroximidosulphate in the quantity 
of a centigram-molecule was dissolved and directly heated with 
a half molecule in centigrams of copper sulphate. 

The solution (either 1, 2, or 3) being in a small flask 



HYDBOXYAMIDOSULPHATES BY COrPER SULPHATE. 505 

connected with a tube receiver holding bromine water kept 
cold, was heated, sometimes quickly, sometimes slowly, either by 
a spirit lamp or in a bath of sulphuric acid, the solution being 
finally boiled for some minutes, so as to drive all sulphur dioxide 
into the bromine water. Before heating, air was removed from 
the apparatus by a current of carbon dioxide. In one experiment 
the apparatus was made entirely of glass. The oxidised sulphur, 
dioxide was weighed as barium sulphate. 

The boiled-out copper solution was titrated with X/10 soda 
(free from sulphate), using lacmoid paper as indicator. The 
imperfection of this operation was proved beyond doubt on cal- 
culating out the nature of the changes which had occurred, but 
it was serviceable and the best available under the circumstances. 

To the boiling hot solution and precipitate of barium sul- 
phate, barium chloride was added in excess, the total precipitate 
collected, well washed, and transferred to a pressure tube in 
which it was heated with hydrochloric acid for three hours at 
150°. The barium sulphate was again washed on the filter, then 
ignited and weighed. The second filtrate and washings contained 
sulphuric acid, the quantity of which was estimated as barium 
salt. To make this part of the analytical process intelligible, it 
must be explained that barium amidosulphate, although itself 
quite soluble in water, is partially precipitated along with barium 
sulphate, even in presence of hydrochloric acid (this Journ., g, 
283). At 150°, the precipitated amidosulphate hydrolyses, yielding 
barium sulphate and ammonium sulphate in molecular proportions. 

The copper filtrate from the crude barium precipitate was 
evaporated to a small volume, heated with hydrochloric acid for 
some hours at 150°, and mixed with barium chloride. The 
precipitated barium sulphate represented the principal quantity 



oOG DIVERS & HAGA : DECOMPOSlTIOxNT OF OXYAMIDOSULPHATES. 

of amidosulphate sulphur, the full amount of which was ascer- 
tained by adding to it twice the quantity of that in the ammo- 
nium sulphate extracted by hydrolysis from the crude barium 
precipitate. The sulphur from the hydroxyamidosulphate, ob- 
tained as sulphate, w^as found by subtracting from the total the 
sum of the quantities of sulphur present as (a) copper sulphate 
taken ; (b) barium sulphate from the hydrolysed barium amido- 
sulphate which had been precipitated along with the barium 
sulphate by barium chloride ; (c) sulphuric acid added for 
hydrolysing the hydroximidosulphate, when that salt had been 
started w^itli ; and (d) in the same case, sulphuric acid resulting 
from the hydrolysis of the hydroximidosulphate to hydroxy- 
amidosulphate. 

Hardly any attempt was made to estimate the amount of 
nitrous oxide liberated. To do so would only have been useful 
as a check on the accuracy of the determinations of the amido- 
sulphate, and for that purpose the two substances would have 
had to be estimated in the products of one experiment. This, 
it did not seem possible to do. An experiment in which hydroxy- 
amidosulphate was decomposed gave 55.3 per cent, of the 
nitrogen as nitrous oxide, as against 56. G calculated from the 
equation most in accordance with amidosulphate and other sul- 
phur determinations. The method of measuring the nitrous oxide 
and nitrogen was to expel air from the apparatus by a current 
of carbon dioxide continued for some time, and then heat the 
copper salt and boil out the gases which were collected over 
mercury and potassium hydroxide and measured. The alkali 
was then replaced by absolute alcohol to dissolve the nitrous 
oxide, and the residual gas measured. 



Observations on the Development, Structure and 
Metamorphosis of Actinotrocha. 

By 
Iwaji Ikeda, Bigahishi. 



Wlfh Plates .YA'T'-.YA'X. 



Introductory. 

Since tlie discovcrv of ActlnotrocJia hy Johannes Mulleiî 
in 184'.), tliir^ peculiar larval form and its mother animal, 
Phoronis, have been made the subject of investigations by many 
distinguished authors such as Wagener ('47), Gegenbaur ('54^ 
Krohn ('54), Schneider ('62), ]\[ETScnNiKOEF ('72, '82), E. 
B. Wilson ('81), and Foettinger ('82). Among more recent 
writers Caldwell ('85), INFc'Intosh ('88), BENTL\:\r ('88), 
Roule ('90, '96), Cori ('91), and E. Schultze ('97) may be 
mentioued as having published impoi-tant contributions ; whiU- 
Masterman ('97) has ]nade quite an elaborate study of the 
animal with the view of establishing its relationship to Balano- 
fj/o.<i<in.s and the Chordata in general. As, however, in spite of all 
these works there still existed many gaps and unsatisfactory points 



508 I. iKEDA : 

in our knowlodo'e of tliis interei^tino' animal, tlio investia;ation, 
of wUieli an account is given in the followiui;- pages, was under- 
taken, and tliongli the residts are far from exhaustive, I liope 
they will hel]) to advance our knowledge of the sul^jeet. 

^[v studv was hegun in tlie sunnner of LSDS during a stav 
at the ]\risaki JMarine Biological Station and later was continued 
at that Station as well as in the Zoological Institute of the Science 
College. 

At Al)uratsubo, a small inlet close to the Station, is found 
a species <>f FJioronix, which has been named hv Dr. ()ka ('97) 
r. ijimai.''' Its colonies adhere to the overhanging ledges of rocks 
near the shore. As the water at the place is always cahn and at 
low tides recedes so as to almost expose the ledges, the animals 
can be easily collected. During the greater j^art of the year, eggs 
and young embryos, clustered together, in what may conveniently 
be called ernhryonal 'masses, are found adhering to the lophoplioral 
crown of the adult, one on each side of the median line. These 
furnished materials for the study of fertilization, segmentation and 
the earlv larval stashes. The larva' in the Actinotrocha stage are 
found swimming in the inlet and are caught with the surface net. 
As will later be fully described, there occur four kinds of the 
larvce, which ]io doubt represent as many species, including the 
common Phoronis ijimai. 

The specimens, both adult and larval, were killed with the 
saturated solution of corrosive sublimate in l^o acetic acid or with 
Flemming's fluid. Of the various colouring methods tried on the 
sections, double-staining with eosin or safranin and Delatleld's 
hcTmatoxylin gave the most satisfactory results. 

*For a discussion of the status of this species, see Supplementary Nofes. 



ON DEVELOPMENT ETC. OF PHOllOXIS. Ö09 

Before proceeding further, I beg to tender my sincere thanks 
to Professors Mitsukuki and Ijima for their kind supervision 
of my Avork and for their ])aiiistaking revision of my manuscripts. 

Contents. 

I. The Eurly Development uf the l*liuruiii,s ]jarva. 

((. Xotes on Fertilization. 

b. Notes on Segmentation. 

c. Gastrulatiou and Mesohlast-Formatiou. 

(1. Further Observations on the Development of the Lirva. \\\) to 
tlie Period when it becomes free-swimming. 

II. The Structure of Actinotrocha. 

a. The External Appearance of Actinotroclui. 
h. The Internal Structure of Actinotrocha. 

1. Body-Divisions and Body-Cavities. 

2. Organs of Ectoblastic Origin, 

3. Organs of EntoLlastic Origin. 

4. Organs of Mesoblastic Origin. 

III. Metamorphosis. 

IV. Supplementary Notes. 



I. The Early Development of the Phoronis Larva. 
a. Notes on Fertilization. 

Phoronis is, as is well known, a hermaphrodite, in which hotJt 
iliG mah and female sexual elements mature at nearhj the same 
time. Put few authors seem to have studied tJu> animal duriug its 
breeding season, so that our knowledge of its sexual organs and, 
consequently, of its fertilization has remained very imperfect, as 
was pointed out l)y Cora ('91). The only existing statement as 
to how and wliere fertilization is accomplished in Phoronis is that 
of Kow'ALEAVSKY ('07). Tlus author thought that fertibzation 



510 I. IKEDA : 

took place in the bodv-cavities, and accordingly, as CoRi reinai'ks, 
lie nuist have believed that ;r;elt-lertilization prevails in Phoro- 
nw. Cori considers this as hi^hl}^ improbable, bnt does not 1)rini;- 
forward any positive facts in contradiction of it, his inference 
being di'awn solely from facts ol)servcd in other marine jMetazoa. 

Since Kowalewskv's valuable researches ('67), it has 
generally been accepted that the ne[)bridia serve also as oviducts. 
Thus Uenham says that he saw an ovum iittached to one side 
of the ne])hridial funnel and further mentions that Kowalewsky 
observed eggs moving through the nephridial canal towards the 
exterior. Unfortunately both ol)servers failetl to elucidate what stage 
of develo})ment these eggs are in. 

In PJioronis ijimal mature sexual elements are constantly 
discovered tbroughout aljout oue half of the year (frimi November 
to May or June). ]]y carehdiy examiniug a living colony ofthat 
species during this])eriod, it will s(X)n be j)erceived that some indivi- 
duals ditter slightly from the rest in the aspect of the fool or 
body. We see in them a moniliform series of small white specks 
shining through the skin in the uppermost part of the bod_y. 
These are the ova ready to escape to the exterior through the 
nephridia. It nuist have been such individuals that were observed 
by Kowalewsky and Bexham. The Injdy-cavit^^, in which the 
ova lie, corresponds to the rectal chamber near the anterior end 
of the body. I have endeavoured to ascertain whether these ova 
are fertilized or not, and have at last succeeded in ascertaining 
that they are in a stage prior to the extrusion of polai- globules, 
— the primary oocytes, in Bon'ERl's terminology. In the fresh 
state, they are spherical or somewhat elliptical in shape and ])er- 
fectly opaque by virtue of the abundant yolk-granules contained in 
the vitellus. It is characteristic of these ova that the nucleus, which 



ON UEVELOrMEXÏ ETC. OF THOKOXIS. 511 

i« situated, not in the eeiitrc, luit near the }teiiphery, is al- 
ways in tlie nieta- ur ana-phase of Karvokinesis (fii;". 17). Tn 
sncli an ovum the ehroniosonies are constantly fcund to he i^ix in 
nuiiiher, each ])ein_u- dinnh-hell shajted witli the t\V(j ends directed 
towards the poles. Fii;-. 18 represents a portion ot" the section 
pussin,!;' through the eipiatorial })lane (jf the nncleai' li,L;ure. It is 
evident that these eggs are in preparation for the extrusi(ai of tlie 
first polar glol)ule. As shown in the ahove figure, the finely 
granular protoplasm of the vitellus contains thickly and uniformly 
distriljuted yolk granules, which have a strong affinity for eosin. 

That the eggs in question are mature is further demonstrated 
Ity the fact that I succeeded in artificially fertihzing them and in 
reai'ing out of them normal endjryos which grew to certain ad- 
vanced stages of development. 

If we now examine the emlnyonal masses, which, as has 
heen mentioned, are found attached one on each side of the ten- 
tacular crown of the adult Phoronis, we find that the end)ryos 
which are farthest away from the nephridial pores are the most 
advanced in development and that they are found in successively 
younger and younger stages as we a})pr()ach the pores, until we 
reach such eggs as have just l)eeu fertilized or perhaps even such 
as have not yet Ijcen fertilized at all. But even the youngest eggs 
found in the mass present an appearance very different from 
those found in the l)ody-cavities, the former being invariably at 
a stage after the ex[)ulsion of one or two polar glof'ules. In the 
e'S'i taken frtan the mass and shown in liü'. VJ, two iiolar globules 
have already been formed ; these nvv situated close together just 
inside the vitelline mend)rane. 

On the other hand, if we examine by means of serial sections 
through the ])osterior region of an adult, where the stomach and 



ül2 I. IKEDA : 

the sexual orn-aus lie grouped together, a, uumljer of large eggs 
are fre<|uently tbuud, Hoatiug freely iu the eoelouiic fluid of the 
l)ody-eavities. These eggs do not differ in any res])ect from those 
in thi' ncphiidial region as regards tlie size, the appearance of 
the karyokinetie Jigure, or the nunil)er of chi'oniosonies. 

The facts aI)ove stated plainly point to the followijig con- 
clusion : — Tlie oo(joiii<i fall into lite body-car it ics by a dehiscence of 
the ovarian vjalU and Jiere derclop until they reach the stage of 
primary oocytes. These travel yradually upwards to the nephridial 
region, retaining memiwhile the nuclear figure formed for the 
production of the first polar globule. Reaching that region, the 
primary oocytes are destined sooner or later to be carried by way 
of the nephj'idia to the exterior, where tJicy become fertilized by 
spermatozoa from oilier individuals. 

Keserving an account of the s})ermatogenesis and ovogenesis 
lor a future occasion, I may here refer t(j a few facts observed 
by me relative to the process of fertilization. AVhen the two 
sexual elements arc ai'tiiicially brought togethe]', numl)erless sper- 
matozoa soon attach themselves to the surface of the ovum. About 
10 minutes afterwards, the lii'st })olar globule makes its appear- 
ance, followed soon afterwards by the second, ^[eanwdiile a small 
clear spot, probably marking the place where the male element has 
entered, ap])ears on the surface of the egg ; it is however ol)serv- 
al)le for only a very short time. Both ligs. 1*.' and 20 are 
sections of ova taken from the emlnyonal mass. The ovum given 
in fig. 19 is fully mature and ready to l)e fertilized ; close to the 
j)olar globules vests the large female pronucleus. The ovum 
re})resented in fig. 20 belongs to a stage of fertilization in which 
the two ]ironuclei stand closely side by side. The larger female 
])ronucleus has a nuclear meml)rane irregular in contour. The 



ON DEVELOPMENT ETC. OF PITORONTS. '"JIS 

iiitonscly stained eliromatin pieces are in l»oth nuclei dispersed 
witliont :iny apparent order througliont tlie finely granular nncleai- 
substance. At one sjiot outside the male jironuclens, tlierc is 
visil)lc a small and clear arclioplasmic (?) s]^ace surrounded l»y 
a set of exceedinii'ly fine radial rays. The two polai- liiohulcs of 
this eg,!;" were distinctly yisihle in other sections which haye 
not been fii;ur(Hl. 



h. XoTES ON Segmentation. 

Our knowledge of the mode of segmentation in JVinron'is is 
far from 1)eing satisfactory. Metscitnikoff ('82) gives no account 
of the process. Foetttnger ('82), if one may judge from his 
figures, seems to haye seen the egg undergoing to taland unerpial 
segmentation. According to CALD\yELL ('82), the segmentation 
^'' proceecU with consiclerahle regularity " [I.e., ]>. o74) ; Roule ('90) 

says 'T ovule féconde .mhit une segmentation totale fort reguliere " 

[le., ]xll47). E. Sciiultze ('97) simply says ''Ich mh das Ki 
sich total und unäqual furchen ' {I.e., p. 6). 

jVIy ohseryations of the process were made on eggs found in 
the eml)ryonal mass as w^ell as on those artificially fertilized. As 
the former showed comparatiyely rarely the earlier stages of 
the segmentation, it was necessary to haye recourse to the latter 
for filling uj) the gaps of ol)seryation. 

Soon after the formation of the second ])olar glohule and the 
disappearance of the micropyle-like spot the first cleayage line 
makes its appearance, passing on one side of the ])olar glohidcs 
(figs. 1 and 21). At this stagx» T can not ])erceiye any ditlercnee 
in size and structure l^etween the two hlastomeres. The second 



.■)14 I. IKEDA : 

cloiivngo ])lanf' passes at riglit angles to the first (fig. 2 h). It 
is a remarkable peculiarity of Phoroivh eggs that two sister ])las- 
toiueres derived hy the division of a mother blastomere, never 
undergo the next division siinultaneonsly, so that l)etween any two 
consecutive stages having an even mnnl^er of blastomeres there 
intervenes an intermediate stage with an odd number of the same. 
This ])henoinenon oecui's even at the second cleavage ; thus just 
before the egi;' attaiiis the four-cell stage, there exists a stage of 
tlircc cells, siicli as is seen in fig. 'Id. Among tlie later stages, 

lli(tse of ?">, 7. *> cells are of constant occurrence. C<mseqn- 

ently it is scarcely admissible to say that the segmentation ])ro- 
ceeds with considerable regulai'itv. 

Caldwell ('82) has assei'ted that the first differentation of the 
future blastoderm into the ectoblast and entoblast is observable 
as early as in the four- cell stage. He says: " ^/ the stage of 4 
sera/tientafion-spherrs a dirhion into tiro >iiiiallfr clear and 
two larger opa(j}ie celh Indicate'^ the future ectoderm and endo- 
derrn " (/.r., ]>. ?uA). At the corresponding stage of riioronis 
ijiniai I have not been able to discover any appreciable difterence 
in the size of its cells (see fig. 2 />). Following the 4-cell stage, 
the division of the blastomeres in the ecpiatorial ]>lane puts the egg 
on the way to the 8-cell stage. According to my own observations, 
the alxn^e mentioned difference in size of the blastomeres becomes 
iii'st ]X'rceptible at this stage. Fig. 3 shows a side view of an egg 
with S blastomeres ; it will be seen that the up])er four Idastom- 
eres are veiy slightly smaller than the lower four. I could not 
liowever, at that period, recognize any difference in the cell-con- 
tents of the two classes. 

The iri-egularity of division, which, as 1)efore mentioned, be- 
comes more and more pronounced as segmentation advances, tends 



ox DEVELOP^IEXT ETC. OF PHORONIS. 515 

to ii;r:i(lnnlly ohsr-nvo tlie orderly ;irr:iii<;-oinont of tlie ec'll^<. At tlie 
ir)-cc'll stiiiTO tlio reo-ular arnniii'emont is still, thono-li less distinctly, 
iiiaiiitained, while at the 32-pell stage it is quite distiirlted (fig. 4), 
From this period on, the polar globnles can no longer be detected. 

In the earlier stages of segmentation, the blastomeres are 
found in close contact witli one another, leaving no noticeable 
space or segmentation-cavity between them. After they have 
increased to about 32 in number, the blastocœle and its opening 
to the exterior {ùii:. 4, /V.r.) become recognizable. The emliryo 
at the morula stage is somewhat oblong in shajx- and has 
a (juito spacious blastocœle, and the blastoccelic pore (/>/.c.) 
is distinct on the ventral side (fig. 5). However, this pore dis- 
appears at an advanced morula stage, and a])parently the vitelline 
membrane also, nearly simultaneously with it. At any rate both 
have altogether passed out of sight at the next stage, that of the 
blastula. In fig. 23, which represents a median section of a young 
morula (the outline of which has undergone mutual compression 
by the crowding together of emlu'vos), the pore (Ij/.c.) is cut through 
and appears as a slit-like passage between two of the bounding 
blastomeres. 

In the blastula (fig. 2-3) the wall consists of cylindrical cells 
and encloses a tolerably wide blastoca^lic ca\ity, which is now at its 
greatest development. In this stage, the bilateral symmetry of 
the future larva is already established. It has an oblong ]>lano- 
convex form, the fiatteued face of which corresponds to the future 
ventral f;ice ; and its ends, one somewhat bi'oader than the other, 
indicate respectively the future anterior, and posterior, ends. 
The cells of the wall are all cylindrical in form, as shown in fig. 
2Ö, those of the ventral side l)eing slightly larger than those on 



016 T. IKEDA : 

tlio convex dorsal side. The nuclei in all tlio Idnstodermal cells 
are always situated in a ]>eri plierai position. 

Plasmic corpii^cle><. — A notewortliy fact with regard to the 
l)lastula is that in its older stage a certain nunil^er of small and 
non-nucleated ])lasniic spheres is almost constantly met with in the 
blastocœlic cavity (fig. 2ö, pl.co.) These have l)een first described 
hy FoETTiNGER uuder the name "corpuscules mésodermiques." 
According to this author, these corpuscles are free nuclei imbedded 
in a connnon ])r()t(»]dasmic mass which is supposed to fill up th(> 
Itlastocd'le, each coi'jnisclc bocoming a mesoblast cell, after appro- 
priating to itself a certain portion of the surrounding protoplasm. 
This view of Foettinger, which certainly can not be accepted 
at the present day, was, I believe, partially due to the then 
defective technique. His method consisted in pouring dilute acetic 
acid over the living eml)ry(», and this, as the author himself was 
well aware, is highly detrimental, in that it frequently In'caks up 
llie blastomeres into fragments. The corpuscles described by him 
fi'(jm so early a stage as that with only 8 ))lastomeres must have 
l)een simply produced l)y fragmentation, the result of his drastic 
treatment. The connnon protoplasmic mass supposed to be present 
in the blastocœle, was probably nothing l)ut a coagulum. 

Again the mesodennzellen which Metscknikoff ('82) found 
m the blastocœlic cavity of the blastula are certainly not true 
mesoblast cells l)ut rather certain spheres similar to Foettin- 
(ier's " corpuscules mésodermiques," as was rightly pointed out l)y 
Caldwell. Eecently E. Schultze ('97) pul)lished a short paper 
entitled " Ueher die 3[esodcrinhiIdung bel Phoronis,'" in which he 
writes as follows: — ''Schon, auf dem Stadium- der rundliehen 
Blastida sehen leir einige Jlesodermze/len im Blastoeed riuf sitzten " 



ON DEVELOPMENT ETC. OF THOKONIS. 517 

[I.e., p. 6). It seems to me that Schulze lias fallen into the 
same mistake as Metschnikoff. 

Lastly, Caldwell ('82) has entertained a view quite different 
from those of other writers. According to him, the bodies in 
(juestion are not present as sueh in the blastoeœle, l)ut are in 
reality only the enl ends of l)lastoderm eells projecting into the 
cavity and as such of course have nothing to do with the true 
mesoblast. 

In the Phoronis studied by me, the plasmic corpuscles are 
present only in the highly advanced blastula (fig. 25, pl.co.). 
They are usually round in shape and very much smaller in size 
than any of the l)lastoderm cells, but as to structure, they do not 
show any deviation from the latter, except in the important res- 
pect that they have no nucleus. Although I tried with them all 
the available nuclear stains, the presence of any chromatic sul)- 
stance in them could in no instance be detected. In an earlier 
stage such a free floating sphere has never been met with. In- 
stead of it, some unusually elongate Ijlastoderm cells (jol. co., fig. 
24), such as were found by Caldwell, were discovered protruding 
their inner end into the blastocœle. The nucleus of these cells 
commonly lies in the periphery as in all other cells of the more 
ordinary sha]^e. In my opinion, the proximal ends of the elong- 
ate cells l)reak off from the main ce]l-l)ody and fall into the 
blastocœle, where they undergo degenerati(ai, breaking into ever 
smaller and smaller spheres. By examining serial sections of an 
advanced stage like that of fig. 25, it is easy to convince one's 
self that there exists no connection whatever l)et\veen the spheres 
and the blastoderm cells. The spheres, or the plasm ie corpuscles, 
are clearly distinct Ijodies and not mere ends of blastoderm cells 
cut olf in the process of microtomizing as was supposed by 



518 I. IKED A : 

Caldwell. Their small size and the total al)seiice of imelear 
substance make it easy to tlistingnish them from the trne meso- 
blast cells. 

The corpuscles are still frefjuently discovered in the l)]asto- 
cœlic cavity at the beginnini^' of gastrulation, together with a few 
mesoblast cells. But in an advanced gastrula they have whollv 
disappeared, })Ossibly having been absorbed \)y the Ijlastoderm cells. 
I think the temporary co-existence of the plasmic corpuscles and 
of the true mesoblast cells in the blastocœle of the gastrula, has 
led some previous authors to confound the two elements. As to 
the signihcance of the corpuscles, I can at present offer no opinion 
'^ uule-^s they be merely an excess of supply of nourishment analo- 
gous to food yolk " as has been suggested by Caldwell ('82, I.e., 
p. 18). 

C. GaöTKULATIÜN and m esq BL AST-Foil m ATION. 

In this section, I shall iirst describe what I conceive to be 
the tj'ue history of gastrulation and mesoblast-formation, and then 
pass on to a discussion of the views of other writers. The two 
developmental processes are so intimately related to each other, 
that it seems best to treat lliem together. 

First as lo external changes. The hilalei'nl synnnelry of the 
|>lano-convex Mastula becomes more clearly marked than hefore 
when the gastral invagination begins on the ventral or the flat- 
tened side. The initial depression occurs over the whole ^'entral 
wall, so that a saucer-shaped emljryo is produced. At first it is 
so shallow as to be perceived with difficulty in the surface view. 
Soon it deei)ens, becoming deepest at a point somewhat nearer to 
the 1)roader end than to the narrow qi\<^i of the embryo. The 



ON DEVELOPMENT ETC. OF rilOllONIS. ol9 

deepest portion may eoiiveiiientlv l)e called the central depression. 
Fig. 6 represents the ventral view uf an enil)ryo in which the 
invagination has 1)eeome visil)le from the ontside, the central 
depression being most deeply shaded in the iigure. In a slightly 
more advanced stage, as the original wide depression grows deeper, 
the external ojK'ning is gradnally drawn together and at a certain 
stage (lig. 7) Itecomes transformed into an ahiKtst triangular 
1)lastopore situated at a j)osition sHghtly anterior to the centre, as 
was the case with the central depression. The anterior side of 
the triangular l)lastopore is S(jniewhat ronnded by curving uniforndy 
outwards, while posteriorly the two other sides gradnally approach 
each other so as to meet at a point which may be called the apex 
of the triangle. Leading backwards from this apex, there runs in 
the mediaii line the so-called ])rimitive groove. This latter and 
also the triangular shape of the blastpore are occasioned, in my opi- 
nion, sim})ly by the blastopore, originally broadly open, becoming 
narrowed by the special activity in the lateral posterior parts of 
its posterior section. In other words, the cell-multiplication of the 
ectoblastic layer , is carried on especially in the last mentioned 
parts, so that there the i)ressnre, which is exercised by the ecto- 
blast towards the invaginated layer, is more marked than in the 
anterior and lateral Ijordcrs of the blastopore. As the result of 
the alxn'e phenomenon, the deHnitive blastpore is pushed further 
anteriorly, and consecjuenth', the archenteric cavity deepens in the 
posterior direction, as shown in fig. 29. The above consideratioji 
is snpported by the resnlts of actual measnrements of the size of 
the embryos concerned. In spite of the fact that the embryo has 
developed consideraldy the body-length does not show any signi- 
ficant increase, remaining all the while at about (J. 12 mm. on an 



520 I. iKEDA : 

average. This «liows that the growth is lost in the curvutiire of 
the body. 

AVhen the growing larva reaches the stage represented in 
fig. 8, tlie l)last|)ore assumes a narrow transversely directed, slit- 
like form. That portion of the larval body lying in front of the 
hlastpore — whidi is the })ersistent larval mouth — protrudes more 
or less prominently forwards and ventrally, so as to acquire the 
form characteristic of the prcoral lol)c of Actinotrocha. In such 
an advanced gastrula, the primary gut-cavity is well established 
and can be plaiidy traced through the wall in the surface view. 
If the larvœ of such an early stage of development l)e taken out 
of the em])ryonal mass and set free in water, they will swim al)out 
by means of the well developed cilia, which cover the whole 
external surface. 

Fig. U re})resents a side view of a larva, in which the ])re- 
oral lobe has grown to a very considerable size. The ncphridial 
pit, which is an ect(jblastic invagination just in front of the 
posterior end of the gut, is now distinctly visible from the out- 
side. In short, the larva, may l»e said to possess the inceptive 
characters of an Actinotrocha. 

I will now proceed to descril)e the internal changes accom- 
[)anying gastrulation. The earbest symptom of this process can 
1)0 seen in sections before it can be detected from the snrface. 
It consists at first in a peculiar dis2)osition of those blastodermic 
cells which constitute the ventral })ortion of the l)lastula wall. 
This ])ortion not only shows a shallow concavity, but also the cells 
comp()sing it become, as figs. 26 a and 2(5 h show, ii'i'egnlarly 
arranged on account of mutual pres«urc, as a result of which some 
of the cells are even forced out of file so as to fall into the 
Ijlasiocœle. These liberated cells have usually a round sha[)c and 



ON DEVELOP:\rEXT ETC. OF PITOROXIS. Ô21 

of course contain cacli a distinct nucleus. Sonic other cells are 
apparently in tlie process of being pushed out and have a cluh-likc 
shape, the narrowed end heini;' still inserted between the cells 
of the layer to he invaginated. A further symptom of incipient 
invagination consists in the circumstance that the nucleus in uK^st 
cells of this portion has no longer a peripheral jwsition, hut is 
situated in the middle or rather nearer to the inner end of the 
cells. Moreover, the nucleus is frequently met with in the form 
of the karyolvinetic figure which shows that the cells are dividing 
and increasing in numhei- in the layer to he invaginated. Tltr 
celU ini>^]i€x} out into the hladoccele arc nothing else than me><ohlast 
celh, so that it may he dated that the niesohJast-fornuition begins 
siinuUaneoudy uùth the gadrulation. 

At the beginning of gastrulation we can thus distinguish two 
parts in the blastoderm wall, riz., the mesentohlad and the ecto- 
hlad. The former corresponds to the whole of the portion to be 
invaginated, while the ectoblast forms all the remaining portion of 
the embryo. The mesentoblast is com])Osed of large and irregular- 
ly arranged cells, while the ectoldast is of taller cylindrical cells 
regularly arranged in a single row (figs. 2<) a and 2(3 h). The 
mesentoblast, as the name indicates, is destined to give rise to 
both the entoblastic and mesoblastic elements. The characteristic 
disposition of its constituent cells indicates its l)eing the source of 
mesoblast proliferation . 

The mesoblast ])rolifcration becomes more and more accen- 
tuated in activity as the gastral invagination gradually deepens 
(see fig. 27), ])ut the mesoblast cells thus formed do not adhere as 
a lining epithelium to the ectoblast, until what are called the ant- 
erior diverticula have been formed on both sides of the blastopore. 
The blastocœle, in which tlie mesoblast cells are at first loosely 



Ö22 I. IKED A : 

scattered ai »out, is liciicefortli liroatly reduced in extent and finally, 
as the development of the areheiiteron proo-resses, is almost o]»li- 
tei'ated, especially aloni;' tlie dorsal and lalei'al portions of the 
end)ryo where the ectohlast and the i;'ut come into direct contact 
with each other (see fii^s. 29 and oO/>). 

File's. 28 a-c show three cross sections throuii-h different parts 
of a larva of nearly tlie same stage as that represented in fig. 6, 
in which the invagination has hecome recognizahle in the surface 
view. Fig. 2S n passes thi'ough the central depression which be- 
comes gi'aduallv shallower posteriorly {i'liXi^- 2<S <?* and 2.S r). As 
these figui'cs show, the mesohlast cells are at this stage still being 
proliferatc<l uniforndy from evei'v part of the mesentoblast and do 
not yet form a lining epithelium to the ectohlast. When the blasto- 
pore has taken a triangular slia])e (fig. 7) and the primaiy arclien- 
teric cavity has somewhat bent itself towards the hind end, the 
]V)sterior ] »order of the blastopore has travelled a certain distance 
in an anterior direction. Tf we examine serial sections of this 
region, a narrow and shallow groove is detected running for a 
short distance immediately behind, and from, the meeting point of 
the lateral blastopore lips. Also at about this stage, a paired in- 
vagination, the anterior diverticula of Caldwell, a]ipears along 
the side of the anterior portion of the archenteroiL These points 
will become clear from a consideration of figs. :10 a-r, which 
are drawn from serial ti'ansverse sections of an eml)rvo 
slightly older than that shown in fig. 7. Tn fig. 30 a, showing 
the I'ight-hand side of the )>lastopore, we notice a lateral infolding 
{mit. dir.) of the archenteric wall a short distance inside the 
blastopore lip. Here the component cells are irregularly arranged 
and their entire disposition reminds one of the mesentoblastic 
layer. Indeed some indubitable mesohlast cells are found pressed 



ox DEVELOPMENT ETC. OF PIIOROXIS. 023 

against tlio tip of tlio divorticnlniii. Xo (lonl)t tlio mf'so])last is 
hero arising, not l)y dii-ect coll multiplication, but l)y the ])usliiug 
in of tlio colls of tlio clivorticuluni. This is more clearly illustrated 
in fig. 31, "which shows a transverse section through the lilasto- 
pore of a more advanced larva ; hero the mesol)last cells almost 
fill uyi the blastocodic cavity on both sides of the blastopore. 
In fig. 30 b, a transverse section just l)ehind the closure of the 
blastopore, the most anterior portion of the primitive groove before 
montionod is cut across. Here the wall of the o-roove underlvinçr 
the gut is formed of uiutually compressed cells, some of which are 
evidently migi'ating into the blastocœle (on the left-hand side in 
the figure). If the sections are followed further posteriorly, the 
groove still persists, but no mesoblast cell in the actual immigrating 
process can be discovered, although there are those which have 
been previously pushed out and are now floating between the two 
primary germinal layoi's at this region. Still more posteriorlv the 
groove entirely disaj^pears and the entoblastic and ectoblastic 
layers are separated from each other by the com|)ai'atively wide 
blastocœlic cavity (fig. 30 r). At this stage, therefore, the greater 
p'.rt of the archenteric wall has ceased to conti-il)ute towards the 
mesoblast-formation ; in other words, it has lost its mesentoblastic 
nature. The mesoblast is now being produced only from two 
limited regions, viz., anterior divei'ticula and the ventral groove. 
In a slightly more advanced larva, the ventral groove is still 
present for some distance immediately behind the blastopore, but 
the layer wliich forms the groove has ontirelv ceased to give rise 
to mesoblast colls (fig. 32, which is taken from a transverse sec- 
tion very near the l)lastopore). It ap]K'ars to me that this groove 
is to be regarded as Init the posterior portion of the original 
mesentoblast, wdiich, owing to the flict that the central depression 



524 T. IKEDA : 

is eeeentrically jJnced nearer to tlie niiterior end, has to traverse 
a longer distance l)efoi'e it can l)e reflected inwards, and thns on 
its inward conrse lags beliind the anterioi- and antei'o-lateral por- 
tions. Eventually all the cells of the wall of the groove that are 
left behind after proliferating the mesohlast cells, are without 
doul)t invaginated and forai a part of tlie ent<)1)last. The groove 
then entii-ely disappears. I could not discover any remnant of it 
in any pai't of the posterior region where, according to Cald- 
well, the ectoblast and the entoldast are said to stand in fusion 
to give rise afterwards to the aims. In such an advanced stage, 
the antei'ior diverticula have also ceased to give off mesohlast 
cells and have become straightened out, their walls acquiring a 
normal epithelial character (entoblastic). 

From the facts above adduced, it may 1)e concluded that 
hoik the anterior <liverticiiJn and the ventral groove, présent at <i 
cerkitn developmental Hage of the Phoronis embryo-'^, are remnanti^ of 
tlie original mesentohlast ivhich at an. earlier dage occnpied the 
the whole extent of the gaxlral invagination. TJicy are, therefore, 
merely temporary, and destined, sooner or later to split into meso- 
blastic and entoblastic cells. 

As will be seen in figs. oO a-e, the ectoblast and the archen- 
teric walls are l)rought together into such close contact, especially 
along the dorsal and lateral regions that scarcely any interspace 
is left between them. In the embryo given in fig. 8, the cavity 
of the rudimentary preoral lobe is filled with mesohlast cells pro- 
duced from the original mesentoblastic layer. So far as I can make 
out, these show no difference whatever from those proliferated 
from the anterior diverticula : both are indistingm,shably mixed 
together. Though most of the mesohlast cells in the preoral lobe 
lie loose during the active period of the diverticula, there are 



ON DEVELOPMENT ETC. OF PHOKONI«. 025 

iuiiud a few that have already apposed themselves ilatty to the 
ectoblast (see fig. 29 j, while the cavity behind the blastopore still 
remains without a mesoblastic lining. This last condition persists 
till the period when the nephridial invagination makes its appear- 
ance. The state of things in question is to be seen in. lig. 20, 
which represents a median sagittal section through an embryo (jf 
nearlv the same staçre as i'ar. cS. 

Soon afterwards the anterior diverticula and the ventral 
groove entirely disappear and the preoral lobe begins to bend more 
distinctly downwards. Meanwhile an unpaired ectoblastic invagi- 
nation appears at the posterior end of the larva, on the ventral 
side of the blind end of the now greatly elongated gut. It appears 
at first as a shallow depression (fig. 33, 7iep. p.) of purely ectoblastic 
nature, having nothing to do with the mesentoblast. It is from 
this invagination that the future nephridia of Actinotrocha de- 
velop and hence I shall call it the nephridial pit, in preference 
to the name " anal pit " of Caldwell, who tor the first time des- 
cribed this structure. I have very frequently noticed signs of 
vigorous cell-division in the cells of the pit wall, evidently only 
for enlargement of the pit itself, since the axis of the karyokinetic 
spindle is always placed paratangentially to the wall. I have 
moreover often noticed peculiar ectoblastic cells round in shape and 
in process of multiplication, situated just outside the edge of the 
entrance to the pit (fig. 33). 

In larvte of the stage of fig. *) tlie ne])hridial pit can l)e well 
seen in surface views. This stage further attracts our special at- 
tention on account of several inq)ortant developmental processes 
taking place in it. First to be noted is the fact that from the 
posterior end of the primary gut a small and short evagination 
protrudes itself touching the ectoblast with its blind posterior 



ü26 I. IKEDA : 

end. This hollow protu1)er;ince is the riuUmeiit of the intestinal 
canal of Actinotroeha. In loni^itudinal section it is shown in lig. 
37 {hiL). 

In lii;'. ol, ]'e])rcsentini;' a sliiihlly o))li(|ne frontal section of 
a larva of nearly the same stafj;e as that of iig. 0, we see below 
the |)it-like ne])hndial sac, which is (|uite free from the gnt. The 
ectoblastic wall of the jjrc.'oral Ljbe is at this stage somewhat 
uniformly lined with flattened mesohlast cells, while in the cavity 
heliind the blastopore the mesohlast cells are for the most part 
freely scattered, though a few^ have already begun to arrange them- 
selves against the ectoblast laver in this region. In fis;. So, a 
transverse section through thc^ p(jsterior end of u larva of nearly 
the same stage, the nephridial pit appears as a single flattened 
sac {ncp. p.) lying in front of the intestine [i/d.) ; the ectoblastic 
Avail is internally lined with a few isolated and flattened mesol)last 
cells. In a slightly nujre advanced stage, the ectoblast behind 
the blastopore, and in a less complete degree the gut wall 
also, shows a similar mesoblastic lining, though a few mesohlast 
cells still remain free, esjiecially in front of the nephridial 
sac. 

In order to facilitate comparison with the statements of other 
writers, I will here add a few words on the change of form under- 
gone l>y the ne])hridial pit. AVhen in a larva slightly older than 
that of fig. *J, the preoral lol)e and the future intestinal portion of 
the gut have become considerably elong.ited, the nephridial ])it, 
which has meanwhile become deeper than l)efore, begins at its 
inner blind end t(j divide into two latei'al branches. Each of the 
latter corresponds, as will be fully demonstrated further on, to tlie 
nephridial canal of Actinotroeha. Fig. o.S, a frontal section of a 
larva at this stage, shows the Ijifurcation just alluded to. The 



ON DEVELOPMENT ETC. OF PlIOllONIS. 527 

relation of the uiipuired nepliridial sac to tlie gut will 1)e Ijest 
imderötood from tlie median saiiittal section uiven in li*;'. ol . 

I may licre Itc allowed to put in a short histiH'ical l'evicw of 
the niesol)last-foi'mation in the PJioronu larva. 

KowALEWSKY ('67) attril)uled the origin of tJic mesohlast to 
the eetoljlast. 

INIetschnikoff ('82), Foettixger ('82), and E. Schuetze 
('97) confounded the plasmic cor])Uscles with the true mesoldast, 
and none of them was aware of the presence of the anterior diver- 
ticula. 

Caldwell ('85) made many interesting observations on the 
mesoblast-formation. According to his view, there exists no meso- 
hlast Ijefore the closure of the l)lasto])ore lips (lateral), but it arises 
later from three distinct sources, viz., 1) the anterior paired 
diverticulum (entoblastic), 2) the posterior paired diverticulum 
(ectoblastic) and 3) " the primitive streak " connecting the above 
two structures. Further it has been declared by him that the 
body-cavities of the larva arise in two different regions. As to 
the preoral body-cavity, he writes as follows : " From the time 
when two or three me.^oblast eelh arc budded off from the diverti- 
cula on either side, a cavity is prese?it in each mass thus for/Ned. 
These cavities are the two halves of the hodij-eaviitj (preoral) 
{I.e., p. '374). On the other hand with respect to the ])osterior 
Ijody-cavity, he states that " if is formed independently in a 
paired mass of cells which y row out to the end of the first formed 
sacs and remain separated from septum " [I.e., p. o7(>). Thus he 
regards the preoral b(xly-cavity as ai'ising after the enterocœlic 
type. Lastly the author puts forward in his recapitulati(jn the 
opinion that the blastopore gives rise to both the mouth and the 
anus. 



ü2S I. IKEDA : 

Houle ('90) al^o dinliiiguislied two s(H't.s of mcsoblabt cells in 
view of tlieir different origin and fate : '' Meseneliynies primaires " 
and '* initales niésoderniiqnes." Both are derived from the " pro- 
toendoderme" whieh foi'ins the primaiy arehenterie wall. The 
lattcu' gives rise to cells grou})ed together into two compact 
" bandlettes mésodermiques," Avhich are regarded as homologous with 
the mesodermal l)ands of Annelid larvœ. In reality these hands 
are, as have been pointed out by Sciiultze, nothing else tlian the 
posterior ])aired diverticula of Caldwell. 

From the account of the mesoblast-formation given in the 
foregoing lines, it is evident that the first stages of that process 
are observable from the vei'v Ijeginning of gastrulation (ligs. 2(] a 
& b), and long before the l)lastopore takes the small triangular 
sha})e. ( )n this j)oint my observations stand at variance with 
Caldwell's. Nor can I agree with that author in the o}>inion 
that the mesoblast produced from the auterior diverticula (even 
thouL!:h consisting of onlv two or three cells) incloses an enterocudic 
cavity. As already described, the cells in question, after l)eing 
bud(h'd oil', lie loose in the blastocœle together with jU'eexisting 
mesoblast cells and without forming a wall to a special cavity of 
any sort. 

As to the ventral groove, Metsciinikoff ('82) was tlie fust 
to refer to this structure and wrote as follows : " In pd.^seiuler 
Lage dc6 Kmhryo^ Icann man cine in Verbindung mit detii Bla^to- 
porus beßndliehen FurcJte {/ongiéiidina/e) waln'nehinei), (reiche zum 
Tliniercnde des Embryox hinzieht und sich nur auf dein Ehtodenii' 
beshränlct. Diese Furclie erhöht den bilateralen Bauplan des 
Embryos erscheint indessen als eine vergängliche Bildung, welche 
man auf sjnUeren Stadiunh rergebens suchen vAlrde'''' {I.e., p. oOl). 
According to Caldwell, this groove, which he calls " the primi- 



ON DEVELOP^FENT ETC. OF PHOROXIS. 529 

tivo streak," is pvodueed l)y a fusion of tlio Idastoporo lips ; tlie 
cells along the fusion line differentiate after multi]ilieati()n into the 
epiblast, the hypoblast, and the niesoblast. And the i'a])id gi'owth 
of the epiblast in this region soon obliterates the groove, leaving 
however its posteriormost portion as the " anal ])it." But such, as 
I have tried to show, is not the case, for the so-called primitive 
streak entirely disap])ears leaving no trace whatever, htng l)ef()re 
tlie nephridial or anal ])it makes its appearance. Therefore there 
exists no direct genetic relation between the primitive streak and 
the anal pit. 

. Caldwell's view that the two nephridial pouches give off 
the mesoblast, which eventually lines tlie posterior l)ody-cavity, 
can not ])e sustained ; for, according to my own observations, that 
l)ody-cavit3' with its mesoblastic lining wall is already j'jresent before 
the nephridial pit divides into the two pouches. It is true that 
the cells floating in the posterior l)ody-cavity arc in some sections 
found aggregated at the 1)1 ind ends of the pouches as shown in 
Fio". ^(S. This is a condition which mio-ht mislead one to the 
conclusion that mesoblast cells are here in process of proliferation. 
But solid cell accunndation in such a section is to be considered 
as simply due to the obliteration by compressiou of the lumen of 
the nephridial pouches. Fig. oO taken froiu an obli(|uely cut 
sagittal section thi'ough a larva of this stage, shows no wander- 
ing cells in front of the ]iou('lies (of which only the right one is 
seen in the figure) ; in this case there is certainly no doul)t about 
the matter. 

Finally as to the anus, Caldwell mentions " a solid cord of 
cells" which he considers to be the posterior remnant of the 
primitive streak. According to him, this acquires a lumen and 
forms a fine canal leading fi-om the primary gut cavity to the 



,ü30 I. IKEDA : 

exterior. However, it seems elear to ine tlmt tliis eord is nothing 
else than an early stage of the intestinal ontgrowtli independently 
produced at the posterior end of the gnt. Moreover, in Phoronis 
ijimai, the gut cavity does not eonie into eoinnninication with the 
exterior at so early a developmental stage as Caldwell observed ; 
in that species, the anus first opens at a definite stage when the 
larva l)ears two pairs of larval tentacles, 

E. SciTULTZE ('97) rejects Caldwell's views in regard to 
the anal ])it, ]»ut regards it as a rudiment of the future ventral 
])Ouch of Actinotrocha. This is, however, certainly not true, since 
the ventral ])oucli is a thing that has a distinct origin and appears 
at a much later stage of larval development. 



<J. Further Observations ox the 

])EyELOrMENT OF TUE LaRVA. 

Rome authors have recorded that the larva swims about 
abroad at such a stage of development as is rcj-a'csented in fig. <S. 
However in Phoronu iju/ifii, the larva lies hidden in the lop- 
hophoral loops of the mother until it has acquii'cd at least two 
pairs of larval tentacles. 

In the larva shown in fig. 9, the somewhat ])rominent preoral 
lobe hangs over the larval month. Local ectoblastic thickenings 
occur at two places, ?;/-., at the centre of the upper surface of 
the ])reo]-al lobe and along the mid-ventral line near tlie posterior 
end of the bod3\ The former is the future nerve ganglion ; the 
latter, the rudiment of the first pair of larval tentacles. The 
nephridial invagination at the posterior end is still shallow. 

At a little later stage, the tentacular thickening divides into 



ox DEVELOPMENT ETC. OF PHOP.OXIS. 531 

two iiioro prominent ridsio?^ running on oaeli sitio obliquely ant- 
eriorly. The preoral lobe grows rapidly so as to hang down on the 
ventral side and as a consequence of this an œsophageal canal is 
formed (fig. 37, œs). The œsophageal wall is, therefore, ecto- 
blastic in origin and is composed of strongly ciliated columnar 
cells. About this period the nephridial invagination l)ecomes 
completely divided into two lobes at the proximal end, as T 
have already described (figs. 37 and 38, 7irp. p.). In more ad- 
vanced larvte, the pit is split throughout its entire length into 
two nearly parallel canals, each of which opens independently to 
the exterior. Figs. 39 a-c show tln-ee transverse, though not 
consecutive, sections Classing through the posterior region of a larva 
at such a stage. In the first of these figures, the two cell-masses 
{nep. c.) on either side of the stomach represent the uppermost 
portion of the nephridial canals. In the second figure, each of the 
cell-masses encloses an easily distinguishable lumen. The two 
canals finally open to the exterior each Iw a small pore {nep. o.), 
as seen in the third figure (only one pore is cut through in the 
al)Ove figure, the section l)eing slightly oblique to the main axis 
of the Lirval body). In the above figure we see an ectoblastic 
cell-mass separating the right and the left nephridinl canals (nep. c). 
How is this partition 1)rouglit al)Out ? I think it is caused bv re- 
evaginntion of the distal unpaired portion of the nephridial pit, as 
by that process the pit wall forming the above portion is graduallv 
transferred to the body-surface of the larva. 

Meanwhile the oesophagus becomes more and more elongated, 
while the paired tentacular thickenings bulge out each into two 
perceptil)le prominences. The latter represent the rudimentary state 
of two larval tentacles, each of which has internally a cavity con- 
tinuous with the ])ostoral body-cavity. Fig. 10 represents a larva 



532 I. IK EDA : 

witli two paii',^ of ns yet very sliort Inrval tontneles ; this i;^ the 
most ndvaneed (levelopmental stage to be met with in the em- 
bryonal masses. Fig. 40 is a median sagittal section of such a. 
highly advanced larva. Here the a^sophagns (œs.) and the in- 
testine [hit.], wliicli latter now commnnicates with the exterioi- by 
the small anns {an.), are highly developed, so that the three parts 
of the alimentary tract (œsophagns, stomach, and intestine), may 
be said to l)e almost complete. The JiervO ganglion (fig. 40, (/I.) 
is well differentiated from the ectoblast of the preoral lol)e, pre- 
senting itself in section as a ronnd, well marked mass principally 
composed of nerve fibres. I have l)een nnal)le to ascertain whether 
a proctodœnm is j'troduced at all, and if so, what j^art of the post- 
gnt it gives rise to. 

The preoral ])ody-cavity is, at this stage of development, still 
very incompletely separated from the postoral cavity l)y a few 
mesoldast cells (fig. 43, '//ic.^\). The nephridial canals (fig. 41, 
7iep. c.) are now distinctly separated and removed from each other, 
and are fonnd in a cross section to l)e situated laterally to the 
intestine {inf.). One on the right-hand side of the above figure is 
cut through at its extei'ual opening, while on the other side the 
nephridium is represented by a thick mass of a few ectoldastic 
cells. This lateral shifting of the nephridia becomes more and 
more pronounced with the advancement of larxal developmeiit. A 
slightly advanced state of the nephridia is shown in fig. 42, where 
the nephridial canals {ncp. c.) are now seen tolerably long and have 
a wall composed of a single row of cubical cells. It is often observed 
that some mesol)last cells connect the canals with the splanchnic 
walls (see the above figure). These cells seem to be the first in- 
dication of the future collai'-trunk septum. Besides, a certain 
number of mesenchvmatous cells, which later uudoubtedlv become 



ON DE\'ELOrME:S'ï ETC. OF i'HORONIS. 533 

the excretory cells of Actinotroclia, is always found attached to 
the 1)liiid ends of the nephridial canals. Caldwell say>s that he 
saw the excretory cells aggregated around the apex of each canal 
and that they had numerous plasmic processes, giving them a 
strong resemblance to the perforated cells known in J^chiiirua. It 
seems, however, highly prol)al)le to me that this strange appear- 
ance of the excretory cells is an artefact, since, as I shall point 
out later, the same cells in Actinotroclia are certainly not provided 
with any such processes. 

I have very frequently detected some gigantic meso1)last cells 
Iloating freely in the ]:>ostoral 1)ody-cavity of larvie with (jne or 
two pairs of tentacles (tig. 44, corp.). They are round and 
nucleated and contain numerous large yolk-spheres. After repeated 
examination I have come to regard them as mother-cells of blood 
corpuscles which are found as corpuscle-inassess in the collar cavity 
of Actinotroclia. This point will again Ije treated of in detail in 
the proper [)lace in the following section. 



II. The Structure of Actin otrocha. 

a. External Appearance. 

It can scarcely be dou])ted that each species of the PJloi'o- 
n'idœ has a characteristic lorm of Actinotroclia peculiar to it. Bonie 
of the previous observers {c. g., Wilson and Masterman) have 
mentioned two distinct types of larvie as occurring in the same 
locality. Among the larva^ which I ol »served at Misaki, I was 
able to distinguish four ditlerent types, each of which had a 
characteristic form and a more or less delinitc topographical 



534 I. IK EDA : 

di.stril)utioii. I will designate these types Ijy the letters .1, i>', C, 
and JJ. 

Type A (lig. 13). The larvje of this type were principally 
coUeeted in Aburatsubo and belong in all pr<)bability to the species 
Fhoronia ijiitml, wliich, as I have said, is fonnd in the same 
locality. The body is comparatively short and thick, measnring 
abont 1.-1.'3 mm. in total length. The larval tentacles of a fidl 
üTown larva never exceed 1(3 in nnndjer. 

Type B (fig. 14). This is a larger form than the preceding 
(al)ont 2-2.5 mvi. in length). The body and the intestinal canal 
are long and slender. The fnll grown larva has abont 28 tenta- 
cles which are mnch more slender than those of Type A. Pecnliar 
to it is the sensory spot (.so.) sitnated jnst in front of the ganglion 
{(/L). The larvae were fonnd in greatest abnndance near Kitsune- 
saki, a point at the month of the inlet Moroiso. 

Type C. (figs. 1-3 a & l>). This type is distingnished from 
all the others by several characteristic points. In size of body 
it is intermediate between Types A and B (nsnally l.bmm. in 
length). The body is relatively short and thick. The nnmber 
of tentacles, so far as I know, ranges from 16 to 24. A pair of 
flask-shaped glands [gld.) is fonnd one on either side of the 
ganglion [gl.]. A pair of retractor mnscles {i'ct\) runs longi- 
tudinally through the trunk cavity from the tentacular ring to the 
apex of the anal cone. Compared with the first two types this is 
much rarer. 

Type D (figs. 12 and IC). This is a rare form of which I 
have obtained only seven specimens in all. It is enormously large 
in comparison with the others (4.-5. mm. in total length Jind 
1. mm. in width). The ]>reoral lobe is disproportionately small, 
while the trunk is Ion-' waA thick. The tentacles are remarkably 



ON DEVELOrMENT ETC. OF PlIOROXIS. OOÖ 

iiLiiiKTüUS, sometimes reaching 48 in niiiiiber. In ;■, single living 
specimen, the skin of the trnnk was of a light orange colour ; the 
sul »dermal circular muscles were especially well developed in the 
trunk but intei'ruj)ted at four longitudinal clearly marked zones. 

The youngest swimming larva I have ever obtained was of 
type A. It was already supplied with four pairs of tentacles 
which, however, wTre still short. The l)ody measured about 0.5 
//i/ii. in lençrth. The trunk was short and showed a sliirht 
characteristic curvature, the concavity being turned toward the 
dorsal side.^ The thickly ciliated hood was comparatively large; 
the ganglion and the perianal ciliated belt were already well deve- 
loped. In the surface view of this larva during life, I was not 
aljle to detect the ventral pouch nor the c(jrpuscle-masses. 

At about the stage vvitli five pairs of tentacles, the trunk 
becomes elongated and straightend out. The nephridia may then 
be seen in their characteristic bouquet-foim, and the ventral 
pouch appears as a solid ectoblastic thickening. Neither the 
corpuscle-masses nor the retractor muscles are yet to he seen. 

As the larva grows, the number of tentacles increases in pairs 
proceeding from the ventral side toward the dorsal ; hence, the 
most dorsally situated tentacles are the youngest and the shortest. 
In larVcC with 12 tentacles and Ijclonging to type A, the ventral 
pouch is deep enough to l)e plainly visible from outside. We 
always notice from this stage on a pair of the retractor nmscles 
which extends Ijctween the ganglion {(//.) and the dorsal inner side 
of the tentacular circle {ret. in figs. 12, lo, 14, and 1Ö). 

The larval organisation of types A and J> is nearly com- 
pleted in the stages with 14-1(> tentacles. J^et me next give a 
somewhat detailed description of the external aj)pearance of Actino- 
troeha in ueneral. 



536 I. IKEDA : 

TJie prcoral lube. This is u .structure wliicli looks like ;i 
broad liood with its concavity directed downwards. It almost 
entirely covers the upper anterior part of the collar/" when not 
influenced ])y external circumstances. Masïer.aeax has made the 
statement that in its natural attitude the hood has its leuii'th 
disposed parallel to the ]>rincipal 1)ody-axis. However, if the 
larva be examined in the living state, it will at once l>e discovered 
that its normal disposition is horizontal. It l)ecomes turned up 
only as the result of preservation. Its whole surface is covered 
with cilia, most strongly developed along the free margin which 
constitutes the preoral ciliated belt. In the full grown larva, the 
ganglion (and also the sensory spot in ty[)e B) has also a set of 
sjK'cially long cilia on the outside. Numerous line and refractive 
]ierve fibres are seen radiating from the ganglion {gl.) to the free 
margin of the lobe {pre. hel.) (figs, 13 and 14). 

Masterman has described and figured two ectoblastie struct- 
ures which are said to be situated on the ventral wall of the hood 
and which he has named the " oral " and the " pharyngeal " 
grooves. These he compares, as to their function, to the gill- 
slits of the Chordata. I can not but think that that writer has 
liere fallen into a very grave error, which might have been 
avoided, had he examined the structures in question in living 
specimens. Among the preserved s})ecimens I have fre(piently 
noticed those in which the lower or oral wall of the hood Avas 
])roniinently bulged out in front of the mouth. \i\ consequence 
of that prominence (iig. IC), j>ro//(.), thei'c was produced on either 



■•'■1 ;i(Uipt tills iKUiie ul' Mastcnnuirs tn ilfimk' lliat portion dftlic I.-irval Ijotly whk'li lichi 
in iVoiit ui'tlie tt'iUacular circle and bohind the preoi-al Idhe. 



ON DEVELOPMENT ETC. OF PnORONIS. 537 



side of tlio mouth n transverse gvoovo, wliicli was visible wlicn 
viewed from the ventral side. I l)elieve it was to grooves of this 
kind that Masterman assigned the ahove im^iortant significance. 
Tn mv opinion, they are simply artificial j^rodnctions due to preser- 
vation. 

Tlir Collar. The form of the collar as a whole may be 
compared to a cylinder obliquely truncated at the posterior end. 
Its posterior border is fringed with a regular row of tentacles, 
while antei'iorly it is joined to the hood by a narrow neck. The 
number of tentacles (larval) varies according to the different stages 
of growth and also aecording to the type to which larva) l)elong. 
They are most numerous in type D, most individuals of wliich 
bear 40-48 tentacles (figs. 12 and 1()). The rudiments of the 
adult tentacles make their appearance as bud-like ectoblastic 
ijiickenings immediately below the base of the larval tentacdes. 
An exception to this rule is found in the case of larva) belonging 
to type I), in which the adult tentacles are represented by a loeal 
ventral thickening of the wall of the larval tentacles at their 
jn'oximal portion (see fig. 58 c/, s. t.). It is very probable that 
the numbc]' of the larval tentacles corresponds to that of 
the adult. In type A, at any rate, I have ascertained that the 
full grown larva and the worm just metamorjdiosed bear the same 
number of tentacles, namely 10. 

The trunk. This portion, whic4i is the shortest of the three 
regions in early larval stages (fig. 10), comes with growth to 
occupy the largest part of the larval body and assumes a long 
cylindrical form. Its anterior boundary is the tentacular cii-cle ; 
the posterior end is girdled with the ])erianal ciliated belt which 
serves as the larval locomotorv or^'an. 



538 I. TKEDA 



h. The Internal Structure of Actinotrocha. 

1. Body-Divhiom nml Bod y -Cavities. 

I have oiKloavoTod to show in tlio proeodinii; pn2;es, that the 
l)0(lv-('avitio!^ of Actinotroclia do not arise from the enteric diver- 
ticula, as was insisted upon l)y Caldwell, hut that they are 
simply produced by mesohlastic cells applving' themselves to and 
forming the lining of the ectohlastic, and the entoblastie, wall. 
They may, therefore, he classed under the '* j^^^^^'^^o^^^l^ " or 
" schizocœle " of Hertwig. ]\Ioreover, tlie body-cavities of Acti- 
notrocha as a whole do not in their genetic relation correspond to 
those of the adult, as I shall attempt to elucidate in the sequel. 
Dui'ing the metamorphosis, the greater part of the former (the 
preoral cavity) is almost wholly lost, while the other part (the 
collar-cavity) is transformed into a vascular space, so that what is 
known by the same name in the adult is of an entirely new origin. 
Thus we see that the larval l)ody-cavity of Actinotrocha, ?'. e. the 
ti'unk cavity, is the only jx'i'tion that persists among the body- 
cavities of the adult, in wliich it is known as the foot or infraseptal 
cavity. In correlation with this cii'cumstance are observable cer- 
tain changes in the position of the ne])hri(lia and of the vascular 
system. As described by Caldwell, the nephridia of Actinotrocha, 
which are not pi'ovided with an internal opening, lie for the most 
part in the collar cavity, while after the metamorphosis they are 
found wholly in the infraseptal cavity of the worm. Moreover, 
the paired corpuscle masses which are f )und only in the collar 
cavity of the larva, are no longer seen in the same cavity of the 
adult. These chanws to a cei-tain extent at least establish the 



ON DEVELOPMENT ETC. OF PHORONIS. ôoO 

fact th;it some profound elianges in tlic iir»ani;cinent of the hody- 
ravities must occui' during the metamorpliosis. As is ac- 
knowledged l)y all, the suprasepttd cavity of Phoronis is greatly 
reduced in size as compared with that of the larva, and contains 
almost no organ except the blood vessels. Tlie infraseptal cavity 
is, on the contrary, very wide, and contains many important 
organs, e. //., the alimentary canal, the sexual organs, and the 
main part of the vascular system. Thus it becomes necessary to 
make distinctions between the body-cavities of Adinotrocha and 
those of the adult and to call them respectively by different names. 
The former may be termed the larval ]>ody-cavities, and the latter, 
the adult body-cavities. 

Most pi'cvious writers have not taken any particular notice of 
the relation which exists between the external body- division s and 
the body-cavities of Actinotroeha, so the words " hood " and " foot " 
do not denote anvthino- but mere external features. The idea of 
segment was first introduced by Caldwell ; he considers the larval 
body as diN-ided into three parts : (1) the preoral lobe set in front 
of the septum, (2) the ti-unk portion situated behind the septum, 
and (3) the foot or invaginated pouch. According to this view, 
the body-cavity is divided by the septum into two contiguous parts, 
vh., the preoral cavity in front of, and the trunk cavity behind, 
the septum. Masterman divides the entii'c liody into three 
portions, viz., the preoral lobe, the collar, and the trunk. These 
three divisions are not only externally marked by their respective 
forms, but also l)y the presence of two transverse septa or mesen- 
teries. Thus we see, the preoral lobe of Caldwell comprises l»oth 
the preoral lobe and the collar of ^NTastermax. 

Whatever may be the value of j\[asterman's Diplochorda 
hvpothesis, T fool inclined to acco])t with some modifications, his 



040 T. IKE DA : 

view of the Ixjdy-divisions. The external appearance of the three 
]iortions I have already descril^ed in brief. As to the internal body- 
cavities corresponding to tliese external portions, I can not agree 
with Masterman, when he says that they are completely separated 
from one another ; for, as I shall soon show, the septum which lies 
between the preoral and the collar cavities is always an incom- 
plete formation, at least in all the Adinotrochœ which I have 
observed. Besides, T have been unable to detect the first and third 
pairs of nephridia, which are said to exist in the preoral, and the 
trunk cavities (Master^afan). Therefore, I can not regard the body- 
divisions 0Î Actinotrocha as " segments " in the sense ofthat author. 

The septa or mesenteries are very delicate in structure and 
can hardly be recognized in living specimens. I have, therefore, 
had to study them mostly in sections. I shall hereafter call the 
two septa the preoral, and the postoral, septa. 

Larval Preoral Body-Cavity. The larval prcoi'al l)ody-cavity 
fills u]) the interior of the hood, in which there is no entoblastic 
organ. Innumeral)le mesench\miatous filtres traverse the cavity 
(figs. 45, 49, 63 a, m./.). A few blood corpuscles are also 
frequently discovered in this cavity (fig. 49, corp.) ; this fact is, 
I believe, one of the proofs of the correctness of the view which I 
now propose to consider. 

I have already spoken of the incomplete formation of the preoral 
septum. Whether this is a mere sjiecific difference or not, remains 
to me uncertain, as I have had no chance of examining the larvae 
investigated by Masterman. In sagittal sections of the larva at 
any stage of its growth, the septum can constantly be traced so 
long as the œsophagus is contained in the sections. In figs. 4ô 
and 63 a, a slender cellular strand {mes'.) behind the ganglion {gl.) 
represents the septum in cross section. It extends between the 



OiSr DEVELOPMENT ETC. OF PHORONIS. 541 

upper and tlie lower walls of the liood. Thus it will be seen that 
the septum completely separates the preoral cavity from the collar 
cavity just behind the ganglion. But, when we come to secticms 
passing through a more lateral regi(jn to either side of the gan- 
glion or of the œsophagus, the upper portion of the septum Ijecomes 
abruptly indistinct. In fig. 54, which shows a sagittal secti<)n 
through the right-hand side of the oesophagus of a larva of 16 
tentacles (type A), the septum {tiles'.) near its ventral attachment 
is indicated by a comparatively thick layer of cells, while the 
dorsal portion is divided into fine protoplasmic branches, of which 
some extend to the upper wall of the hood and others stop short 
of it. As the relations of this septum are somewhat complicated, 
I will try to make them clear by referring to a series of cross 
sections (not continuous) through the hood of a larva of type D 
(figs. 59 a-d) and also to the annexed wood-cut. The latter is a 
diagrammatic representation of the Aclinotrocha hood and its 
neighbouring part, as seen from above, i. c, in horizontal projection. 
The dorsal side is above and the ventral, below\ Nearly in 
the centre is the nerve ganglion. Below it and concealed from 
sight, is the mouth, from which the œsophagus leads downwards. 
The little stellate markings, scattered over the greater part of the 
figure, are supposed to represent meseuchymatous cells, wdiich, with 
the branched and reticulate fibres arising from them, pervade 
the preoral l)ody-cavity, except in a small space immediately in 
front of, and l)elow, the ganglion. This free space I shall call the 
230sterior recess of the preoral cavity. The line a b c d e f g h, 
curved somewhat like the letter M, indicates the position of the 
preoral septum. The part shown in full line represents that por- 
tion of the septum which is complete in structure and the part in 
broken line, that portion of the same which is incomplete. All 



i4^ 



1. ikEDA i 



tlic .sjnicr ill front of tlio f^cptuin is llic prconil cavity, while back 
of it lies the colhir cavitv. 



Denser accumulation of filires 
in front of tlio r(_'«-u;-.s. 



1 cjivitv 



I'rcorul belt. 




recess of 
:il cavity. 

gliou. 



IV _ 

I'reora,! 
jueseutery. 
Ocsopluij^us 

( 'ollar cavity 



Tentacles. 

If we now study tlie scrii's of sections (fi.i;s. -39 a-d'^), the 
nature of this septum will l)eeome clear. Section öd a is the most 
anterior of the four and ])asses through the hood at al)out the 
plane of the line I-I in tke ahove wood-cut. The Avhole of the 
p]-eoral cavity is Idled Avith the filjres of branching jneseuchymat- 
ous eells except in the ventral median part {p.r.). This is the 
beginnini;- of the posterior recess of the ))reoraI cavity, which is 
limited anteriorly by a faint menil)raiieou« layer consisting of pro- 
to})lasniic li])res only. Section od b is from about the plane of 
the line II-II. This contains the ganglion (y^.) and ou each side 



* Unfortunately in the scries of sections tissues have undergone considerable disturbance 
by the action of the killing reagent, but the relations of the layers remain unaltered. 



ON DEVELOPMENT ETC. OF PHOROXLS. .j4o 

the anterior end of the eolhir cavity {coLrX Below the gaiijjjlioii, 
the posterior reeess (/>./'.) is .seen to ]n\\v a (■oin])lete wall, that is 
to say, the posterior septum is tally developed. In the ahove 
fleure, the two wide s])aees lying on Ixjth sides of the posterior 
reeess correspond to the two anterior horns of the collar cavity 
projecting forward (marked b and (/ in the ■wood-cut). And "we 
can certainly see that on the dorsal side as Avell as laterally there 
is no distinct partition or continuation of the preoral septum 
which, according to Mastermax, should entirely divide the preoral 
and the collar cavities at every point. As the figure shows, the 
dorsal portion of the mesentery {mes') is decomposed into line 
protoplasmic processes which join with those of the fihrous mesen- 
chymes dis^iersed through the preoral cavity. In section, öd c, 
passing though the middle of the ganglion (the line III-III 
of the wood-cut), the collar cavities {col.c.) are much wider and 
have become united below the œsopliagus {œs). The septum (as 
the wall of the posterior recess, ]).)•.) in this region is a little more 
definite in form than in the last figure ; the posterior recess (p.r.) 
is distinct as before. In section 59 d passing through the line 
IV-IV of the wood-cut, the posterior wall of the recess (79. y.) is 
obliquely cut and appears in the right-hand lower corner as a 
membraneous slice, the recess being distinctly boundetl by the 
septum {mes\). Outside of it are seen, one on each side, the 
sections of the retractor nuiscles (ret.), of which more will be 
said later. The collar cavity (eol.c.) is now very spacious, but the 
septum laterally remains in the same condition as before. 

From the above descriptions, it will be clear that the preoral 
septum is complete only in the median ])ortion (indicated by the 
full line e d ef in the wood-cut), while in the more lateral part 
on each side, it is at the best a loose open reticular membrane, 



544 t. iKEDA : 

tliroiigli ^vllit'll the t'œlomic fluid of the preoral and the collar 
cavities is j^ut in free circulation. 

A questionable structure has l)een described from the preoral 
cavity 1)y Mastekman under the name " subneural sinus," and is 
compared to the structure l^earing the same name in the Hemi- 
chorda. According to him, the sul)neural sinus is an interstitial 
space left between the two laminae composing the preoral septum, 
just under the ganglion and alcove the so-called " subneural gland." 
Anteriorly and laterally, it is said to be surrounded liy the preoral 
cavity, and posteriorly, by the collar-cavity ; its upper and lower 
walls are claimed to be directly formed of the ectoblast without 
a peritoneal layer. Further it is said, that the sinus connnunicates 
mid-dorsally with the dorsal blood vessel on the oesophagus. 
After repeated examinations of the larvœ of the four different types, 
I am convinced that Masterman's subneural sinus is identical 
with what I have called the posterior recess of the preoral cavity. 
It has nothing to do with the tissue-space in the jn-eoral septum, 
])ut is clearly a part (jf the preoral Ijody-cavity, which is free 
from the mesenchymatous fil)res. Besides, I can not in any way 
detect the presence of the dorsal vessel on the œsophagus, a vessel 
which connects the suljueural sinus with the dorsal vessel on the 
stomach. A view similar to mine as above expressed was given 
Ijy HarmePv in his paper on Cephalodisus ('97)- 

Masteeman has further given an interesting description of the 
" proboscis pores," situated on each side of the ganglion. They 
are compared to the proboscis pores of Balanoglossus and are 
said to fulfill the same function as the collar nephridium of 
Actinolrocha. In the larvîc studied l)y me, the only things that 
bear even a remote resemblance t(j them, are the flask-shaped glands 
which are seen on the upper face of the preoral lobe of the larva 



ON DEVELOPMENT ETC. OF PHOIIOXIS. 545 

belonging; to type C. But the position of tliese glands in rdatidii 
to the ganglion as well as their histological structure at once 
reveal their true nature. The internal openings of the organs 
were described bv Masterman as follows : " Just ivherc the 
preoral mesoblastie ivall slopes away on rlther side of the slnnx 
there are a pair of thickenings, which traced forwards, show 
themselves to be the commencement of a pair of internal openings^' 
{I.e., p. 307). The paired thickenings referred to by him are 
apparently nothing else than the points of attachment of the 
retractor muscles in the collar cavity, as will be seen in fig. 59 d 
{ret.). Further details respecting these muscles will be given 
later. 

Larval Collar Cavity. The collar-cavity is a comparatively 
wide space extending between the preoral and the postoral septa. 
It is produced anteriorly into two horns, embracing between them 
the posterior recess of the preoral cavity. It is perfectly separated 
by the postoral septum from the trunk cavity. The postoral sep- 
tum, or simply the septum, as it is more commonly called, is 
stretched obliquely transversely between the splanchnic and the 
somatic walls, along a line a little l)elow the tentacular circle (figs. 
45, 48, mes^. Its dorsal attachment on the splanchnic layer is, as 
represented in fig. 45 (mes.), found at the plane of the junction 
of the œsophagus with the stomach, while ventrally the attach- 
ment lies much further below. In frontal sections of the larva, 
the septum (fig. 48 mes^ is seen on either side of the stomach 
and its somatic insertion lies just under the tentacles, so that each 
tentacular cavity is continuous with the larval collai- cavity (fig. 45). 

The adult collar cavity, or the supraseptal ca\'ity, is already 
formed in the fully developed larva of every type, as a ring-space 
runnino" alons; the inner side of the tentacular circle and above 



Ö40 I. IKEDA : 

the septum (see figs. öS n and d, s.r.c). Tliis, togetlier with 
.several other larval organs in the larval collar cavity, liad hetter 
be treated at a more suitable place in the sequel. 

Masterman has described a d(jrsal mesentery running along 
the mid-dorsal line of the œsophagus, and separating dorsally the 
larval collar cavity into two lateral halves. In the Actinotrochae 
of all the types observed by me and at every stage of the larval 
growth, no such mesentery is present. It is true that the liody 
walls and the œsojihageal walls very frequently come close together, 
especially in the young larva after preservation, so as to greatly 
narrow the collar cavity in this region (figs. 49 and 50 a). But 
a mesentery is never to l)e found. Its absence is quite clear in 
the large Actinotrocha belonging to type D, in which the skin 
and the œsophagus lie w^ell separated l)y a consideral)le space 
(figs. öSrt and 58/>). 

Trunh cavity. The trunk cavity occupies the interior of the 
third body-division — the trunk. It is completely separated 1)y the 
joostoral septum from the collar cavity, and since the septum is oblique 
in position, it extends dorsally nearly to the l^ase of the œsopha- 
gus. The ventral mesentery extends along the median ventral line 
of the body wall and of the alimentary canal, and is wholly con- 
fined to the trunk cavity. In fig. 4'"5, which shows a median 
sagittal section of a young larva of type A, a portion of this 
mesentery (ernes.) is represented as a thin cellular membrane 
extending between the alimentary canal and the ventral pouch 
ip.o.), the latter being still shallow at this stage. The whole 
extent of the ventral pouch is stretched by the ventral mesentery 
to the skin as well as to the digestive canal. This relation re- 
mains the same as the pouch grows in length and finally winds 
around the digestive canal. A tr«nnsverse section through the 



ON DEVELOPMENT ETC. OF PHORONIS. 547 

trunk of a Inu'hlv advanced larva of type C', is ^iven in fig. ~>7 «r, 
in wliicli tlie nnieh elongated and eonvoliited poneh is seen cut 
into several sections (po.), connected witli one another l)y tlie 
mesentery (I'./nes.). 

Very freqnently it happens that the ])eritoneal niesoblastic 
ephhelinm, which lines the perianal ciliated helt, is detached from 
the ectohlastic wall. 'J'liis is a pnrely artificial appearance cansed 
hv the killing reagent. It seems probable that Mastekman has 
erroneonsly considered the space thus formed by splitting to be a 
vascular space (the " perianal sinus "). The same author states, 
though with nnich reserve, that he has discovered a third pair of 
nephridia in this trunk cavity, which is considered to be a modi- 
fied part of the b(xly-cavity, and also to l)e rudiments of the adult 
nephridia. I can at jn-esent say no more than that these are cer- 
tainly absent in every type of the Actinotrocha studied by myself. 

2. Orgam of Ectohlaxtlc Orir/in. 

The epidermis of Adinoiroclm is represented by a single 
laver of cubical or cylindrical cells, those of the collar wall and 
of the upper and the lower walls of the hood l)eing ja'ovided with 
well developed cilia. Besides, there are three specially ciliated 
regions : the preoral belt, the tentacles, and the perianal helt. The 
last is the lai'val locomotory organ ; on it the cilia are very long, 
thick, and somewhat bristle-like when in active motion. At places, 
where cilia are strongly developed, [e.g., the nerve ganglion, the 
sensory spot if jiresent, the ciliated ])elts, etc.) the constituent cells 
are cvUndrical, the nucleus generally lying near the basal end. 
The body wall of the trunk region is very thin and is formed of 
greatly attenuated cells (es]X'cially slender in the advanced larva*). 



048 T. TKEDA : 

Numerous uuifclliiLir glands are found in the Actinotrocha not 
only all over the two surfaces of the preoral lobe, l3ut also in the 
oesophageal wall as well as in the inner ectoblastic wall of the 
ventral pouch. They are also, though less ahundantly, distributed 
over both the collar wall and the tentacular wall. The glandular 
cells are all pear-shaped, the nucleus being found always oppressed 
to the base of the cell (figs. 49 and fiicJ, in.gl.). In their 
staining reactions, the secretory contents of the glands agree with 
those of mucin. Tt has been often noticed that living larvdo 
remain adhereing to the objects they have touched with the hood, 
and that metamorphosed larvfc behave similarly with the tip of 
the e vagina ted pouch. 

There exists still another, paired, multicellular gland ^vhich 
is observed only in the larva^ of type C (figs. 1"), gld. and fig. lr)c). 
Tt is situated on 1)oth sides of the median line on the upper sur- 
face, and somewhat near, the neck of the preoral lobe. Tt has the 
shape of a round flask with a short neck (fig. 1") e). The appeai'- 
ance of the section through the body of this gland reminds us of 
the chorda dorsalis in Vertebrate embryos : it presents to view a 
mesh-work of protoplasm, a small numl)er of nuclei behig found 
here and there closely pressed against the rc^ticular beams or the 
nodes of these (figs. 50 a-c)."' Each of the meshes corresponds to 
one gland cell. In fig. o6 h, which shows an oblique median 
section of the body of the gland, a comparatively wide round 
space exists in the centre, surrounded by the gland cells which 
are arranged more or less radially. This space, when traced \\\)- 
wards, passes into a short and very narrow tubular canal, finally 
to lead to the exterioi- by a small aperture (Fig. f50 c). Since 



* Bv nn >inf()rli)nnte ovcrsit^'lil, Fi<4'. 50/^ Ikis liail its miinlu'i- oniittod in the plate, 



ox UEVELOi'ilENT ETC. OF I'UOIIONIS. 541) 

the iiet'k portion of the gluiid is very sli(jrt, it is difficult to 
prepare u good longitudinal section of it, in which the canal may 
be seen opening to the exterior. Fig. 50 c represents the terminal 
part of the emptying canal, which, as can Ije ascertained hy regu- 
lating the focus, leads to tlie external pore. Mr. Fizuka tells 
nie that similar glands of ectohlastic origui arc constantly found 
on the superior ramus of a parapodium in certain FoIycJiœla. 

Ventral Pouch. As the ventral pouch is one of the juost 
characteristic structures of Aclinolrocha, its form and fate have 
been fully studied Ijy many previous oljservcrs. In the 8-armed 
larva of type A, an ectodermal thickening l)elow the tentacular 
row represents the origin of the pouch. At the 10-armed stage 
the thickening becomes more conspicuous, l)ut no invagination has 
as yet taken place. For the first time in the 12-arnied stage, the 
wall at the thickenins; begins to sink inwards and backwards 
(fig. 4ü, ^90.). The invagination is lined with a mesoblastic layer, 
and, as before noted, is for its whole length suspended by the 
ventral mesentery, joining it to the somatic and the splanchnic 
walls. As the growth of the larva advances, the pouch becomes 
more elongated and bends on itself around the alimentary canal 
)ligs. 18 and '")7 a, po.). In fully developed larvie of whatever t\']^)e, 
the inner or ectoblastic wall is thrown into small wavy folds 
(beginning at the distal portion near the pouch pore), while the 
mesoblastic layer becomes muscular, so that at the end of larval 
life, it forms a thick muscular sheath whose constituent cells stand 
vertically t(j the inner wall. As to the form and positi(jn of the 
pouch pore, I can öfter no details in addition to what has been 
observed by Metschnikoef and many other authoi'ities. 

Kervous System. The iiervous system of Actinotroclta, like 
that of Fhoroiik, is of a very low devcloj)ment, being rc])resented 



OOO I. IKEDA : 

merely by u local dillerciitiatioii of tlic cctoLlastic cells into iiei'- 
VOU8 elements. The epidermis over l)otli the ganglion (fig. 14. gl.) 
and the sensory spot (.so.) is strongly ciliated, so that the organs 
ai'e easily recognizahlc in the living larva. The earliest stage 
in Avhich I tbnnd the ganglion was a 4-armed larva of type ^1 
(fig. 40 (/l.). In it, the ganglion consisted of only a few ganglion 
cells and nerve fibres. 

Although tlie ganglion and some nerves directly ])roceeding 
from it can l)e detected with tolerable distinctness in the living- 
specimen on account of their peculiar refractivity, the peripheral 
nerves are as a o-eneral rule so verv line and delicate, that thev 
can not be satisfactorily made out by means of any ordinary 
process. With fair success I have had recourse to vital staining 
with methyl-blue. Larva' of type i> have been principally em- 
ployed for this purpose. They are left for about l'")-20 minutes 
in a weak solution of methyl-ldue in sea water and immediately 
afterwards treatetl with ammonium molybdate. Sometimes, I have 
made supplementary ol)servations on larvœ lyiug alive in the 
methyl-blue solution under the cover glass, but this can l»e con- 
tinued for only a slujrt time, since a general o\'erstaining of other 
tissues soon takes ])lace. 

Fig. GO (V^' shows the dorsal view of the anterior half of a 
larva of ty])e JJ, which was treated in the abcjve way. The nerves 
are shown in blue. The results obtained as to their distribution 
dilfer in many important points from those (»btained by MasïEK- 
MAN. AVhcther this dilferen<'e is due to the technique or is 
actually existant in the species studied, is diilicult to ascertain. 

* As the larva sliown in tliis figure was compressed by the eover-glass, the rim of the liooil 
wliieli appears like its free margin is in reality tlie line along Avliieli tlie hood was bent and 
reOected over l)y pressure. Tlie line drawn close to the peripheral dots in blue rej)resents the 
true edge of the Jiooil. 



ON DEVELOPMENT ETC. OF THOtlONlS. ÔÔi 

jMasteeman says iiotliiug of the method (■iii[)l()y('d in liis inves- 
iiiratioii, ami uiitortuiiati'lv there exists no other study lliaii his 
with wliich lo eoiiii)ai"e mv resiüts. 

As mav l»e gathered from the ai)ove-meiitioiicd figure, I eau 
discover no collar nerve ring, nor dorsal or ventral eounnissui'e. 
Besides, in spite of repeated efforts, I have always failed to make 
out the presence oi the so-called perianal nerve ring. The collar 
ring and the dorsal commisure, if they can Ije s(j named, are repre- 
sented by a small number of parallel fibres, which spring directly 
out from the posterior corner of the nerve ganglion. In evei'y 
case examined, they could be traced no further than a short dis- 
tance from the ganglion. Sometimes, I ha\e Ijeen al)le to discern 
in sections the main nerves (commonly 3 in number), which run 
close together and parallel to one another along the mid-dorsal 
line of the trunk, 1)ut they were confined to only a few sections 
posterior to the first pair of tentacles. On the other hand, a 
vei'v complex and l:)eautiful system of nerve iiljres could be seen 
on the preoral lobe. The fibres are here exceedingly numerous 
and fine, radiating from the ganglion on all sides towards the 
free margin of the preoral lol)e. In the median line and 
anteriorly to the ganglion (r/l.), th(! fibres appear as three longi- 
tudinal })arallel strands on Avhich the unpaired sensory spot (so.) 
is situated not far from the ganglion. After passing through the 
sensory spot the strands fray out into fnie fibres which continue 
their c(jurse towards the free margin of the preoral lobe. The 
fibres emanating iVom the ganglion do not all show a regular 
radial arrangement, but there are some that aiisiug from the 
lateral edge of the ganglion, soon take an anteriorly directed course. 
Sometimes there were not wanting, especially near the ganglion, 
indications of anastomosis between the fibres. H(jwever, it seemed 



552 1. IKEDA : 

to nie more i)ro1)a])le' that these appearaiiecs were caused simply 
by the juxtaposition of intersecting fibres. 

The nerve endings in tlie preoral ciliated Ijclt deserve special 
notice. In fig. (30«, there is shown a row of small dots alonir 
the margin of the band. A portion of the latter more highly 
magnified is shown in fig. (10 A. Here each fibre ends in a small 
knob wliich is devoid of any lateral process. At first sight luider 
l(jw magnification, the row of knoljs appears like a deeply stained 
ring. Huspecting that there might exist lateral processes connect- 
ing knobs, I have repeatedly made observations and experiments, 
but without having ever been able to demonstrate such a connec- 
tion between them. 

I can not l)ut thiidc it very strange that post-ganglional nerve 
fibres, if such really exist in the forms of the collar ring and of 
the dorsal and the ventral commissures, should not be revealed l)y 
the method ado])ted. The negative result may be considered due 
to incomplete develo})ment of nervous elements in the collar and 
in the trunk region ; but other anatomical relations prove to a 
certainty that the larvie investigated were fully grown. ^Vs I am 
not (juite sui'e that my method was not in some respect im])erfect, 
1 leave the matter undecided for the present. 

According to ]\Iasterman, there is an ectodermal depression 
directed inwards and backwards, just in front of, and under, the 
ganglion. He calls it the " neurojwre," comparing it to tlie 
neuropore of Ami^liio.uis and even to the medullary canal of 
Vertebrates. I nmst say 1 was much disappointed in failing to 
detect in the ActinotrocJuc studied by me this structure of so nmcli 
theoretical interest. As a matter of fact, it happened very fre- 
quently, while ol »serving livhig larvjc, that tlie ganglion was 
retracted dee})ly inwards by an active contraction of the two 



ON PEVELOT.MKXT ETC. OF TJIOROXTS. •>'>■> 

retractor musflcs in the collar cavity (fig>:. lo, 14, !•"), ret.) pro- 
diicing at the same time a deep depression just in front of the 
ganglion. It is also of almost constant occurrence that the gan- 
glion is withdrawn inwards on the application of reagents, so as 
to produce a shallow pit or groove in front of, or helow, the 
ganglion (figs. Go r^ f/L). A cpiite similar fact is always ohserved 
in Lovén's larva. From these circumstances T am much inclined 
to regard the " neui'opore " of Mastermax not as a I'cally exist- 
ing structure, Init as an artefact. 

As to the tentacles, I have at present nothing to add to what 
is already known about them. 

o. Organs of Entohladh- OrUjhi. 

In the fully grown larv;e the alimentary canal is a long and 
straiirht tuhe ; it begins with the mouth which is overliung bv the 
preoral lobe, and ends at the anus in the centre of the anal cone 
sin-rounded by the perianal belt (figs. 12-10). Of the whole ali- 
mentary tract three parts may Ije distinguished : the Oesophagus, 
the Stomach, and the Intestine. 

Oesophagus. In the eml)ryological part of this article I have 
said that the oesophagus of Actinotrocha is of ectol)lastic origin, 
so that the original gastrula mouth is to be sought at the jnnc- 
tnre of the œsophagus with the stomach. The resophagus (Figs. 
45, 48, and 49, œs.) is a comparatively short and narrow canal 
with a wall composed of densely ciliated cylindrical cells, among 
which are scattered numerous unicellular glands [m.gl.). Thus 
the wall does not differ in structure from that of the hood or of 
the collar. 

]\Iastermax lias described an unpaired ectodermal invagina- 



5r)4 T. TKEDA : 

tion situated in front of tlio moutli and just under the ganglion. 
It is called the " suhneural gland." Here again I am not in a 
position to confirm his view. In spite of repeated examinations 
on living sjiecimens, I have heen unahle to discover any structure 
which has the slightest resemhlance t(^ the sul)neural gland. To 
judge from my own ol^servntion, the " sul)ueural gland " as well 
as hoth the " oral-" and the '' pharyngeal grooves " of this author 
are products of his fixing method. In preserved specimens, it is 
fi'equently noticed tliat the lower wall of the hood is hulged out 
and downwards in front of the mouth (fig. K), prom.), and, as a, 
result of this, there is hrought ahout on the wall behind the 
prominence a depression, which appears on sections as a tolerably 
deep pit (fig. ()3 a). 

Sloiiiach. The stomach forms the largest and widest portion 
of the alimentary canal. It is especially long in the larvie of type 
I), in which it extends l)elow nearly to the plane of the perianal 
l)elt. Tlie greater part of the stomach wall is composed of 
cylindrical cells with short cilia whose spherical nucleus is usually 
situated in the centre of the cell (figs. 45, 48, and oO a, dow.). 
But the anterior portion of the wall along the mid-doi'sal line and 
the posterior portion iiear the intestine greatly differ in their 
constituent cells from the remaining parts. They consist exclu- 
sively of tall ciliated cells which contain elongated nuclei, and 
are, in a word, of the tpso])hageal type (fig. 45). In the full 
gi'own larva, the ventro-latei'al portions of the stomach wall form 
two digestive areas placed in the neighbourhood of the septum. 
Here the cell boundaries are indistinct and the nuclei are im- 
bedded in a common mass of protoplasm, in which remains of 
various unicellular organisms ai'e enclosed. 

From the anterior end of the stomach a pretty wide and 



ox DEVELor:\rEXT ETC. OF piior.oxrs. ooö 

unpaired diverticulum ]n'otrucles itself forwards (fig. 14, dir.). The 
position of the organ is wholly ventral to the œsophagus {œs.), 
and the form is like that of a sac eomiiressed in the dorso-ventral 
direelion ((igs. 4"), 4*.>, 50«, and 63 r;, div.). The iutcrnal cavity 
is continuous with the stomach cavity. The roof of the diverticulum 
in the fresh state generally shows a reddish hrown tint. This 
coloration is (hie to the superposition of the fundamental brown- 
ish coloui- on the hfemoglobin of the blood corpuscles which, in 
advanced larvcT, overlie the organ in either one, or two masses. 
The cells which compose the diverticular wall are tall and slightly 
curved, and are ciliated on the free ends (fig. ÖO «, div.) In fully 
grown larvœ of every type, each cell constantly contains a single 
small round vacuole in its distal end (fig. <)! />). The vacuoles can 
not ])e stained bv most of the staining reagents. I have seen 
them in the diverticular wall of a highly advanced larva belonging 
to type A, which had already evaginated the ventral ponch ; even 
in this case, they were found only one in each cell (fig. 61 b).' 
The whole of the diverticulum is lined externally with the thin 
j)eritoneal layer (see the above figure). 

Many previous observers have noticed this organ and have 
called it by various names : — 

J. MuLLER ('46)— " Blinddarme " (paired), 
Gegenbalte ('54) — " Haufen der Leberzcllen," 
Wagexer ('47) — " Leberblinddarme," 
Clapaeède ('63) — " A dark mass with glol)ules " (after 

Mastermax), 
Metschnikoff ('71) — " brown specks," 
Wilson (A.G.) ('81)— ''glandular lobes of the stomach," 
Masterman ('97) — " Xotochord " (paired), 
TvOULE ('981 — " Isotochord " (unpaired). 



556 T. iKEPA : 

Thus it will be seen that while some authors have apjxirently 
confounded the organ with the overlyins: corpuscle masses, others 
have considered it to ])e a glandular appendage of the stomach, 
and still others have regarded it as a skeletal structure. Accord- 
ing to Mastermax, who maintains the last mentioned opinion, the 
stomach wall is produced, in the antero-lateral region, " into two 
remarkal)le diverticula which in the fully developed larva lie as 
a pair of elongated organs, Notochords, laterally to the oesophagus " 
('97, /.<?., p. 302). The organs are said to soon undergo a remark- 
able metamorj^hosis, i.e., vacuolization. The vacuoles are produced 
successively one after another at the distal ends of the cells and 
are arranged alternately in several layers. On account of these 
facts Masteeman rejects the view that the organ is of a glandular 
nature, and holds that it is to Ijc compared in function and structure 
with the notochord of the Chordata. In 1898 Roule published 
his third paper on Äct'inoirorha, in which he denied that the organ 
is double in number and lateral in position to the œsophagus, 
but admitted the vacuolization in the larva of Phoronis sabatieri 
{~P. 2)-^''fmwopkiIa CoEi). 

I can not at present decide wdiether the variations in the 
mnnber of the diverticulum and in the deoTce of vacuolization are 
of s2:)ecific value or not. For the present I must be content with 
simj^ly noting that the stomach diverticulum in the larva^ studied 
by me is constantly unpaired and undergoes no farther vacuolization 
process than the production of one vacuole in each celh 

Intestine. The intestine which leads to the anus is a slender 
canal whose wall is composed of a layer of somewhat cylindrical, 
ciliated cells with round nuclei (figs. 4.") and 48, iniÀ. 



ox DEVELOPMENT ETC. OF i'HOllOXI.S. 557 



4. Organs of JlcsoOladic Orujln. 

As the UR'Soblastic organs have been l)ul httle studied in their 
development, so their strueture and fate after metamorphosis are 
very imperfeetly known. Although I have endeavoured to make mv 
study of the organs as exhaustive as possilde, some important 
questions remain yet unsoh'ed. The prineipal organs to be des- 
eribed in this place are the muscular elements, the vascular system, 
and the iicpliridia. 

Nephridia. 1 will treat these imder the mesoblastic organs, 
for, though the nephridial canals are of ectoblastic origin, the 
organs as a whole bear intimate relations to the mesoblast. ]\Iost 
of the earlier (jbservers overlooked the presence of nephridia in 
the larva. The first discoverer was AVagexer ('48), whose descrip- 
tion is, however, very meagre and gives us no exact idea of the 
organ. Caldwell in his preliminary note ('82-'83) has given a 
detailed description of the nephridia. According to his view, the 
nephridial canal at no time during larval life, opens into the bodv- 
cavity. 

Master.aiax ('97) has descril)ed the excretory system of the 
larva in detail and has suggested an hypothesis which seems to 
me to l)e an extraordinary one. Each of the three " segments " 
of the larval body, he concludes, is provided with a paired organ 
which perf )rms the excretory function. The three ])airs of organs 
are called respectively the '' proboscis pores," the "collar nephridia," 
and the '' trunk ne[)hridia," Of these, however, the presence of 
the iirst and the third is, as I have before pointed out, very 
doubtful. The second pair, or the collar ne[)hridia, are the organs 
which I consider to be the nei)liridia. ^Mastekman's views on the 



DOS i. IKEDA : 

structure of the uepliridial canals are m the main similar to those 
of Caldwell, except in one important point, rlz., that the canals 
are said to open ])y means of funnels into the collar cavity. 

When a larva of any type is examined in the living state, 
the proximal ends of tlie organs are seen, as descriljed l)y Wageister, 
as t\v(j bonquet-shaped masses which are formed 1)V a crowchng 
together of the excretory cells (fig. 13, neph.). They are placed 
symmetrically one on each side of the stomach and in front of 
the postoral septum. Each of them consists of two ])arts, the 
nephridial canal and the excretoiy cells. The former is com- 
posed of a layer of cubical cells, and contains a narrow lumen 
which ends l)lindly at the internal end and distal ly leads to 
the nephridial pore lying on either side of the pouch pore. 
The greater part of the nephridial canal, together with the excretory 
cells, rests on the upper surface of the postoral se2)tum. Fig. 50 h 
shows a cross section of the larva, passing through the left neph- 
ridial canal near its internal blind end, where the excretory cells 
adhere. In the above figure, a small cell mass inep.c) on the right 
of the figure, shows the cut end of the nephridial canal which is 
attached to the septum (//^e.s.). In the figure, the left canal {nepx) 
containing a small lumen, is found applied to the somatic walls. If 
traced a little downwards, these two canals become attached to and 
imbedded between the two layers of the somatic walls, and are no 
more to l)e seen in the collar cavity. Such a state is represented 
in fig. 00 c, in which the two canals {ncp.c.) are wholly imbedded 
in the somatic walls on both sides of the stomach {.siiii.). This 
.condition is more distinctly shown in figs. 47 (a-c), which are 
taken from serial longitudinal sections of a larva of type A with 
12 tentacles. These figures sbow only one jujrtion of the skin, 
where the nephridial canal and the somatic attachment of the 



ON DEVEt0r3IENT ETC. OF PlIOKONJ«. ôÔO 

postoral septum {iiicf^.) arc situated. In üg. 47 a, one portion of 
the peritoneal layer of the stomach wall is als<j represented. Xow 
we see in the first two hgures of the above series, that the nephi-i- 
dial canal {iiep c.) which is here imbedded in the somatic la}'ers, 
lies distinctly below the septum {iiics.). So, in the third ligure 
the nephi-idial ])ore {ncp.o) is seen as a small pit in the trunk 
Avail, which is situated considerably Ijclow tlie septum. The in- 
fraseptal position of the nephridial pores has also been acknowledged 
l)v Caldwell. '1'1iou<>;1i Mastermax has made no direct state- 
ment ou this poiut, it may safely be inferred from his figures, 
that he must have regarded the pores as lying in front of the 
septum. 

Fig. '31 '( represents a longitudinal section through the middle 
of the sup]'ase])tal port i ou of the nephridial canal. Here the canal 
appears as a couiparatively long tube with a narrow lumen; it is 
invested throughout with a thin mesoblastic ejnthelium. At its 
upper extremity where the lumen disappears, a certain number of 
s})indle-shaped excretory cells is found aggregated together. In 
fig. ül b, which is taken from the same series as fig. 51 a, the 
canal has wholly disappeared from the section, leaving only a 
bunch of the excretory cells (e.vo.c.) adhering to the septum {mes.). 
AU of these spindle-shaped cells have their nuclei in the swollen 
ends. I have never found either among, or in, the neighbourhood 
of the cell bunch any perforated excretory cells Ijeariug many 
processes, — cells which are said to have l)eeu j)resent in the Acli- 
notrocJia studied by Caldwell. 

Masteemax considers that each Ijonquet of the excretory cells 
is composed of a cellular mass traversed by a system of minute 
funnels ; and that these funnels communicate with the main canal 
of the ncphridium as well as with the collar cavity. But I may 



o(30 î. iiCEbA : 

say with certainty tliut, at least in the lavva^ Htuclietl 1)y nie, 
there existed no snch fnnnel-system nor any sneh free eonnnuni- 
eation between the collar cavity and the nephridial canals. The 
same negative resnlt was also reached l)y nie in my examination 
of the just metamorphosed larva of tyjie A. Thus in tig. 0-4/ 
>vhicli is drawn from a section through the tip of the nephridium, 
the excretory cells [exec.) still remain compactly grouped on the 
1)1 ind ti]) of the canal {m'p.c), hut are not traversed by any sort 
of canal-systems. 

Muscular Sydcm. The muscular system of Adinotrocha re- 
mains in a low state of development, which may account for the 
fact that most previous observers have paid no particular atten- 
tion to it. It was therefore of much interest to discover two pairs 
of t(jlerably well developed muscles, which had hitherto remained 
apparently unknown. They show a strong resemblance to the 
retractor muscles which have been known in many forms of 
Trocho2?hora larvœ. 

Though the longitudinal and circular muscles of the liody 
wall may be observed with tolerable distinctness in the living larva, 
they are usually very poorly preserved after hardening. They are 
all subdermal in position and very delicate in structure, so that 
as a rule they can not be satisfactorily distinguished from the 
underlying peritoneal layci'. However, in certain jn'cparalions of 
the entire larva the circulai- muscles of the upper and lower 
walls of the preoral lobe and of the trunk body wall could he 
detected as line deeply stained fibres. In the same way the 
lonuitudinal muscles of the collar wall, esiK'ciallv in the larva, 
of type V, were fairly traceable. The larva of that ty})e also 
exhibited a peculiar arrangement of the circular muscles of the 
trunk, in that these formed four, e(iuidistaiit, hMgitudinal series 



ox DEVELOPMENT ETC. OF PITOEOXTS. '3<U 

Mvoimd the poi-ipliery of the trunk (see fig. 12). Tii the Livva 
of type C I have always found a comparatively thick layer of 
circular muscles. In fig. öl b, which shows a portion of the 
trunk wall containing the nephridial canal [nepx.), the muscles 
are represented as a thin filn'ous layer [rir./j/.) intercepted 
hetween the ectoderm and the peritoneal epithelium. The fioor 
of tlie month just opposite the stomach diverticuhnn, is always 
associated with a particularly well developed nmscular sheet. The 
mesenchymatrms unicellular fibres which traverse the preoral cavity 
are to be regarded as a kind of primitive muscles. The most 
highly developed parts of the muscular system of the somatic, and 
of the splanchnic, walls are to be found in the muscular sheaths 
of the ventral pouch and of the dorsal wall of the stomach in 
the advanced larv<r of all types. Each of them is formed of 
a thick layer of enormously elongated muscular cells which stand 
vertically to the ectoblastic, or the entoblastic, wall as the case may 
be (figs. Ö8 c and ()8 e, in.sh.). The sheath of the stomach wall 
is thickest along the mid-dorsal line of the stomach ; it is shown 
in fio\ 58 c and fio-. <).3 e, the former figure beins; taken from a 
cross section and the latter from a longitudinal section through 
the dorso-antei'ior regiyn of tlie stomach. The muscular sheath 
(or the external wall) of the ventral pouch is essentially similar 
to that of the stomach. 

The Betracior 3ü(SirIes can be constantly detected in every 
type of the larva as two slender threads on l)oth sides of the 
œsophagus (figs. 12, 13, 14, l-j, ret.). They spring from the hind 
lateral corners of the ganglion {(jL) and run divergently downwards 
until they insert themselves in the collar walls between the first, 
and the second, tentacles. In order to obtain a clear idea of the 
position of these muscles, it is necessary to study them in sections ; 



002 T. IKEDA : 

the larvœ of types B and D are best suited for this purpose, as 
the muscles in these are remarkably large and hjng. Figs. 63 a-c 
are taken from serial sagittal sections through the right side of 
the nerve ganglion. In figure a, a median section, we see beliind 
the ganglion nothing 1)ut tlie preoral septum {mex\) In the next 
figure, h, the septum is found to have shifted to a more anterior 
position and its dorsal termination is accompanied by a strong 
muscle-band {ret.). This band corresponds to that portion of the 
retractor muscle which is nearest to its anterior insertion on the 
postero-lateral side of the ganglion. The two muscles are shown 
in fig. .39 (/ {ret.), a cross section through the posterior recess 
(p.r.) of the preoral cavity, where they spring directly from the 
septum (^ine^\). In fig. Go c, which shows a more lateral region 
than fig. 63 b, the muscle is found to have retreated far backwards, 
touching with its posterior portion the œsophageal walls {pes.). The 
posterior insertion of the muscles on the somatic walls are best 
studied in serial cross sections of the larva. Figs. 58 a and b show 
two cross sections passing tlirough the mouth {a) and through the 
middle of the œsophogus (/;). In both figures the muscles {ret.) 
are found on l)oth sides of tlie œsophagus {(cs.). A little further 
down they soon detacli themselves from the œsophagus and begin 
to traverse freely the l)ody-cavity (larval collar cavity), and after 
that thev again apply themselves to the skin on each side 1)etween 
the first and second tentacles {f and /" in fig. 58 b). 

There is also present another pair of muscles, which can be 
discovered only in the larvae of type C. They are so very long 
as to equal the entire length of the trunk (fig. l^^b, ret'.). They 
arise on each side from tlie somatic walls just above the nephri- 
dial pore and run straight downwards traversing the trunk cavity, 
and ending at the terminal portion of the intestine. Fig. 57 a 



ox DEA'i:rX)PMEXT ETC. OF PIIOrtONIS. ')()0 

represents a transverse section through the middle portion of the 
trunk where tlie stomach {sl//i.) joins the intestine {int.). There 
tlie muscles appear as two small striated masses {/'ef.') lying on 
l)iHh sides of the intestine. As represented in fig. 57 h, which is 
taken from the portion of the body wall c()Dtaiiung the nephi'idial 
canal {nepj'.), each of the nuiscles is in its origin traceal)le to the 
circular muscle layer {cir.vi.) which is sul)dermally interposed het- 
ween the ectohlast and the peritoneal epithelium. When these 
two muscles, in their downward course, reach the level of the 
terminal portion of the intestine, they fuse together into one on 
the dorsal side of the intestine. I think that the above men- 
tioned muscles are a kind of retractors serving to contract the 
trunk of the larval body. I can not help thinking that the 
" Afterbänderung " of Wagener ('47) is probably simj^ly the post- 
erior portion of the nuiscles in question in the proximity of the anus. 
Vascular sydern. It is a well known fact that the closed 
vascular svstem of Fhoronis offers one of the greatest obstacles to 
the idea entertained by some naturalists that the animal is of the 
Polyzoan type. ]\Iany writers are, therefore, much inclined to 
attrihute the simple body organization of Fhoronis to secondary 
adaptation, and to erect the animal into a distinct order very closely 
related to the C'hordata. Putting aside for the present all the- 
oretical speculations, it is of great importance in ascertaining the 
phylogenetic relation of the animal to note that one portion of the 
larval body-cavities is transformed into a blood vessel, ar.d that 
the simple and rudimentary vascular system of u4ctinotrocha uudei'- 
goes a wonderful change and suddenly attains the high organi- 
sation seen in the adult during metamorphosis. 

Krohn ('50) proved that the " Leherzellen " of Wagener 
nud Gegexbat'r were really blood corpuscles. Hfwvever, he did 



504 I. IKEDA : 

not discover any l)lood vessel in Acilnofrocha ; lie thonglit that 
the blood vessels of the metamorphosed worm arose in the cor- 
pnsele masses of the larva. 

Claparède ('63) mentioned a rino-like vascular canal nnder 
the tentacalar row of the larva, hut did not explain its nature. 

Schneider ('62) discovered two vessels in Actinoiroclid, which 
ran parallel aloni;- the mid-dorsal line of the stomach. 

jMetsciinikoff ('71) descril)ed and figured in a larva of 10 
tentacles the "feinen Häutchen" situated just ahove the invagina- 
tion pou<'h, Avhich was said to 1)e the " Gefässanlage." Besides, 
it is stated that he saw a ventral " sinusartigen Schlauch " which 
covered the greater part of the stomach and communicated an- 
teriorly with the collar cavity. According to his view, this 
Schlauch should give rise to the ring vessel of the adult. But 
what are really meant hy the " Schlauch " and the " Häutchen " 
is not clear from his text and figures. 

Wilson ('81) confirmed the main points of Metschntkoff's 
observations, l)ut disproved the 25i"^^ence of a blood vessel along 
the intestine, and also the free communication between the pseudo- 
hœmal space and tlie perisviceral cavity. According to this author, 
there are two sorts of corpuscles : the one kind floats in masses in 
the perivisceral cavity, and the other (the pseudohœmal corpuscles)