DEPARTMENT OF MARINE BIOLOGY

OF

CARNEGIE INSTITUTION OF WASHINGTON
ALFRED G. MAYER, DIRECTOR

PAPERS FROM THE TORTUGAS LABORATORY

OF THE

CARNEGIE INSTITUTION OF WASHINGTON

VOLUME V

WASHINGTON, D. C.
PUBLISHED BY THE CARNEGIE INSTITUTION OF WASHINGTON

1914

s^

0^^

CARNEGIE INSTITUTION OF WASHINGTON
Publication No. 182

3^

were Tirst issued

APUi 1914

PRESS OF

THE NEW ERA PRINTING COMPANY

LANCASTER, PA.

CONTENTS.

Page.
I. In Memoriam: George Harold Drew. By Alfred G. Mayer 1 plate i-6

II. On the Precipitation of Calcium Carbonate in the Sea by Marine Bacteria,
and on the Action of Denitrifying Bacteria in Tropical and Tem-
perate Seas. By G. H. Drew 2 maps, 4 figures 7-45

III. Preliminary Remarks on the Geology of the Bahamas, with Special Refer-

ence to the Origin of the Bahaman and Floridian Oolites. By T.
Wayland Vaughan 47~54

IV. Building of the Marquesas and Tortugas Atolls and a Sketch of the Geologic

History of the Florida Reef Tract. By T. Wayland Vaughan. . 55-67

V. Some Chemical Characteristics of Sea-water at Tortugas, Florida. By

R. B. Dole 69-78

_^ VI. Observations upon the Growth-rate and (Ecology of Gorgonians. By L. R.

Gary ' 2 plates 79-90

VII. Growth-Changes in Brittle-Stars. By H. L. Clark 3 plates 91-126

VIII. The Early Influence of the Spermatozoan upon the Characters of Echinoid

Larva. By D. H. Tennent 11 figures 127-138

IX. Studies of Jamaica Echini. By Robert T. Jackson 21 figures 139-162

X. The Spermatogenesis of the Mongoose; and a Further Comparative Study
of Mammalian Spermatogenesis, with Special Reference to Sex
Chromosome^. By H. E. Jordan i plate, 9 figures 163-180

XI. Bryozoa of the Tortu*gas Islands, Florida. By R. C. Osburn 23 figures 181-222

I.

IN MEMORIAM.

George Harold Drew,
1881-1913.

One plate.

GEORGE HAROLD DREW

GEORGE HAROLD DREW.

(1881-1913-)

George Harold Drew, only son of George Drew, of Devonshire, England,
was born on October 23, 1881, and died suddenly on January 29, 1913.

From early boyhood chemistry and microscopy fascinated him, and
while yet a child he studied in a disused stable on his father's place, which
was made to serve as a rough laboratory.

Until the age of ten he was taught by his mother, and then he was sent
to New College, Eastbourne, Sussex. Being deeply interested in the scien-
tific side of medicine, he entered Cambridge University, where he became a
scholar of Christ's College, and gained the distinctions of Exhibitioner,
Foundation Scholar, and Prizeman, afterwards becoming Schreiner Uni-
versity Scholar, and a student at Exeter Hospital. During his five years
at Cambridge he took the entire medical course, but afterwards decided
to devote his life to research, inclination leading him to choose the scientific
side of his profession rather than to engage in its active practice.

After leaving Cambridge, he served for a year as University Scholar of
St. Mary's Hospital, London, and subsequently, after working for a short
time at the Port Erin Marine Laboratory, he went to Plymouth and associ-
ated himself with the Marine Biological Laboratory, wherein most of his
scientific work was accomplished. He was one of those chosen by Captain
Scott to serve upon the scientific staff of the British Antarctic Expedition,
but he declined the honor, preferring not to break the continuity of his
studies at Plymouth.

From 1910 until 1912 he held a Beit Memorial Fellowship for Medical
Research in cancer; in the summer of 191 2 he was appointed John Lucas
Walker research student in the University of Cambridge, and on January
I, 1913, he became a Research Associate of the Department of Marine
Biology of the Carnegie Institution of Washington.

His connection with the Carnegie Institution of Washington began in
191 1, when he accompanied the expedition to Jamaica, afterwards studying
at the Tortugas, and upon a second visit to America in 1912 he spent a
month at the temporary laborator>' established at Golding Cay, Andros
Island, Bahamas.

He was to have been a member of the projected expedition to Torres
Straits, Australia, in 191 3, and his preparations were being matured at the
time of his sudden death.

The saddest losses that the world sustains are often those it heeds the
least, and all who knew him realize that this young man, with his keen

3

4 Papers from the Marine Biological Laboratory at Tortugas.

insight for research, his broad culture, and well-rounded special training,
was destined to become one of England's great men of science — a position
which his rare personal charm would not only have graced, but would have
rendered him powerful for advancing the intellectual welfare of his country.

In 1910 he carried out some interesting observations upon the repro-
duction and early development of the kelp {Laminaria) of the coast of
Devonshire, proving that there is an alternation of generations in the life
cycle of this seaweed ; for the free-swimming sexual cells fuse in pairs and
then settle down upon the bottom and develop into a chain of cells, any
single cells of which may develop into a Laminaria plant.

His published works cover a period of only four years, yet in this brief
time he was the author or joint author of fifteen papers.

His earliest paper, in 1909, is upon parasitic and other diseases of fish,
in which he describes various forms of tumors and cancers.

This paper was followed by one of an experimental nature written in
association with de Morgan as its joint author, and it was shown that, in
Pecten, injurious bodies implanted in the tissues soon become surrounded
by an agglutinated layer of blood-corpuscles and these are replaced by a
mass of fibrous tissue which encapsules the foreign body very much as such
bodies are incased by mesenchyme in injured vertebrates.

In 191 1 Drew announced that, as a result of injury, cysts lined by
columnar ciliated epithelium could be formed from the fibroblasts in Pecten.
He also found that the blood corpuscles of Lamellibranchs are capable of
ingesting and destroying bacteria, and that if the animal be wounded so
that blood escapes, some of the corpuscles adhere to the cut surfaces and
then send out slender processes which fuse with those from other corpuscles,
thus forming a net-work the fibers of which finally contract and close the
wound.

He also found that the repeated application of a strong solution of iodine
to a circumscribed area of the skin of Fundiilus produces an inflammatory
tumor having some relation to certain abnormal growths seen in nature in
fishes.

While at Plymouth the subject of the cause and nature of tumors and
cancers engaged his major interest, and he did much histological work upon
the fate or growth of transplanted tissues. His last work was upon the
culture of living tissues from the frog and the dogfish in plasma outside the
body of the animal. In this study he achieved decided success, having
with his usual ingenuity devised an improved technique which enabled him
to grow tissues upon microscope slides free from bacterial contamination.
The paper describing these results was completed only a few days before
he died and is published in the Journal of Pathology and Bacteriology,
Cambridge.

Some of his most interesting work, however, was done while associated
with the Department of Marine Biology of the Carnegie Institution of
Washington.

George Harold Drew. 5

In 1 9 10 Sanford, and also Vaughan, published the conclusion that a
considerable portion of the calcareous muds in the bays and sounds of
southern Florida were precipitated out of the sea-water in some unknown
manner. It remained for Drew, in 191 1, to discover that there is in the
warm surface waters of the West Indian and Florida region, and especially
in the limestone mud itself, a bacillus which deprives the sea-water of its
nitrogen, thus causing the calcium to combine with the dissolved carbon
dioxide and to form the finely-divided limestone mud so characteristic of
coral-reef regions. Drew isolated this bacillus and found that it became
inactive in even moderately cold water, and thus it functions only in warm
or tropical seas, thriving best at depths of less than 100 fathoms. In the
surface waters of the Bahamas and Florida it is the most abundant marine
bacillus.

Drew hoped to continue these studies and to extend them to the Pacific,
for this calcium precipitation is an important factor, and has resulted in the
formation of vast beds of limestone apparently far greater in bulk than that
formed by corals.

The complex problem of the chemical balance of the constituents of sea-
water and their solvent powers under various conditions was being enthusi-
astically considered by Drew, and had he lived it was the hope of this Labo-
ratory that he might have had exceptional opportunities to continue these
studies — for truly he was the exceptionally brilliant man, for the coming
of whose like again the world must wait, we fear, for many years, for there
are but few in each generation who are gifted with his rare genius for
research.

Alfred Goldsborough Mayer.

BIBLIOGRAPHY.

1. Some notes on parasites and other diseases of fish, in Parasitology, vol. 2, No. 3, p. 193,

1909; also vol. 3, No. I, p. 54, 1910. Also a review of the same in Journal Marine
Biol. Association, Plymouth, vol. 9, No. 2, p. 246, 191 1.

2. The reproduction and early development of Laminaria digitata and Laminaria sac-

charina, in Annals of Botany, vol. 24, p. 177, 1910. Also reviewed in Journal Marine
Biol. Association, Plymouth, vol. 9, No. 2, p. 245, 191 1.

3. Some points in the physiology of Lamellibranch blood-corpuscles, in Quart. Journal

Microscop. Sci., vol. 54, pp. 605-622, 1910.

4. A table showing certain culture characteristics of some of the commonest bacteria

found in the laboratory tanks at Plymouth, in Jour. Marine Biol. Association, vol. 9,
No. 2, pp. 161-163, 191 1.

5. The action of some denitrifying bacteria in tropical and temperate seas, and the bacterial

precipitation of calcium carbonate in the sea, in Journal Marine Biol. Association,
vol. 9, No. 2, pp. 142-155, 191 1. Also in the Report of the Department of Marine
Biology, Year Book of the Carnegie Institution of Washington, No. 10, pp. 136-141,
1911.

6. Some cases of new growths in fish, in Journal Marine Biol. Association, Plymouth,

vol. 9, No. 3, pp. 281-287, I plate, 1912.

7. A note on some attempts to cause the formation of cytolysins and precipitins in certain

invertebrates, in Jour. Hygiene, vol. 11, No. 2, 191 1.

8. Experimental metaplasia, i. The formation of columnar ciliated epithelium from

fibroblasts in Pecten, in Journal of Experimental Zoology, vol. 10, pp. 349-374, 3
plates, 191 1.

6 Papers from the Marine Biological Laboratory at Tortugas.

9. A note on the application of Giemsa's Romanowsky stain to the blood and tissues of
marine invertebrates, in Parasitology, vol. 4, pp. 19-21, 191 1. . . ,. , „

10. An experimental investigation of the cytological changes produced m epithelial cells
by long-continued irritation, in Journal of Pathology, vol. 17, 1912.

1 1 Report upon the formation of calcium carbonate through bacterial action in the tropical
Atlantic, Report of the Department of Marine Biology, Year Book of the Carnegie
Institutionof Washington, No. II, pp. 136-144. 1913- ^ , , t^ , , j

12. On the culture in vitro of some tissues of the adult frog, in Journal of Pathology and

Bacteriology, vol. 17, pp. 581-583. plates 44-46, 1913. . . . „

13. The origin and formation of fibrous tissues as a reaction to injury in Pecten maxtmus,

by G. H. Drew and W. de Morgan, in Quart. Journal Microscop. Sci., vol. 55, pp. 595"
610, plate 24, 1910. Also a review of the same in Journal of the Marine Biol.
Association, vol. 9, No. 2, p. 595, 191 1. j i u

14. Note on the abnormal pigmentation of a whiting infested by trematode larvae, by

F. W. Gamble and G. H. Drew, in Journal Marine Biol. Assoc, vol. 9, No. 2, p. 243.

15. On the pricipitation of calcium carbonate in the sea by marine bacteria, and on the

action of denitrifying bacteria in tropical and temperate seas, in Papers from the
Marine Biological Laboratory at Tortugas, Publication No. 182, Carnegie Institution
of Washington, 1913.

II.

ON THE PRECIPITATION OF CALCIUM CARBONATE IN
THE SEA BY MARINE BACTERIA, AND ON THE
ACTION OF DENITRIFYING BACTERIA IN
TROPICAL AND TEMPERATE SEAS.

By G. HAROLD DREW.

Two maps and four figures.

TABLE OF CONTENTS.

Page.

Introduction 9

General considerations and previous work lO

Description of apparatus 13

Culture media and methods 19

The investigation of samples of sea- water taken off Port Royal, Jamaica 21

The investigation of samples of sea-water taken around the Dry Tortugas 24

The investigation of samples taken from a point 70 miles west of Ushant Island, France 27
The investigation of samples from the Marquesas Keys, Florida, and experimental pre-
cipitation of calcium carbonate by bacterial agency 29

Some considerations on the physiography of the Tongue of the Ocean and Andros

Island 31

Bacterial investigations in the deep water of the Tongue of the Ocean 34

Hydrographic observations in the Tongue of the Ocean 37

Bacterial investigations of the chalky mud flats which are being deposited to the west

of Andros Island 41

Conclusion 44

8

ON THE PRECIPITATION OF CALCIUM CARBONATE IN THE SEA BY
MARINE BACTERIA, AND ON THE ACTION OF DENITRI-
FYING BACTERIA IN TROPICAL AND TEMPERATE SEAS.

By G. Harold Drew.

INTRODUCTION.

The investigations described in this paper were made in the summers of
191 1 and 1912 under the auspices of the Carnegie Institution of Washington.
The original intention was to study the action of marine denitrifying bacteria
in tropical seas. The discovery, during the course of the experiments, that
these denitrifying bacteria also possess the power of precipitating calcium
carbonate from soluble calcium salts present in sea-water has perhaps, by its
geological significance, somewhat overshadowed the interest of the primary
object of the work. The main contentions raised in this paper are:

(i) That in the seas of the American tropics bacteria exist which are
actively precipitating calcium carbonate from the calcium salts
present in solution in sea-water. It is suggested that this bacterial
action has been a very considerable factor in the formation of
chalk and many other varieties of sedimentary rock, chiefly or in
part composed of calcium carbonate. It is also contended that
the vast deposits of chalky mud now being formed to the west of
the Bahamas and in the neighborhood of some of the Florida Keys
are precipitated by bacterial agency and that a similar process
plays an important part in the cementation of fragments of coral
and other detritus into compact coralline rock.

(2) That the destruction of nitrates by bacterial action in the seas of the
American tropics is far in excess of that occurring in temperate
waters. Hence an explanation is afforded of the relative scarcity
of plant life (and consequently of animal life) in tropical as corn-
pared to temperate seas, in accordance with the terms of Brandt's
(2 and 3)^ hypothesis.

Preliminary notes on this work have already been published in the
Tortugas Laboratory Reports^ for 191 1 and 1912. The chronological
sequence of the investigations will be followed in the account given here of
the experimental work.

' The figures in parentheses refer to the bibliography, p. 45. „ . »-

2 Carnegie Institution of Washington, Year Book No. lo, 1911. pp. 136-141; \ear Book No. 11, 1912. PP-
136-144.

10 Papers from the Marine Biological Laboratory at Toriugas.

GENERAL CONSIDERATIONS AND PREVIOUS WORK.

It is generally conceded that the plankton of tropical and subtropical
seas is far less in quantity than that found in colder waters.^ The zoo-
plankton depends ultimately for its food on the phyto-plankton ; hence any
factor limiting the growth of the phyto-plankton which was capable of
exercising its influence in tropical and not in temperate or arctic waters
might offer an explanation of this phenomenon. It has been shown by
various investigators that this factor is not temperature, light, or salinity,
and it has been suggested that the explanation may lie in the relative
deficiency, in tropical seas, of the nitrates or nitrogenous compounds so
essential for all plant life. A matter of common observation in support of
this view is the remarkable scarcity of algal growth in the shallow waters of
tropical shores as compared with that in temperate regions, and the fact
that in the tropics, wherever sewage or other nitrogenous waste is poured
into the sea, a free growth of algse is found.

There is at present no really reliable and accurate chemical method of
estimating the combined nitrogen in sea-water, hence the above theory
can not be directly put to the test. On the other hand, the existence of
denitrifying bacteria in temperate waters has long been known, and it
would seem a fair deduction that should this bacterial destruction of nitrates
take place with greater intensity and completeness in tropical than temper-
ate waters, an explanation of the relative scarcity of phyto-plankton in the
former would be offered. This suggestion was first made by Brandt (3)
in 1901, and is universally known as " Brandt's hypothesis." He enunciated
it as follows:

If the denitrifying bacteria of the sea, like the closely investigated denitrifying bacteria
of the land, develop a strongly disturbing activity at higher temperatures, only a relatively
small production (of phyto-plankton) would take place in the warm seas in spite of much
more favorable conditions, according to the law of the minimum, owing to the great dis-
turbance amongst the indispensable food substance; whilst, in the cold seas, more nitrogen
compounds would be at the disposal of the producers owing to the retardation or suppression
of the disturbing process. (From the published English translation.)

The presence of denitrifying bacteria has been demonstrated in Kiel
Bay by Baur (i), along the Dutch coast by Gran (9), in the open waters
of the North Sea and Baltic by Feitel (7) and Brandt (2), and in 1909 I
identified several of the species described by Gran in samples of water
obtained from the western part of the English Channel. All these deni-
trifying species have a higher temperature optimum than that of their
natural environment and this is obviously a point strongly in favor of
Brandt's hypothesis.

The chief difficulty in the way of putting the hypothesis directly to proof
lies in the fact that at present no accurate method of determining the nitrate

• For the most recent work, and full discussion of this subject, see The Depths of the Ocean, by Murray
and Hjort, p. 366 et seq., 1912. London.

On the Precipitation of Calcium Carbonate. 1 1

contents of sea-water exists, and hence it is impossible to correlate quantita-
tive plankton observations with direct analyses of the amount of combined
nitrogen present in sea-water in different localities. Much valuable work
on this subject has been done by Raben (15), but he states that his error
in control experiments averages over 30 per cent. An exhaustive study
(as yet unpublished) of all the methods of estimating combined nitrogen in
sea-water as given by various investigators has been made by Mr. D. J.
Matthews, hydrographer to the Marine Biological Association of the
United Kingdom and to the Fishery Branch of the Department of Agri-
culture, etc., for Ireland, and he has come to the conclusion that the limits
of error in all these methods are so large as to make them unsuitable for
investigations such as these. Since chemical methods are at present inade-
quate to give evidence on this hypothetical deficiency of nitrates in warmer
seas, it seemed of interest to investigate the distribution and relative activity
of denitrifying bacteria in tropical waters in comparison to those found in
temperate seas, and it was with this primary object that the present work
was undertaken.

The previous researches most closely related to these investigations are
those of Gran (9), who isolated a number of species of denitrifying bacteria
from the inshore waters of the Dutch coast. He made use of solutions of
nitrates, nitrites, or ammonium salts as the sole source of nitrogen in his
culture media, which contained only a dilute solution of calcium malate as
organic nutrient material for the bacteria. He classifies the bacteria into
four groups, according to their reactions in pure cultures towards nitrates
or nitrites.

1. Those which reduce nitrates and nitrites to free nitrogen without any

ammonia formation.

2. Those which readily reduce nitrates to nitrites. The nitrite disappears

slowly without perceptible formation of free nitrogen and some
ammonia Is formed.

3. Those which can not reduce nitrates to nitrites, but which are capable

of slowly removing the nitrite without perceptible formation of
free nitrogen. Though the nitrites are not reduced, yet they can
serve as the sole source of nitrogen for the growth of the bacteria.

4. Those which can not reduce and are not capable of assimilating either

nitrates or nitrites, but will flourish when ammonium salts are
present.

In Investigations on samples of water taken In the English Channel some
10 miles ofT Plymouth, I was able to recognize species belonging to the
second group of Gran's classification, but could not detect the presence of
species belonging to any of the other groups, and it would seem probable
that these other groups are chiefly composed of littoral forms.

In fluid culture media inoculated with samples of sea-water and kept
at a temperature of 28° C, Gran found that the formation of nitrite was
detectible in from i to 2 days, and that eventually all the nitrate and nitrite
was destroyed in the majority of cases, especially if the cultures were

12 Papers from the Marine Biological Laboratory at Tortugas.

reinoculated at intervals. In my experiments I was able to obtain similar
results in cultures kept at 30° C. after 8 days; in cultures kept at 15° C. the
first formation of nitrite was detectible in from 5 to 6 days, but denitri-
fication never proceeded beyond this stage.

Baur (i) showed that the optimum temperature for growth and denitri-
fication of the species described by him lay between 20° and 25° C. when
the bacteria were grown in fluid culture media containing peptone.

The most important work on the distribution of marine bacteria is that
of Fischer (8) in 1886, 1889, and 1893, but he does not enter into the chemical
activities of the species found, so that the observations do not throw much
direct light on problems of the metabolism of the sea. The variations in
the number of bacteria found in different surface samples from positions in
mid-ocean are somewhat surprising and difficult to account for. Deeper
samples were taken by means of a water-bottle made of brass, but in view
of the now well-known bactericidal action of metals, and of copper in
particular, I do not consider that any great value can be attached to these
observations. With the exception of Fischer's work little seems to have
been published on the general distribution of marine bacteria.

A point that has not yet been considered is the origin of the nitrate supply
in the sea. Nitrates are absorbed by diatoms and the phyto-plankton in
general, and are presumably built up into complex nitrogenous compounds
within the plant. If these compounds, on the death of the plant, are broken
up and the nitrogen again rendered available for use in the form of nitrates,
a series of reactions must be gone through which may well be performed by
bacterial agency, and this also applies to the waste nitrogenous products of
animal metabolism. In addition, it has been shown that nitrates are actu-
ally decomposed by the denitrifying bacteria,, which would thus tend to
keep the nitrate concentration down to the level necessary for their own
existence and would come into competition for this essential with other
forms of plant life. If the bacteria are successful in decomposing nitrates
to the extent of entirely removing the nitrogen from all chemical combi-
nation, as seems probable from the experiments in cultures, it follows that
there must be some source of nitrates in order that the concentration in
the sea may remain constant. The existence of nitrifying bacteria, which
are capable of absorbing and combining with the free nitrogen of the air
and eventually giving rise to nitrates, has been shown by Keding (10) and
Keunter (11), but these have so far only been found on the bottom close to
shore, or apparently living in symbiosis with algae or plankton organisms.
Similarly Thomsen (16) has demonstrated the presence on the bottom of
inshore waters of bacteria which are capable of forming nitrites from
ammonium salts, and others which can convert nitrites into nitrates. It
would seem possible that similar bacteria having a nitrifying action remain
to be discovered in the open sea.

The precipitation of calcium carbonate in the sea by bacterial agency is
apparently a line of investigation that has not previously been suggested or

On the Precipitation of Calcium Carbonate. 13

followed. Both Baur (i) and Gran (9) made use of calcium salts in their
culture solutions in order to obviate the great increase in alkalinity that
resulted if potassium or sodium salts were used, but they have not called
attention to, or apparently realized, the probable significance of this pre-
cipitation of calcium carbonate by bacterial agency as an important factor
in the formation of various sedimentary calcareous rocks in tropical seas.

The subject of the precipitation of calcium carbonate in sea-water has
been dealt with by Murray and Irvinj (14) in 1889 and again by Murray
(Sir John) and Hjort (13) in 1912, and they ascribe the precipitation to the
interaction of ammonium carbonate, derived as an ultimate product of the
decomposition of nitrogenous organic matter, with the calcium sulphate
present in sea-water, according to the equation

(NH4)2C03 + CaSOi = CaCOs + (NH4)2S04

Expressed in the terms of the ionic hypothesis, this reaction can be explained
by the statement that CaCOa must be precipitated when the product of the
concentration of its ions Ca and CO3 exceeds a certain limit; an increase in
the concentration of CO3 ions is produced by the advent of (NH4)2COa,
which is partially ionized into NH4 and COs, and hence the product of the
concentrations of Ca and CO3 ions is increased and CaC03 is thrown out
of solution.

Though this reaction has been conclusively shown to occur under experi-
mental conditions, where nitrogenous organic matter has been allowed to
putrify for some time in sea-water, yet it is obvious that its effect must be
purely local and must be confined to the immediate neighborhood of the
decaying organic body which gives rise to the formation of (NH4)2C03.

In this paper the precipitation of CaCOa in an unorganized state alone
is dealt with. The formation of the calcareous skeletons, tests, and shells
of animals, and the skeletons and platelets of algse, which play an immensely
important part in the constitution of marine bottom-deposits, is beyond the
scope of this work.

DESCRIPTION OF APPARATUS.

In 191 1 the apparatus at my disposal was of a somewhat primitive
nature, as it is difficult to know beforehand exactly what gear will be nec-
essary in a new field of work. In 191 2 a more complete outfit was available,
and the Carnegie Institution's yacht Anton Dohrn was especially fitted for
my requirements.

For deep-sea work the motor trawl winch was modified so as to carry
fine sounding-wire, and a derrick was rigged aft, projecting over the stern
of the boat, over which the wire was led. The motor winch is sunk below
the level of the deck, a commendable arrangement, as it can be covered
over with hatches when not in use and so affords great economy of deck
space, and also has the advantage of bringing the weight of the winch nearer
the water-line and avoiding the unstability that may be caused when a
hea\y winch is fixed on deck.

14 Papers from the Marine Biological Laboratory at Tortugas.

The sounding-wire was 2.2 mm. in diameter and consisted of four
strands of eight wires each, made of high-tensile steel; the breaking strain
was given as 400 pounds, but in practice I have no hesitation in saying that
it far exceeded this figure. The wire was very difficult to kink, and did not
show any tendency to untwist or permanently stretch under a tension of
about 350 pounds; it proved in every way satisfactory.

For measuring the length of wire run out, one of the fathom-measuring
sheaves made by the Telegraph Construction and Maintenance Company
of London was used. This consisted of a sheave containing a steel wheel
about 12 inches in diameter, grooved for and made especially to fit the wire;
the length of wire run out is measured by the number of turns of the wheel
indicated by a dial on the side of the sheave. The dial has two hands show-
ing fathoms and hundreds of fathoms; the hands revolve backwards on
winding in the wire, and so again register zero when the sounding is com-
pleted. The axle of the wheel revolves on simple bearings, avoiding thus the
slight inaccuracy if ball bearings are employed.

Samples of the bottom were obtained with one of the "snapper rods"
disengaging an iron 30-pound weight on touching the bottom. This con-
sisted of two brass jaws closed by a strong spring, and kept apart by a
trigger; on touching the bottom the trigger was released and the jaws
closed on a sample of the bottom; at the same time the 30-pound weight,
which was only held in position through the tension of its own weight, was
disengaged as soon as the tension was relieved on touching the bottom, and
so was left behind as the wire was reeled in.

In order to tell the depth at which bottom was sounded the wire was
led through a pulley connected with a spring balance, which accordingly
registered the tension of the wire. On touching the bottom the decrease in
tension due to the release of the weight was shown on the dial of the balance.
This arrangement was not satisfactory in rough weather, as the rolling of
the yacht caused such varying tensions on the wire that it was not always
possible to tell the exact depth at which the weight was disengaged.

For obtaining samples of water for bacterial analysis a special water-
bottle was designed by Mr. D. J. Matthews. This apparatus is described
in detail by Mr. Matthews (12) in vol. ix, No. 4, of the Journal of the
Marine Biological Association of the United Kingdom, so only a brief
account of it will be given here.

The apparatus employed by previous workers for obtaining samples of
water from the deep sea for bacteriological examination has either con-
sisted of some sort of water-bottle made of metal, or else of exhausted glass
bulbs with a neck drawn out into a capillary tube which could be broken
off at the depth from which a sample was desired. The use of exhausted
glass bulbs presents considerable difficulties for depths as great as 800
fathoms; the bulbs must be strong and very thoroughly annealed, as other-
wise the slight shock caused by breaking the capillary neck is liable under
the great pressure to make the bulb fly into small fragments; another

On the Precipitation of Calcium Carbonate.

15

great disadvantage is the strong probability that the sudden reduction in
pressure, to which the water is exposed as it enters the bulb, would immedi-
ately kill any bacteria in the water. The employment of a metal water-
bottle seemed undesirable in view of the bactericidal action of metals.
In order to settle this point some test experiments were made with various
metals to see if a suitable one could be found. 100 c.c. of water from the
laboratory tanks at Plymouth, diluted i in 100 with sterile sea-water, was
exposed for 6 hours to the action of about 2 square inches of the various
metals, with the following results:

Metal.

No. of
plates.

No. of colonies
of bacteria de-
veloping from
I c.c. after plat-
ing on peptone
agar, counted
after lo days.

Aluminium bronze...
Pure copper foil

I
2
3

I

3

I
2

3

I
2
3
I
2
3
I
2
3

0
0
0

I

0

0

0

0

I

17

12

8

3

2

4

512

S6o

480

<•

•1 11

Silver (coins)

Control experiment . .

It is thus obvious that none of these metals are suitable for the work,
and probably the only metal that could be used would be platinum, which
would be prohibitive on account of the expense.

In order to overcome these difficulties, a water-bottle on a new principle
was designed for me by Mr. Matthews. A purely diagrammatic represen-
tation of its method of working is shown in figure i .

The container of the bottle consisted of a strong glass cylinder B, holding
about 250 c.c; this was closed at each end by thick rubber washers W, W,
through the center of which a short piece of thin-walled rubber tubing
passed, the tubing being sealed at the end within the cylinder. The washers
were carried by the plates Pi and P^. The cylinder was carried in a
skeleton frame, and by sending down two messengers along the sounding
wire 5 it could be first opened and then closed at any required depth. The
whole apparatus was first sterilized by steaming in a "Koch," and then
the cylinder was completely filled with 95 per cent alcohol ; the washers were
kept tight on the ends of the cylinder by strong springs (not shown in the
diagram) so that no leakage occurred and the various parts of the apparatus
were in the position shown in figure l, A. When the apparatus has been
lowered to the required depth, the first messenger is sent doAvn, this hitting
the lever Li, releasing the connecting rod Ci and allowing the plate P2 and

i6

Papers from the Marine Biological Laboratory at Tortugas.

the cylinder B to fall into the position shown in figure i, B, the cylinder
being prevented from falling on to P^ by the check rod T, which falls
through Pi until caught by the knob at its extremity. The cylinder is thus
opened at each end by the first messenger, and the alcohol being of lower
specific gravity than sea-water, diffuses out almost instantaneously, causing
an upward flow of water through the cylinder. On sending down the second
messenger, which hits the lever L2, the connecting rod C2 was released, and
by means of strong springs the plate Pi was forced down on to the top of
the cylinder, which at the same time fell on to the plate P2, and thus the
cylinder with its sample of water was tightly closed at each end by the rubber
washers, the position of the parts being that shown in figure i, C.

Ls

I P,-l

L2

C2

.W,

w.

p.-l

C2 w.

B

Fig. I.

The washers with their attached pieces of thin rubber tubing had sufficient
capability of bulging inwards to allow for the contraction of the alcohol due
to the low temperature at any considerable depth, and to its compressibility
being greater than that of sea-water; and similarly the expansion of the
sample of water, as the apparatus was hauled up, was compensated for
by the elasticity of the thin-walled rubber tubing. It is obvious that
even had a slight amount of leakage occurred, a leakage inward during the
descent of the apparatus would not vitiate the results, as bacteria would
promptly be killed in the 95 per cent alcohol, and similarly on hauling up,

On the Precipitation of Calcium Carbonate. 17

the leakage, if any, would be outward, due to the expansion of the sample
through the regularly increasing temperature and decreasing pressure, so
that the sample would not be contaminated by any of the surface layers
through which it was hauled. There was, however, no reason to suppose
that any leakage occurred, and it appears that the expansibility of the
rubber washers and tubing was sufficient to allow for the small changes in
bulk of the fluids within the cylinder after the first sterilization by steaming.
The action of the alcohol was relied on for sterilization between successive
samples, and both experimentally and in practice this method was found to
be absolutely safe, as all the marine bacteria are very readily killed by
alcohol, and they do not form resistant spores.

After the collection of a sample it was siphoned off into a sterilized glass
bottle by means of a sterilized length of rubber tubing; this method was
considered preferable to any arrangement of taps leading from the collecting
cylinder, owing to the difficulties of cleaning and sterilization which would
be involved. Part of the sample was also siphoned off into bottles which
were returned to Plymouth for analysis for salinity ; these bottles had previ-
ously been thoroughly washed, rinsed with several changes of distilled water,
and then dried in an oven; they were corked with rubber stoppers.

It was found in practice that this design of water-bottle worked extremely
well and gave very little trouble. It is to be noted that the sample of water
collected is only kept in contact with rubber and glass throughout, so that
the bactericidal action of metals is avoided.

Surface samples of water were taken in wide-mouthed stoppered bottles,
holding about 12 ounces; the samples were always taken from the bow of
the boat when moving ahead, in order to avoid any possible contamination
from the sides of the boat.

Some samples from depths up to 80 fathoms were collected off the
Tortugas in 191 1 in retort-shaped glass flasks of about 300 c.c. capacity,
with narrow, recurved, and long-drawn-out necks. These were sterilized,
exhausted, and sealed; they were then lowered in an apparatus in which
the extremity of the neck could be broken off at any desired depth by
sending a messenger down the sounding-wire, when the flasks became
completely filled with water. After hauling up, a little water was shaken
from the neck, and it was then sealed with the blowpipe. This method
avoids risk of contamination from more superficial layers of water as the
apparatus is drawn up, since the changes in pressure and temperature as
it ascends tend to cause a continuous outflow through the narrow neck until
the surface is reached.

A somewhat similar apparatus was used for obtaining deep samples
from the station 70 miles west of Ushant, but the glass bulbs were smaller
and the tube leading from them was bent at right angles to itself. Con-
siderable difficulty was caused by the breaking of the tube owing to the force
of the inrushing stream of water impinging on the wall where it was bent at
right angles.

1 8 Papers from the Marine Biological Laboratory at Tortiigas.

If this form of apparatus is used, all sharp angles in the inlet tube should
be avoided, and it should be so arranged that the inrushing stream of water
spreads itself in a fan-shaped manner over the sides of the bulb, but I do
not consider that any form of exhausted glass flask is suitable even for
depths as small as 80 fathoms.

In Jamaica no apparatus for obtaining deep samples was available, so
the primitive method of lowering a sterilized stoppered bottle with a string
tied to the stopper was employed. At the required depth the stopper was
pulled out until the bottle was nearly full and then allowed to fall back
in place. This method can only be used for very shallow depths, as the
pressure of the water at greater depths makes it impossible to withdraw
the stopper. A source of error is also introduced in that the inrushing
water passes in close proximity to the stopper and its attachment, and may
carry in bacteria which have adhered to them when passing through the
surface layers.

Temperature records were obtained in the Bahamas by means of deep-
sea reversing thermometers specially made by Messrs. Negretti & Zambra
of London. They were tested up to a pressure of 3 tons to the square inch
at the National Physical Laboratory at Teddington, England, and a table
of temperature corrections was furnished for each instrument by the same
institution. These reversing thermometers differ from ordinary thermom-
eters in having a constriction and S-shaped dilatation immediately above
the main bulb, with a somewhat large secondary bulb at the upper end of
the stem. The graduations are reversed, so that the lowest temperature
is marked near the top of the capillary portion. On turning the thermom-
eter upside down, the mercury thread breaks at the constriction and fills
the small bulb at the end of the capillary and also part of the capillary
itself. The thermometer is read in the reversed position, and when certain
corrections have been applied the temperature at which the thermometer
was reversed is recorded. The effect of the pressure of the water is avoided
by having the thermometer sealed in an outer glass case. The lower end
of this case is partially filled with mercury, in which the bulb of the thermom-
eter is immersed, thus allowing for rapid conduction of heat between the
mercury in the thermometer bulb and the surrounding water. An auxiliary
thermometer was sealed up in the same outer case as the reverser, so that
the temperature at which the actual reading was taken could also be re-
corded. In order to calculate the correction that must be applied to the
temperature registered by the reverser, three factors must be known:
(a) Temperature of thermometer at moment of reading.
(6) Kind of glass of which it is made,
(c) Volume (expressed in degrees of the stem) of the secondary bulb and

the portion of the stem below the 0° graduation.
Of these a is given by the auxiliary thermometer, and h and c were
engraved on the back of the stem of each reversing thermometer. All of
the thermometers were made of the glass known as Jena^"" 16 III, and the

On the Precipitation of Calcium Carbo?iate. 19

apparent dilatation of mercury in this glass is ttsoo"- ^1^^ correction to be

(F°+r)(r-/)

applied to the reading of the reverser is given by the formula ; ,

^^ o ^ 6,300

where T = the temperature registered by the reverser, / = the temperature
shown by the auxiliary thermometer at the moment of reading, and V =
the volume (expressed in degrees of the stem) of the secondary bulb and
the portion of the stem below the 0° mark of the reverser.

The thermometers were mounted in pairs in simple metal cases, and
were attached just below the water-bottle. They were suspended in a
vertical position by a catch forming part of the water-bottle; this was
released by the first messenger, when the thermometers fell by their own
weight and so reversed, and were hauled up in this position. This simple
arrangement proved quite as satisfactory as any of the more complicated
reversing frames which are generally in use.

CULTURE MEDIA AND METHODS.

The culture media employed for isolating and counting the bacteria in

plate cultures were the following.

/. Peptotie agar: Peptone, 2.0 grams; potassium nitrate (KNO3), 0.5 gram; sea- water,
1,000.0 c.c; agar agar, 18.0 grams.

In the earlier work less agar was used, but eventually it was found more
convenient to use a stiffer jelly, and this did not appear to appreciably
hinder the growth of the bacteria.

II. Potassium malate agar: Potassium malate (C2H3(OH)<^ ^-.QQprV i.o gram;

sodium phosphate (Na2HP04, 12H2O), 0.25 gram; potassium nitrate (KNO3),
0.5 gram; sea-water, i ,000.0 c.c. ; agar agar, 12.0 grams.

The medium was only filtered through glass-wool, so that a very slight

fioccular precipitate of calcium phosphate was retained.

///. Peptone gelatin: Peptone, 2.0 grams; potassium nitrate (KNO3), 0.5 gram;
sea-water. 1,000.0 c.c; gelatin, 150.0 grams.

This medium was used only at the Tortugas; it was necessary to keep it
cooled with ice to about 20° C, as the temperature of the laboratory some-
times rose as high as 37° C. and gelatin media will not remain solid at these
temperatures.

The following fluid media were used :

I. Gran's medium {modified): Potassium nitrate (KNO3), 0.5 gram; sodium phosphate

(Na2HP04, 12H2O), 0.25 gram; calcium malate (c2H3(OH)^^QQ^Ca),
about 5.0 grams; sea-water, 1,000.0 c.c.

Calcium malate is only slightly soluble in water, so can be added in excess.
Gran used distilled water and added 30 gm. sodium chloride per liter, but
in these experiments sea-water has been used instead.

II. Calcium succinate medium: Calcium succinate (C2H4<' ^qq yCa V 2.0 grams;

potassium nitrate (KNO3), 0.5 gram; sodium phosphate (Xa2HP04, 12H2O),
0.25 gram; sea- water, 1,000.0 c.c.

20 Papers from the Marine Biological Laboratory at Tortugas.

This medium was boiled and filtered before sterilization to remove the
slight precipitate of calcium phosphate. It was found that this medium
with the addition of the phosphate gave a more vigorous growth than if it
was omitted.

///. Calcium acetate medium: Calcium acetate (Ca(CH3COO)2), 5.0 grams; sodium
phosphate (Na2HP04, 12H2O), 0.25 gram; potassium nitrate (KNOj),
0.5 gram; sea-water, 1,000.0 c.c.

Boiled and filtered before sterilization to remove precipitate of phos-
phate.

IV. Peptone calcium acetate medium: Calcium acetate (CaCCHaCOO)^), 5.0 grams;
peptone (Witte's), 0.2 gram; potassium nitrate (KNO3), 0.5 gram; sea-
water, 1,000.0 c.c.

Media II, III, and IV were also made up with the addition of 0.2 gm.
of magnesium tartrate per 1,000 c.c.

The fluid media were made up in 1,500 c.c. resistance glass flasks, and
1,000 c.c. of medium were used for each culture.

For other purposes a simple solution of peptone in sea-water was em-
ployed (2 gm. to 1,000 c.c), and media were also used consisting of this
peptone solution with the addition of 0.5 per cent of various carbohydrates,
such as cane sugar, dextrose, Isevulose, mannite, lactose, etc., with sufficient
neutral red solution to color them, in order to test the acid-forming proper-
ties of the bacteria in the presence of carbohydrates.

The ordinary "Koch" steam sterilizer and iron oven for dry -heat
sterilization were used, and gasoline cooking stoves were found to be the
most satisfactory source of heat. It was found an advantage to use Petri
dishes with porous earthenware covers which enabled the water of con-
densation to evaporate partially; the evaporation could be checked at any
time by covering the dishes with a bell jar lined with wet filter paper. It
was usually found necessary to keep all cultures on tables with their feet
standing in dishes of kerosene, in order to prevent the attacks of ants and
other insects. In all other respects ordinary bacteriological routine was
followed, and the methods need not be further particularized here.

The reduction of the nitrate to a nitrite in fluid culture media was
tested for by the addition of 5 c.c. of 10 per cent sulphuric acid and 2 c.c.
of a I per cent solution of metaphenylene diamine hydrochloride to 25 c.c.
of the culture. The production of a brown coloration (due to the formation
of Bismarck brown) is an indication of the presence of a nitrite, and is an
extremely delicate reaction.

The diphenylamine and brucine sulphate reactions were also used when
testing for the presence of nitrates.

The formation of ammonia was tested for by the addition of 5 c.c. of
10 per cent potassium hydrate and 5 c.c. of Nessler's reagent; the white
precipitate formed on the addition of the potassium hydrate does not
appreciably interfere with the test, though it renders it somewhat less
delicate.

On the Precipitation of Calcium Carbonate. 21

Under expeditionary conditions, and in the absence of the somewhat
elaborate apparatus that would be necessary in order to estimate chemically
the amount of denitrification in cultures, it was only possible to compare the
rate of denitrification in different cultures by noting the time taken for the
first appearance of the nitrite reaction and the time taken for all trace
of nitrite or nitrate to disappear. It seems that the rate of denitrification
in culture media inoculated with equal volumes of samples of sea-water
must be a function of the number of bacteria in the sample, the temperature
at which the cultures are grown, and the specific power of denitrification of
the individual species of bacteria. Considering the rapid multiplication of
bacteria when the food supply is plentiful, up to a maximum determined
chiefly by the accumulation of the waste products of their own metabolism,
it appears that the factor of the number of bacteria in the sample may be
neglected within the limits of these experiments. For example, the number
of bacteria in 1,000 c.c. of Gran's medium at the end of 24 hours would
probably be much the same whether it were inoculated from a sample con-
taining 8 or 16 bacteria per i c.c; similarly it was a matter of experience
that the first trace of nitrite formation was observable at about the same
time, whether 5 or 10 c.c. of a given sample had been used for inoculation.

Consequently It would appear that for purposes of comparison, and
within the limits of the experiments described, if the temperature be the
same for the cultures compared, the rate of denitrification is a measure of
the specific denitrifying power of the particular species of bacteria.

In the work on the bacterial precipitation of calcium carbonate, the
precipitate (which was often so fine as to tend to remain in suspension)
was usually obtained by centrifuglng. It was either preserved in small
bottles with some of the culture fluid, or else washed first with distilled water
and then with absolute alcohol, and finally allowed to dry. These precipi-
tates were sent to Dr. F, E. Wright, of the Geophysical Laboratory of the
Carnegie Institution of Washington, who with great kindness reported
on their mineralogical properties.

INVESTIGATION OF SAMPLES OF SEA-WATER TAKEN OFF PORT ROYAL.

JAMAICA.

The work at Port Royal was done in May 191 1, but was of a very pre-
liminary nature. It was necessary to depend on a sailing-boat for obtaining
the samples; yet, owing to the remarkable regularity with which an on-shore
wind springs up every morning, but little difficulty was experienced from this
cause. No apparatus for obtaining deep samples was available, but samples
were taken from a depth of 6 fathoms by means of a bottle from which the
stopper was pulled by a line and then allowed to fall back Into place. A
measurement of the rate of denitrification in fluid culture media inoculated
with samples of sea-water was made, but Isolation of the bacteria on solid
media was not attempted. The following method was employed:

22 Papers from the Marine Biological Laboratory at Tortiigas.

Samples of sea-water were collected in sterilized stoppered bottles from
the surface and from depths of 3 and 6 fathoms from positions about 5 miles
from shore, where, from a consideration of the wind and tide, the water was
probably under truly oceanic conditions and unaffected by the neighboring
land. 10 c.c. of these samples were added to 1,000 c.c. of Gran's medium.
The cultures were kept in a moderate light and the temperature varied from
25° to 31.5° C. The average temperature during the growth of each culture
was noted.

In a typical culture made from surface water, and for which the average
temperature was 29° C, the first indication of the formation of a nitrite, as
given by the metaphenylene diamine reaction, appeared after 27 hours;
after 38 hours the brown color produced in this reaction was very intense,
the culture became cloudy, and on testing with Nessler's reagent slight
ammonia formation was apparent. After 48 hours the culture became
very cloudy and a scum of bacterial growth developed; the nitrite and
ammonia reactions remained unaltered. After 63 hours the nitrite reaction
was somewhat less marked, the ammonia reaction was unaltered, and
bubbles of gas began to appear. After 72 hours many bubbles of gas were
being produced, and the nitrite and ammonia reactions were very slight.
After 86 hours the bubbling had ceased, and no nitrite or ammonia was
present in the culture. Testing the culture for nitrates by the brucine
and diphenylamine reactions then showed that no nitrate was left in the
solution.

In the absence of a gas-analysis apparatus the nature of the gas evolved
could not be determined, but considering that it was non-inflammable, did
not turn lime-water milky, and that the nitrate originally present had been
destroyed, it seems strongly probable that this gas was pure nitrogen.
Thus in 86 hours 0.5 gm. of potassium nitrate had been decomposed by
bacterial growth. If a further 0.5 gm. of potassium nitrate were then added,
it was rapidly decomposed, and this could be repeated many times until
the other constituents of the culture medium were used up.

It was found that the rate of denitrification varied somewhat with the
temperature, and that in cultures kept at a temperature of between 10° and
12° C. no growth or denitrification occurred. The denitrification was
always more rapid in cultures from water taken from a depth of 3 or 6
fathoms than from the surface. It was also more rapid with samples taken
from the thick muddy water of a mangrove swamp, where organic matter
was plentiful.

The bacteria present in the cultures were very minute, actively motile
bacilli with rounded ends.

An abstract of the behavior of a few of the cultures is given below.

I. Sample collected 5 miles south of Port Royal, wind southeast, force 4, tide rising. Sample
taken from surface. 1,000 c.c. of Gran's medium was inoculated with 10 c.c.
of sample.
After 20 hours a slight cloud developed in the culture and faint nitrite reactions were
given.

On the Precipitation of Calcium Carbonate. 23

After 36 hours a dense cloud developed in the culture and strong nitrite reactions were

given.
After 60 hours a dense cloud and scum developed in the culture and strong nitrite and

faint ammonia reactions were given.
After 70 hours a dense cloud, scum, and bubbles developed in the culture and faint nitrite

and faint ammonia reactions were given.
After 84 hours culture was less cloudy, with much scum, no nitrite or nitrate reaction,

very faint ammonia.
Average temperature at which the culture was grown, 30° C.
Sample collected from same spot under similar conditions, from a depth of 3 fathoms.

1,000 c.c. Gran's medium was inoculated with 10 c.c. of sample.
After 20 hours a slight cloud developed in the culture and faint nitrite reaction was given.
After 27 hours a denser cloud developed in the culture and strong nitrite and faint

ammonia reactions were given.
After 38 hours a dense cloud and scum developed in the culture and strong nitrite and

faint ammonia reactions were given.
After 48 hours a dense cloud and scum developed in the culture and moderate nitrite

and faint ammonia reactions were given.
After 63 hours a moderate cloud, thick scum, and bubbles developed in the culture and

faint nitrite and faint ammonia reactions were given.
After 72 hours a slight cloud and thick scum, no nitrite or nitrate, and very faint

ammonia reaction.
Average temperature at which the culture was grown, 29° C.
Sample collected from a spot 6 miles south of Port Royal, wind ESE., force 4, high tide

(slack). Taken from surface. 1,000 c.c. Gran's medium was inoculated with

ID c.c. of the sample.
After 20 hours a slight cloud developed in the culture, no nitrite reaction was given.
After 27 hours a slight cloud developed in the culture, faint nitrite reaction was given.
After 38 hours a dense cloud developed in the culture, strong nitrite and faint ammonia

reactions were given.
After 48 hours a dense cloud and scum developed in the culture, strong nitrite and

faint ammonia reactions were given.
After 63 hours a dense cloud and scum developed in the culture, moderate nitrite and

faint ammonia reactions were given.
After 72 hours a moderate cloud, scum, and bubbles developed in the culture, very

slight nitrite and faint ammonia reactions were given.
After 86 hours a moderate cloud and scum, no nitrite or nitrate and very faint ammonia

reaction.
Average temperature at which the culture was grown, 29° C.
Sample taken from surface water of the large mangrove swamp lying northwest of Port

Henderson. 1,000 c.c. of Gran's medium inoculated with 10 c.c. of sample.
After 20 hours no cloud or nitrite reaction.
After 24 hours slight cloud and slight nitrite reaction.

After 40 hours strong cloud and scum, strong nitrite and slight ammonia reaction.
After 75 hours cloud, scum, and bubbles, no nitrite or nitrate and slight ammonia

reaction.
Average temperature at which the culture was kept, 30° C.
Subculture from culture (i). 1,000 c.c. Gran's medium inoculated with 5 c.c. of culture

(i), and kept at a temperature of 10° C. to 12° C. by means of ice.
After 100 hours the culture was quite clear and gave no nitrite reaction. It was then

removed from the ice and kept at the room temperature, which averaged 30° C.
After 107 hours a dense cloud developed in the culture and strong nitrite and faint

ammonia reactions were given.
After 120 hours a dense cloud, scum, and bubbles developed in the culture and moderate

nitrite and faint ammonia reactions were given.

24 Papers from the Marine Biological Laboratory at Tortugas.

After 131 hours a faint cloud, scum, and bubbles developed in the culture; very faint

nitrite and faint ammonia reactions were given.
After 146 hours a faint cloud and scum, no nitrite or nitrate, and very slight ammonia

reactions were given.

Twenty cultures were made from samples of water taken well out to sea
from Port Royal, and the process of denitrification followed through with
each. All gave very similar and consistent results, but the rate of denitri-
fication decreased rapidly with the temperature at which the cultures were
grown; thus at an average temperature of 27° C. the first trace of the nitrite
reaction appeared after about 40 hours, and denitrification was complete
after about 100 hours.

The results of precisely similar experiments that I made with samples of
water taken from the English Channel near Plymouth in the autumn of
1909 showed that there the process of denitrification was very much slower,
and was never complete at the room temperature (17° C). The first trace
of the formation of a nitrite in cultures in the modified Gran's medium, as
detected by the metaphenylene diamine reaction, occurred about the fifth
day, and a large proportion of the nitrite and nitrate always remained, even
in the oldest cultures. In similar cultures incubated at 30° C, denitrification
was complete by the eighth day at earliest, but uniformly consistent results
were not obtained, as in some of the cultures complete denitrification was
never obtained, even after several months.

It would thus appear that even under similar temperature conditions
the marine bacteria in the seas off Jamaica are much more active in causing
denitrification than those found in the English Channel, and since the rate
of denitrification is a function of the temperature, it follows all the more that
the destruction of nitrates by bacterial agency in the seas around Jamaica
must be far in excess of that occurring in the cooler waters of the English
Channel.

THE INVESTIGATION OF SAMPLES OF SEA-WATER TAKEN AROUND THE

DRY TORTUGAS.

The Dry Tortugas consist of a group of eight small keys, the largest of
which (Loggerhead Key) is only about three-quarters of a mile long by one-
eighth of a mile wide. They are situated about 150 miles from the main-
land of Florida, and form the extreme western end of the chain of the
Florida Keys. The lOO-fathom line lies some 30 miles to the south and
southwest of the islands, and then trends round in a northwest direction;
beyond the lOO-fathom line the depth increases with moderate rapidity
until depths of from 1,000 to 1,400 fathoms are reached. To the east,
northeast, and north as far as the coast of Florida, the water is shallow,
the soundings showing from 20 to 30 fathoms in most places. Beyond the
loo-fathom line to the southward the influence of the Gulf Stream begins
to make itself felt, though the region of maximum current velocity here lies
nearer the coast of Cuba. The Tortugas Keys are of purely coral forma-
tion, consisting entirely of broken shell and coral sand, with no soil. The

On the Precipitation of Calcium Carbonate. 25

greatest elevations are the hurricane ridges, which are not more than 15
feet above sea-level, and during a hurricane the islands are sometimes
completely submerged. There is no vegetation on the smaller keys, but
Loggerhead Key, on which the Marine Biological Laboratory of the Car-
negie Institution of Washington is situated, is partially covered with a
growth of bushes and coarse grass. There is no fresh-water supply on the
islands.

From these considerations it is obvious that the risk of contamination of
samples of sea-water taken a few miles from the Keys through land bacteria
is very small, and that such samples may be taken as being truly oceanic.

The motor yacht Anton Dohrn and smaller motor boats made the col-
lection of samples an easy matter, and the well-equipped laboratory made
possible fuller investigations than those attempted in Jamaica.

A number of cultures were made in Gran's medium under conditions
exactly comparable to those made at Port Royal, and the rate at which the
process of denitrification proceeded was observed. The results agreed
almost exactly with those obtained at Port Royal, so need not be described
in detail. It thus seems that the denitrifying power of the bacteria in the
seas around the Tortugas is the same as that of those around Jamaica.

Cultures were also made on various solid media, and pure cultures of
the various species of bacteria were isolated by plating in Petri dishes
with peptone agar. Samples of surface water taken from various positions
around Tortugas as far as possible removed from influence of the land, and
collected on sunny days, gave an average count of 14 colonies per i c.c.
of sample. Counts of several plates from the same locality, and from
different localities, showed a somewhat remarkable agreement as to the
number of colonies present, the highest count ever obtained being 20 and
the lowest 8 per i c.c. Allowing for experimental error, this shows great
uniformity in the distribution of bacteria in the sea around Tortugas.

The colonies appeared to be of two kinds when grown on peptone agar,
one much more plentiful than the other. Subcultures made from these
colonies in Gran's medium showed that the bacteria forming the most com-
mon type of colony produced an active denitrification, while the others
grew very slowly in this medium and produced no denitrification.

The characteristics of the denitrifying form are as follows :

On the potassium malate, or peptone agar media, colonies are visible as
minute white specks after 6 to 8 hours when the room temperature averages
29.5° C. After about 18 hours the colonies are well developed; they are
white in color, circular but with finely irregular outline, and have a granular
appearance. Superficial colonies are much elevated at first, but as growth
proceeds they spread rapidly over the surface of the agar. Deep colonies
remain small, circular, and discrete.

Growth is somewhat more rapid on peptone agar than on the potassium
malate agar, and the older colonies develop a brownish tinge in the center
when growing on the former medium. On gelatin peptone (5 per cent

26 Papers from the Marine Biological Laboratory at Tortugas.

peptone in sea-water and kept at between 20° and 25° C. to insure the
medium remaining solid), growth was very slow; in stab cultures growth
proceeded slowly from the surface downwards, leaving a funnel-shaped
depression of liquefied gelatin.

Acid formation occurs in dextrose, laevulose, mannite, and cane sugar,
but not in lactose media.

Growth is inhibited at a temperature of 10° C, but takes place slowly
at 15° C.

Growth is much retarded by exposure to bright sunlight, but the bacteria
are not killed by a 10 hours' exposure.

The bacteria are facultative anaerobes, but growth under anaerobic
conditions is very slow.

In Gran's medium growth is rapid, but no growth occurs if the potassium
nitrate be omitted, or if the calcium malate be replaced by calcium carbon-
ate. Growth in a pure solution of peptone in sea-water is very slight, but
becomes abundant if potassium nitrate be added, when denitrification
quickly ensues. The most rapid growth was produced in sea-water con-
taining 2 per cent peptone, i per cent potassium malate, and 0.5 per cent
potassium nitrate, and in this clear medium a slight floccular precipitate,
presumably of calcium salts derived from the sea-water, was soon formed.
Growth was also rapid at first in a solution of 5 per cent potassium malate
and 0.5 per cent potassium nitrate in sea-water, but growth apparently
ceased in this medium after a few days and denitrification was never com-
plete; a slight precipitation occurred and the solution was found to have very
definitely increased in alkalinity.

This bacterium does not appear to have been previously described, and
I propose for it the name of Bacterium calcis, owing to its power of pre-
cipitating calcium carbonate from solutions of calcium salts. This point
will be dealt with later in the paper.

The characteristics of the scarcer non-denitrifying form of bacterium
found on the agar plates are as follows:

Growth on the potassium malate agar medium is very slow and indefi-
nite. On peptone agar growth is somewhat slower than in the case of the
denitrifying form.. On the surface, circular cream-colored colonies are
formed, having a brownish center; the edges are smooth and regular, and
the colony remains discrete and does not tend to spread over the surface.
The deep colonies are smaller and usually ovoid in shape, and of a somewhat
darker color than those on the surface.

No growth was obtained on gelatin media.

Acid formation as shown by the neutral red reaction occurs in dextrose
and laevulose, but not in cane sugar, lactose, or mannite media.

Growth takes place slowly at 10° C. No visible growth occurred at 0°
C., but cultures were not killed by 24 hours' exposure to this temperature.

Growth is retarded by light, and cultures are killed by 4 hours' exposure
to bright sunlight.

The bacterium is a strict aerobe.

On the Precipitation of Calcium Carbonate.

27

Free growth takes place in Gran's medium, but develops much slower
than in the case of the denitrifying form. No growth occurs if the potassium
nitrate be omitted entirely, but takes place freely if a mere trace in excess
of that normally present in the sea-water be added, though no denitrification
results. Attempts were made to discover whether this bacterium had any
nitrifying or denitrifying action in various culture micdia, but uniformly
negative results were obtained. Nitrites were neither oxidized to nitrates
nor reduced to ammonia or free nitrogen, and ammonia salts were un-
affected. No growth was obtained in any culture medium that did not
contain at least a trace of nitrates, so it was not practicable to ascertain
whether the bacterium had a nitrifying action without the necessary facili-
ties for quantitative work.

On one occasion samples were obtained from various depths up to 90
fathoms at a point near the Gulf Stream region, 25 m.iles south of Tortugas.
Exhausted glass flasks with capillary necks which could be broken off at
the required depth were used for the purpose.

These samples were plated in the peptone agar medium and counted,
with the following average results:

Depth in Denitrifying Non-denitrifying
fathoms. °^^^ {Bacterium^ forms.
calcis).

No. of colonies
developing from
I c.c. of sample.

0
10
40
60
90

9

25

2
5
5

2
4
2
3
6

II

29

4

8

II

These figures are probably not very reliable, especially for the greater
depths, since it is possible that many of the bacteria were killed by the sud-
den reduction of pressure to which they were exposed as the water entered
the exhausted bulb.

INVESTIGATION OF SAMPLES FROM A POINT 70 MILES WEST OF
USHANT ISLAND. FRANCE.

This spot was chosen as it is sufficiently far out in the Atlantic to be
largely removed from the influence of the English Channel water. The
object was to investigate truly oceanic bacteria, and previous work in 1909
had shown that the bacterial flora of the Channel water was relatively very
complicated, probably owing to the presence of littoral forms. The Marine
Biological Association of the United Kingdom very kindly sent their
steamship Oithona from Plymouth for this work, and gave me every facility
both on board and in their laboratory. As in Tortugas, the deep samples
were collected in exhausted glass flasks, and accordingly, as previously ex-
plained, the results obtained from the deep samples can not be considered
to possess any very great degree of accuracy.

Attempts were made to plate the sample in peptone agar on board the
boat, but the result was not satisfactory, as owing to the motion of the boat

28

Papers from the Marine Biological Laboratory at Tortugas.

the jelly set in irregular waves and lumps. Consequently the samples were
kept on ice, and cultures were made from them at Plymouth 24 hours after
collection. It is clear that in future attempts to make plate cultures on
board a small boat in rough weather, a very delicately swung table will be
necessary, or else the roll-tube culture method must be employed.

Three plates on peptone agar were made from each sample, i c.c. of
the sample being used for each plate. The plates were kept at the room
temperature, averaging about 20° C, and the colonies were well developed
after 48 hours; they appeared to be all of one kind. A count gave the
following results :

Depth
in fathoms.

No. of colonies
developing from
I c.c. of sample.

0
10
20
30
SO
70
80

7
9
6
5
6
30
20

The increase in the number of colonies at 70 and 80 fathoms is somewhat
remarkable, but no conclusions in this respect can be drawn from one series
of observations.

The cultural characteristics of this bacterium are as follows :

On peptone agar, after about 36 hours at 20° C, the colonies are white
in color, circular, with a finely serrated outline and a coarsely granular
appearance. Superficial colonies grow very rapidly, and may spread as a
whitish semi-transparent growth of irregular shape over the surface of the
agar. The deep colonies remain small, globular, and discrete. In old agar
cultures a brownish tinge is developed, and the color may diffuse through
the substance of the agar. On gelatin peptone growth was rapid ; in stab
cultures growth proceeded from the surface downwards, leaving a funnel-
shaped depression of liquefied gelatin, and eventually all the gelatin became
liquefied.

Acid formation, as shown by the neutral-red reaction, took place in
dextrose, mannite, and Isevulose, but not in cane sugar or lactose media.

One thousand c.c. of Gran's medium, inoculated on board with 10 c.c.
of a surface sample immediately after collection, and kept at an average
temperature of about 20° C, showed the first trace of nitrite formation
after 70 hours. After 84 hours a very strong nitrite reaction was obtained,
and a slight ammonia reaction was given with Nessler's reagent. The proc-
ess of denitrification, even after the lapse of weeks, did not extend beyond
this, and no bubbles of gas were formed. Other experiments made with
subcultures from agar and gelatin media gave similar results, so that it
appears that this bacterium can not entirely break down nitrates at a
temperature of 20° C. The optimum temperature for denitrification pro-
duced by this bacterium appears to be about 20° C, as the process was less

On the Precipitation of Calcium Carbonate. 29

rapid at average temperatures of 17° C. and 25° C. At a temperature of
2,2° C, rapid growtli took place, but no denitrification resulted.

It should be noted that these temperature observations were only made
with subcultures from colonies on peptone agar and peptone gelatin media,
and there is reason to believe that the power of denitrification becomes
diminished after cultivation on such media. Further and more accurate
temperature experiments are required in which the culture medium is
directly inoculated with freshly collected samples of water.

This bacterium appears to be closely related to the Bacterium calcis,
its chief points of difference being :

1. Lesser denitrifying power and lower temperature optimum for denitri-

fication.

2. More rapid growth on gelatin media.

3. Absence of acid formation in media containing cane sugar.

INVESTIGATION OF SAMPLES FROM MARQUESAS KEYS. AND EXPERIMENTAL
PRECIPITATION OF CALCIUM CARBONATE BY BACTERIAL AGENCY.

The Marquesas Keys constitute a coral atoll which forms part of the
long chain of keys separating the Gulf of Mexico from the Straits of Florida.
Within the atoll the water is very shallow and the bottom consists of a
fine chalky mud many feet deep. Samples of the water from the lagoon
of the atoll were sent to me at Plymouth by post, and examined 14 days
after collection.

On plating on peptone agar, an average of 800 colonies per I c.c. of the
sample were obtained. These colonies appeared to be all of one species,
and were identical in appearance and in all cultural characteristics with the
Bacterium calcis previously described as occurring around the Tortugas.

A suspension of these bacteria from a culture on peptone agar was made
in sterile sea-water, and a similar suspension containing roughly the same
number of bacteria was made from a third subculture on peptone agar of
the bacteria obtained from the station 70 miles west of Ushant. i c.c. of
each of these suspensions was then added to 1,000 c.c. of the modified
Gran's medium. Some of these cultures were kept at an average tem-
perature of 20° C. and others at 32° C., with the following results:

At 20° C. cultures from Marquesas showed trace of nitrite after 45 hours.

At 20° C. cultures from Marquesas gave strong nitrite reaction after 53 hours.

At 20° C. cultures from 70 miles west of Ushant showed trace of nitrite after 140 hours.

At 20° C. cultures from 70 miles west of Ushant showed strong nitrite reaction after 162

hours.
In both cases a slight amount of ammonia was recognizable by Nessler's reagent

when the nitrite reaction was strong, but decomposition of the nitrite did

not proceed further even after 14 days.
At 32° C. cultures from the Marquesas showed trace of nitrite after 18 hours.
At 32° C. cultures from the Marquesas gave strong nitrite reaction after 22 hours.
At 32° C. cultures from 70 miles west of Ushant never gave nitrite or ammonia reaction.
The cultures from the Marquesas showed a slight amount of ammonia formation,

but the decomposition of the nitrite did not proceed further.

30 Papers from the Marine Biological Laboratory at Tortugas.

From these experiments it appears that the bacteria from subcultures
from the Marquesas have a much greater denitrifying power than those
from subcultures from a point 70 miles west of Ushant, and that as the
bacteria from the Marquesas appear to be of the same species as those
investigated at the Dry Tortugas, their power of causing complete denitri-
fication in Gran's medium has been lost by successive cultivations on
peptone agar.

The presence of the thick layers of fine chalky mud within the Marquesas
Keys, and elsewhere in many places near the Florida coast, led to a con-
sideration of the possibility of its precipitation by bacterial agency.

Since these bacteria grow freely in Gran's medium, the calcium salt of
a simple organic acid is a sufficient source of organic food for them, and it
seems probable that they would thrive in sea-water containing the products
of decomposing vegetable matter, provided that the nitrate supply and
conditions of light and temperature were suitable. Such conditions should
be especially well fulfilled by the drainage into the sea of a well-wooded
country with a calcareous subsoil, and the soluble organic calcium salts
would be precipitated as calcium carbonate by the action of the bacteria.
In addition, the elimination of the acid radicle from the nitrate in the process
of denitrification, by whatever stages it may occur, must leave the alkaline
base free to destroy the normal equilibrium of the salts in sea-water, and by
increasing the alkalinity would also result in the precipitation of calcium
carbonate.

To test this theory cultures were made in a medium having the following
composition:

Calcium succinate 2.5 grams

Potassium nitrate 0.5 gram

Sea-water 1,000.0 c.c.

Calcium succinate is soluble in these proportions and the medium is
quite clear. Free growth was manifested by the cloudiness of the medium
48 hours after inoculation, and nitrite formation was apparent.

After 96 hours the medium appeared quite milky and this milkiness was
due to the presence of exceedingly fine particles of a substance which was
soluble in dilute hydrochloric acid with evolution of gas, and was presumably
calcium carbonate. In some cultures these particles settled as a definite
sediment, but in others the particles were so minute that they showed little
tendency to settle and could only be separated with difficulty by centrifugal-
ization. The conditions determining the size of the particles formed could
not be ascertained, as the size varied in cultures which were apparently made
and grown under identical conditions.

The addition to cultures in which the particles of calcium carbonate were
so small as to remain in suspension of any foreign substance, such as finely
poAvdered calcium sulphate or of larger particles of sand, resulted in the
aggregation around them of the particles of calcium carbonate, forming a
concretion around a central nucleus. These concretions were hard and of

= lOO Fathom Line.

("hart Showing General Geographical Relations of the Western Bahama Island.

On the Precipitation of Calcium Carbonate. 31

almost crystalline appearance under the microscope, and were soluble in
dilute hydrochloric acid, with evolution of bubbles of a gas which, when the
operation was performed on a microscopic slide, could be completely ab-
sorbed by running in a solution of sodium hydrate under the cover slip.
Once this process of concretion has been initiated, it appears to progress
independently of the presence of particles which act as nuclei, and a large
concretion may often be found having a number of smaller concretions
around it, or continued into a chain of small spheres, the whole presenting
somewhat the arrangement shown by freely budding yeast cells. The
deposition of this form of calcium carbonate also takes place on the sides of
the flask, and more especially over any area where the glass is scratched or
roughened.

From these results it would seem strongly probable that the layers of
fine and unorganized chalky mud found in the Marquesas Keys are being
precipitated by the action of the Bacterium calcis, and it would seem a
reasonable suggestion that similar bacterial action may have played an
important part in the formation of chalk and other limestone formations
in geologic times. The formation of semi-crystalline concretions around a
central nucleus at first seemed to suggest an explanation of the formation
of oolite grains, but a mineralogical examination, very kindly made by Dr.
Fred. E. Wright, showed that the concretions did not possess that laminated
structure characteristic of oolite grains and that their crystalline structure
was nearer that of calcite than aragonite.

SOME CONSIDERATIONS ON THE PHYSIOGRAPHY OF THE TONGUE
OF THE OCEAN AND ANDROS ISLAND.

The position of the Tongue of the Ocean is shown in Chart A, which
includes the greater part of the Bahama group and shows its position
relative to Florida and Cuba. The Tongue consists of a long and narrow
stretch of deep water, running in a NNW.-SSE. direction, and except at
its northern end it is completely surrounded by shallows or land. On the
west, for about three-fourths of its length, it is bounded by the coast of
Andros Island ; southwest of Andros it is separated from Santaren Channel
by some 60 miles of shallow water lying over the Great Bahama Bank;
to the south it is separated from the Old Bahama Channel by over 50 miles
of shallows, averaging not more than 3 fathoms in depth; to the east it is
separated from the deep water of Exuma Sound by from 20 to 40 miles of
shallow water of from 2 to 3 fathoms in depth, and by the chain of islands
and cays extending in a NNW. direction from Great Exuma Island. The
mouth of the Tongue of the Ocean lies between New Providence Island
on the east and the northern extremity of Andros Island on the west ; it is
here some 25 miles wide and it maintains this width for the greater part
of its length as it stretches south. Between the southern extremity of
Andros and Green Cay (see Chart A) it narrows to under 20 miles, but
south of this point it expands eastward into an almost circular terminal
basin of about 35 miles diameter. The total length is about 120 miles.

32 Papers from the Marine Biological Laboratory at Tortugas.

To the north, the Tongue of the Ocean is connected by a stretch of deep
water, extending in a northeast direction, with the Providence NE. Channel
and the Providence NW. Channel, leading respectively into the Atlantic
and the Straits of Florida, and thus it is brought into direct connection with
the two main divisions of the Gulf Stream.

The soundings in this area are shown in Chart B. So far as observations
go there seems to be a slight but regular upward gradient from a depth of
1,084 fathoms at the mouth to 740 fathoms in the southern terminal basin.
Along the margins the gradient is extremely steep, and along the coast of
Andros our observations showed that it was almost perpendicular at a
distance of from one-fourth to one-half mile from the shore, but unfortu-
nately no complete observations have been made from which this gradient
could be calculated. When attempting to make soundings in this area, the
sinker (a 30-pound conical iron weight) in every case was caught upon
what was probably some projection on a submarine cliff, and it was im-
possible to free it; after several such attempts and loss of sinkers, the
soundings were abandoned. The entrance to the Tongue by the Providence
NE. Channel is of steadily increasing depth up to 2,200 fathoms at its
junction with the still deeper water of the Atlantic; the Providence NW.
Channel becomes shallower at its junction with the Straits of Florida, and
between Great Isaac and the western end of Bahama Island is only between
200 and 300 fathoms in depth.

The surface water of the Tongue of the Ocean, except along the coast of
Andros, is everywhere continuous with that overlying the thousands of
square miles of shallows forming the Great Bahama Bank and the flats
and cays lying to the north and west of the Exuma Islands, whereas the
deeper water is only connected with the outer ocean by the comparatively
narrow entrance between New Providence and the north of Andros, leading
after a turn of about 90° into the Providence NE. and NW. Channels.

The laboratory from which this work was done was established at
Golding Cay, at the eastern mouth of the South Bight of Andros; this
position was especially suitable for the work, as by running out a few miles
in a direction at right angles to the coast line the middle of the Tongue of
the Ocean could soon be reached, and also the mud flats to the west of the
island were readily accessible by water, as the South Bight runs right
through the island to the west coast. In this region the tides are not
strong, the average rise and fall being from 2 to 3 feet.

Much difficulty was experienced in getting any definite information as
to the set of currents in the Tongue. Our local pilot stated that a current
would set in a southerly direction for weeks at a time, and then without
any apparent reason or change of wind would reverse and set in a northerly
direction for several weeks; but such information derived from the negro
natives can not be relied on. When taking observations on May 8, May 11,
and May 23, we experienced a distinct southerly drift on each occasion,

100 Fathom Line.

Chart of Tongue of the Ocean and Adjacent Banks

On the Precipitation of Calcium Carbonate. 33

but the amount of this drift was not determined, and moreover the drift
caused by the wind was an unknown factor. On May 8 the wind was SSE.,
of force I at 8^30™ a.m. freshening to force 3 at 10^30™ a.m. A rough esti-
mate from the landfall on returning gave the drift of the boat as about 2 miles
south during the 4 hours occupied in working the station; the boat had a
large awning and exposed a considerable area to the wind and had drifted
this distance against the wind, so it would seem that on that occasion there
must have been a strong current setting south.

Andros Island consists of a limestone formation, the exact nature of
which has been dealt with by Dr. T. Wayland Vaughan (17). The greater
part of the island is very flat and is elevated only a few feet above sea-level ;
a few irregular undulations, probably not more than 100 feet high, are found,
especially along the east coast. There is evidence to show that formerly
the level of the land was much higher than at present, and signs of rapid
erosion of the rock are everywhere obvious. One of the most remarkable
features is the absence of soil, even in the well-wooded parts of the island,
the trees and bushes growing directly out of crevices and holes in the rock
and giving rise to practically no leaf-mold. In the numerous "pot-holes"
which occur all over the island, a small deposit of black leaf-mold can be
found, and these "pot-holes" are the favorite places for the cultivation of
sugar cane and bananas. The erosive action of water on the rock is espe-
cially noticeable where the slow drainage from an inland swamp can be
traced in its course to the sea; in such a locality the hard rock is eroded,
honeycombed, and undermined to a most remarkable degree, even though
the amount of drainage, except after the heaviest rains, can scarcely be
more than a slow trickle. Erosion of the rock along the sea-coast, where
it is exposed to the action of the sea-spray, is also very marked. From the
occurrence of this erosion it is obvious that all the water draining from the
land into the sea must contain a high proportion of calcium salts in solution.

Towards the west of the island the land is remarkably flat, and near the
coast consists of white chalky mud which has partially dried and in places
has formed a harder crust on the surface. These half-dried mud flats slope
almost imperceptibly into the sea and are continuous with the submarine
flats which extend some 60 miles off the west coast with an average depth
of 2 to 3 fathoms. The mud forming these submerged flats is very soft,
and near the coast it was easily possible to push a 12-foot sponge-pole down
to its full length into it without touching any harder material. The surface
layer of the mud for a depth of about 6 inches is of a creamy white color,
but below that it is of a grayish tinge and has a slight odor of sulphureted
hydrogen. Unfortunately there was no opportunity of obtaining informa-
tion as to the real thickness of this layer of mud, nor of investigating more
than the surface layers at any distance from the coast.

Microscopical examination showed that this mud was almost entirely
composed of minute unorganized particles of calcium carbonate. Near the

3

34 Papers from the Marine Biological Laboratory at Tortugas.

shore a good deal of organic matter was present, chiefly in the form of de-
caying mangrove roots. Farther out from the shore little organic matter
was noticeable, but it was not possible to examine the deeper layers of the
mud in these situations. The only organic matter that was seen consisted
of the rootlets of a species of Zostera, found in occasional patches some miles
off the coast.

BACTERIAL INVESTIGATIONS IN THE DEEP WATER OF
THE TONGUE OF THE OCEAN.

Continued bad weather during the whole of our stay at Andros greatly
added to the difficulties of this work, and on this account it was only found
possible to work three stations. The last two were worked under the most
disadvantageous conditions, the quick roll of the boat making the filling
of the water-bottle with alcohol and the siphoning off of the sample under
sterile conditions a matter of the greatest difficulty.

The first station worked was 6 miles due east of Golding Cay, the second
14 miles east of Golding Cay, and the third 10 miles ENE.3^E. of Golding
Cay. The three stations were thus at the angles of a triangle which was
nearly equilateral, the base being a little longer than the sides and running
due east and west.

At the first station, worked on May 8, bottom was sounded at 822
fathoms. The sea was calm at first, with a SSE. swell, but became choppy
later. The wind was SSE., force o to i, at 8^30™ a.m., freshening to about
force 3 at io''30™ a.m. The sample of the bottom obtained by the snapper-
rod was of a very stiff clay-like consistency, grayish white in color, and was
composed of very minute unorganized particles of calcium carbonate
containing a few pteropod and globigerina shells. The following tempera-
tures, to which the necessary corrections have been applied, were recorded:

Depth.

Temperature. Depth.

■
Temperature.

Fms.

Surface

10

"C. Fms.
26.90 1! 400

25.90 [! ^

25.14 , 000

22.00

17.13 Bottom (822 fms.)

0/7

{8:93}average = 8.98

{4.62}^^^''*^^ =4.70
{1.94}^^^'^^^ "3.97

SO

100

200

't

These samples, without previous dilution, were plated in peptone agar,
I c.c. of the sample being used for each plate. The agar was cooled to just
under 40° C. before plating. It is very necessary that this temperature
should not be exceeded, as many marine bacteria are very sensitive to heat;
the use of agar at as high a temperature as 45° C. will cause the death of a
large proportion of the bacteria, though in the process of plating they can
be exposed to this temperature only for a very short time. The cultures
were kept in the dark at the room temperature (averaging about 28° C.)
and at the end of 24 hours a free growth of colonies was apparent. At the
end of 48 hours the plates were counted with the following results:

On the Precipitation of Calcium Carbonate.

35

From these counts it is apparent that the number of bacteria falls off at
some point between 200 and 400 fathoms.

The second station was worked on May 1 1 at a point 14 miles due east
of Golding Cay. The sea was calm at first, and the wind ENE., force i,
but later in the day a heavy swell set in and the wind freshened to about
force 4; eventually the weather became so bad that it was impossible to
work, and the station had to be abandoned before it was completed.
Bottom was sounded at 890 fathoms, but there was some stray on the wire,
so that the true depth was probably about 825 fathoms, as shown by the
chart. The bottom consisted of fine white calcareous ooze; no remains of
pteropods were seen, but some globigerina shells were present. The follow-
ing temperatures were recorded :

Depth. Temperature.

Depth.

Temperature.

Fms. ■ OC.
Surface ' 2630 *

Fms.

300

f 1^27 I^^^^S^ =^4.32

400 1 ^;^^ javerage = 9-79

200 1742 j

Bottom < 4.17 \ airoraao = ,1 tc

( 4.13 J ° ^"^ 1

Samples down to a depth of 200 fathoms were diluted i in 100 with
sterilized sea-water before plating. The following results were obtained
after 48 hours' growth:

Depth.

No. of colonies

developing

from I CO. of

sample.

Fmi.

Surface | 16,200

10 13,100

50 14,000

100 14,000

200 15,000

300 14

400 12

The figures given represent the mean of the number of colonies develop-
ing in the two plates that were made from each sample. It is here apparent
that the number of bacteria per i c.c. falls off very rapidly between 200 and
300 fathoms.

36

Papers from the Marine Biological Laboratory at Tortugas.

The third station was worked on May 23 at a point 10 miles ENE.3^E.
of Golding Cay. The wind was east, of about force 4. As it had been
blowing for the previous ten days without intermission, the sea was so rough
that it was only possible to work when steaming slowly ahead into the wind.
This resulted in the production of a great deal of stray on the sounding-
wire, so that the number of fathoms of wire run out is greater than the
actual depth at which the samples were taken; these differences will be
large for the more superficial samples, but small at greater depths, as the
wire strays in a curve whose gradient becomes very steep a little below the
surface under these conditions.

The following temperatures were recorded :

Length of wire run out. Temperature.

Length of wire
run out.

Temperature.

Fms. j °C.

Surface ' 27.10

20 26.50

Fms.
250

350

"C.

{is:oo}a^e'^ee=^4-98

{ 10:84 ^^^'•^s«=^°-8s

100 .... 22 80

160 18.83

At this point the station had to be abandoned, owing to the bad weather.
The samples down to i6o fathoms were diluted i in loo with sterilized sea-
water before plating in peptone agar; the remaining two were plated undi-
luted. At the end of 48 hours the following counts were made, representing
the mean of the number of colonies in the two plates made from each
sample :

. ^, r • No. of colonies

Length of wire developingfrom

run out. jc.c. of sample.

Fms.

Surface is.ooo

20 I 15. 500

100 13,700

160 13.300

250 14.300

350 16

The colonies developing in all the cultures were only of two kinds, the
Bacterium calcis and the non-denitrifying species already described. The
non-denitrifying species formed a relatively small proportion of the total,
and were not found at all in cultures made from samples taken below 250
fathoms. As they appear to be comparatively inactive chemically, and
as nothing is at present known concerning the part played by them in the
metabolism of the sea, they will not be further considered here.

A consideration of these results obtained in the Tongue of the Ocean
shows that the waters down to a depth of somewhere about 300 fathoms
in April 1912 contained an enormously larger number of bacteria than the
water in the neighborhood of Tortugas in June 191 1. The number of
bacteria falls off from about 14,000 to about 12 per i c.c. between depths of
250 and 350 fathoms; the temperature at 250 fathoms was about 15° C.
and at 350 fathoms about 11° C, and it was shown in June 191 1 at Tortugas

On the Precipitation of Calcium Carbonate.

37

that B. calcis will grow slowly at 15° C, but that growth is totally inhibited
at 10° C. It would thus seem that the observed distribution of the bacteria
agrees fairly with what might be expected from the temperature conditions.

As regards these observations on the occurrence of bacteria in small
numbers at depths below 350 fathoms, the possibility of experimental error
must be considered; a leakage into the water-bottle of 0.25 c.c, as it was
being hauled up through the last 300 fathoms, would account for the number
found, and there are also many possible sources of error in the process of
siphoning ofif the sample and making the cultures where a permanent labor-
atory is not available. It is possible that the water below 350 fathoms was
really sterile, though if so the constancy of the results obtained is curious if
ascribed to experimental error. In any case, the small number of bacteria
found at depths below 350 fathoms can play no part in the metabolism of
the sea, since it has been shown that B. calcis is incapable of growth at the
temperatures obtaining at these depths.

The much greater abundance of bacteria in the surface waters of the
Tongue of the Ocean than in the waters around Tortugas may perhaps be
accounted for by the fact that in the Tongue of the Ocean by far the greater
part of the surface water must flow over the immense chalky mud fiats and
shallows which bound it in most directions and, as will presently be shown,
these mud flats are phenomenally rich in bacteria and are probably still
being deposited by bacterial agency.

HYDROGRAPHIC OBSERVATIONS IN THE TONGUE OF THE OCEAN.

The samples of water taken for hydrographic investigations were
analyzed by Mr. D. J. Matthews at Plymouth. With great kindness he
calculated the results, and from his notes the following observations and
conclusions are drawn.

The samples were analyzed for salinity in comparison with the standard
sea-water supplied by the Central Laboratory of the Conseil International
pour I'Exploration de la Mer, and hence the results are strictly comparable
with all other analyses published under the auspices of this International
Council. The following results were obtained:

At first station,

6 miles eas

t of Golding Cay.

Depth.

Temp.

Cl 0/00

1 S<"m

1 O"0 1 <^t

Fms.

°C.

P.ct.

, P.ct.

0

26.90

20.06

36.24

29.12 23.70

10

25.90

20.46

36.96

29.71 24.57

50

25.14 ; 20.43

36.91

29.66 24.76

100

22.00 20.28

36.64

29-45 25.48

200

17.13 20.08

36.27

29-15 26.48

400

8.98 ' 19.58

35.37

28.43 27.43

600

4.70 19.49

35-21

28.30 27.90

822

3.97 19.367

34-98

1 28.11 27.79

Note. — The original surface sample was lost, owing to breakage of the bottle in transit to England. The
analysis was made on a sample taken 3 days later at the same spot.

In this table Cl " 00 means the weight of chlorine in grams found in 1,000 grams of sea -water. S '00 means
the salinity, or total weight of salt in grams found in 1,000 grams of sea-water, aa represents the specific
gravity of the sample at 0° C, and at represents the specific gravity of the sample at the temperature I at which
it was collected, with no correction for pressure.

38 Papers from the Marine Biological Laboratory at Tortugas.

At the second station, 13.5 miles east of Golding Cay, the following
results were obtained :

Depth.

Temp.

CI 0/00

So/w

o-o

fTt

Fms.

°C.

P.ct.

P.ct.

0

26.30

20.2s

36.S8

29.40

24-15

10

26.40

20.33

36.73

29.S2

24.24

so

24.89

20.39s

36.84

29.61

24.78

100

22.63

20.36

36.78

29.56

25.42

17.42

20.12

36.35

29.21

26.47

14.32

19.81

3S.79

28.76

26.7s

400

9-79

19.56

35.34

28.40

27.27

890

4.IS

(No

sample, bottle did not work.)

NAUTICAL MILES
5 10 15

Fig. 2. — Section across the Tongue of the Ocean from Golding
Cay, Andros Island, showing salinities and temperatures.

Owing to the uncertainty of the depths at the third station, due to the
bad weather and consequent stray on the wire, it was decided not to include

On the Precipitation of Calcium Carbonate.

39

these observations in a consideration of the hydrographic conditions, and
to make what deductions were possible from the results obtained at the
two stations given. , . .

Figure 2 shows the vertical distribution of layers of different salmities
at the two stations in diagrammatic form.

5

'

10

°

15°

2C

25

tT

u

_.

^-

lUU

^

^

^'

—

idUU

X

3UU

^

/,

-^

4-UU

/

/

/ .'

7.

•7

buu

/

/ .

bOO

i

700

1
1,

1;

BOC
qnr

—
> —

1;
li
1;

Fig. 3. — ^Temperature curves.
Continuous line = Station i.
Broken line = Station 2.
Dotted line =Cfca//en2er Station No. 27.;

The curves in figure 3 show the vertical distribution of temperatures at
the two stations, and also give the temperatures obtained by Challenger
Station No. 27 in 22° 49' N., 65° 19' W. March 28, 1873, where the depth
was 2,960 fathoms. It is interesting to note that the actual reading at
200 fathoms at the latter station, 17.22° C, agrees more closely with the
temperatures in the Tongue of the Ocean than that taken from the smoothed
curve, which was 18.17° C.

The curves in figure 4 show the vertical distribution of salinities at the
two stations, and also the salinities obtained by the Michael Sars in 37° 12'
N., 48° 30' W., on June 25, 1910.

From these tables and diagrams it can be seen that the surface salinity
increases from west to east very rapidly, 0.34 per cent in 7.5 miles, but the
surface temperature is fairly uniform, between 26° C. and 27° C.

At both stations the salinity increases downwards to a maximum prob-
ably lying between 10 fathoms and 50 fathoms, but more rapidly at Station
I, so that from 10 fathoms to 50 fathoms the salinity decreases from west
to east.

40

Papers from the Marine Biological Laboratory at Tortugas.

Below 100 fathoms the conditions are closely similar at both stations,
as far as the observations go; the salinity decreases fairly rapidly to 400
fathoms and then more slowly to the bottom.

The temperatures decrease rapidly and uniformly from the surface to
about 500 fathoms, then more slowly to the bottom.

35.0

36.0

37

/

/

N

i?

/

/i

'' /

200

300

/

/

/ /

/
/

/

,■'//

/

f

1^ 400
0

r

H

< 500

u.

//

]

/

600
700
800

;'

\

'\

1

Fig. 4. — Salinity curves.
Continuous line = Station i.
Broken line = Station 2.
Dotted line =Michael Sars Station No. 65.

There is practically no thermocline (German "sprungschicht") in the
temperature at any depth and only a poorly marked one, confined to the
upper stratum, in salinity.

The absence of a temperature thermocline is remarkable (see the curves
for the two stations, and the two from the Michael Sars and the German
South Pole Expedition, drawn for comparison).

In general, below about 250 fathoms the temperatures and salinities
agree with the nearest stations of the Michael Sars in the open Atlantic;

On the Precipitation oj Calcium Carbonate. 41

above this depth they are higher, and differ from them and the open ocean
north of the belt of calms, in the absence of a temperature thermocHne, and
in the maximum salinity being found below the surface. The latter points
either to a considerable local supply of fresh water or to a current of lower
salinity from either the Florida stream or the region of equatorial calms.
Unfortunately we have no reliable salinity observations for the two latter.
With regard to the accuracy of the work, Mr. Matthews makes the
following remarks:

THE ACCURACY OF THE OBSERVATIONS.

a. Salinity. — The method of taking the samples from the water-bottle was rather incon-

venient, as a siphon was used; the samples were very small, but well pre-
served. The water-bottle itself might have leaked or closed at the wrong
depth, as was the case with earlier models.^ That this was not so is shown by
(i) The sharp fall in the number of bacteria at between 200 and 300 fathoms.

(2) The close agreement of the salinities at 400 fathoms, the greatest depth at which

they were taken on both stations. Station I gave 35.37; Station II, 35.34.

(3) The close agreement between the bottom salinity at Station I, 34.98 at 822 fathoms,

and the salinity found at the same depth at the nearest position at which we
have modern observations — i. e., Michael Sars Station 65, in 37° 12' N., 48°
30' W., June 1910; according to the curve this is about 34.96.

b. Temperatures. — The National Physical Laboratory correction was given to 0.1° only,

but the readings below 15° are comparable among themselves to 0.05° or
possibly less. The curv^es of temperature for the two stations agree well in
shape below 300 fathoms, but the temperature on Station II is generally
slightly higher than on Station I, as a rule by an amount corresponding to a
difference of depth of about 20 to 25 fathoms.

Below 200 fathoms the curves for both stations agree ver>' closely with
that for Michael Sars Station 64, in 34° 44' N., 47° 52' W.

It is almost certain from the above considerations that the results are
only incorrect by the experimental errors in measuring the depth, in deter-
mining the salinity (0.01° at most), and by perhaps 0.1° C. of temperature.

These observations are sufficient to show that the Tongue of the Ocean
is an area of considerable interest from a hydrographic point of view, and
it is much to be regretted that the continued bad weather during our stay
made it impossible to obtain more observations and samples.

BACTERIAL INVESTIGATION OF THE CHALKY MUD FLATS WHICH ARE BEING
DEPOSITED TO THE WEST OF ANDROS ISLAND.

Samples of the mud were taken from the western entrance of South
Bight (see Chart B), and from points 2 and 3 miles out from the shore;
practically identical results were obtained from all these localities. The
sample at the mouth of the bight was taken in about 4 feet of water, that 2
miles out in 7 feet, and that 3 miles out in 8 feet. The samples were neces-
sarily taken from the surface of the mud.

I The water-botUe only failed once, at about 890 fathoms at Station II.

42 Papers from the Marine Biological Laboratory at Tortugas.

For bacterial examination, one part of this mud was shaken up with
three parts of sterilized sea-water; this was allowed to settle for 15 minutes,
and then the clearer surface layer was diluted i in i ,000,000 with sterilized
sea-water. The diluted fluid was plated in peptone agar, i c.c. being used
for each of the plates. The count of a number of plates after 48 hours gave
40 colonies as an average, and thus the mud itself must contain 40 X 4 X
1,000,000 = 160,000,000 bacteria per i c.c. The actual number in the
mud possibly exceeds this figure, since a large proportion of the bacteria
would probably settle with the larger particles after the first dilution.

The bacteria found in these cultures were nearly all the B. calcis; only
occasionally were a few colonies of the non-denitrifying species seen.

A sample of the water taken from the surface at a spot 3 miles out from
the western entrance of South Bight gave a count of 35,000 colonies per
I c.c, the great majority of these being B. calcis.

Subcultures of B. calcis were made in Gran's medium, and in the calcium
succinate, calcium acetate, and peptone calcium acetate media, whose
composition has already been given. Denitrification in all these media
was rapid and eventually complete, and was accompanied by the precipi-
tation of calcium carbonate. In the last three media, which contained no
solid matter and were quite clear and transparent before inoculation, this
precipitation was manifested after 12 hours by the formation of a thick
white cloud in the fluid, readily distinguishable from the cloudiness pro-
duced merely by bacterial growth. The development of this precipitate
continued rapidly during the first 48 hours, but in many cases it was com-
posed of such fine particles that they showed little tendency to settle to
the bottom of the flask; in other cases larger particles were formed and a
sediment similar in appearance to the chalky mud of the mud flats was
produced. The exact conditions determining the size of the particles
precipitated could not be ascertained, as the size varied largely in cultures
made at the same time, in the same media, and kept apparently under the
same conditions. The addition of magnesium tartrate in small quantities
(0.2 gram per 1,000 c.c.) to the culture media seemed to induce the precipi-
tation of larger particles, but it did not appreciably affect the rate of growth
of the bacteria. In some of the older cultures that had been kept for a
week or more, the sides of the flasks were coated with a thin layer consisting
of extremely minute rhombohedral crystals of calcium carbonate. Occa-
sionally these crystals formed around small bubbles that had remained near
the surface of the fluid, the weight of the crystals eventually caused the
bubbles to sink, and then the contained gas became dissolved ; in this way
a number of small, hollow spheres were formed, their walls consisting of
minute crystals of calcium carbonate. The formation of these curious
bodies occurred especially readily in the calcium succinate medium to which
0.2 gm. of magnesium tartrate per liter had been added. The deposition
of calcium carbonate in a distinctly crystalline form was only noted in old
cultures, and then it was in an amount relatively extremely small when

On the Precipitation of Calcium Carbonate. 43

compared to the precipitate of unorganized and amorphous calcium car-
bonate.

Specimens of the precipitates from some of the culture media were sent
to Dr. Fred. E. Wright, of the Geophysical Laboratory of the Carnegie
Institution of Washington, who described them as follows:

Preparation I: Precipitate from medium^composed of calcium acetate 5.0 grams,
potassium nitrate 0.5 gram, peptone (Witte's) 0.2 gram, sea-water 1,000.0 c.c, filtered
and sterilized, contains two substances: (i) Fine grains of a strongly birefracting, appar-
ently uniaxial, optically negative substance and with refractive index about 1.66. This is
probably calcite. The grains are isolated, and no evidence of spherulitic crystallization
was observed. On treatment with very dilute hydrochloric acid a noticeable evolution of
carbon dioxide took place. (2) Scattered through the preparation are fine needles of a
weakly birefracting substance of refractive index of about 1.525; extinction angle large.
These needles are evidently selenite (hydrated calcium sulphate).

Preparation II: Precipitate from medium composed of calcium succinate 2.0 grams,
magnesium tartrate 0.2 gram, potassium nitrate 0.5 gram, sea-water 1,000.0 c.c, consists
largely of a cryptocrystalline aggregate of a weakly birefracting substance, whose refractive
index is about 1.52 to 1.53. This substance proved too fine for further determination.
Scattered through this substance are rounded and irregular patches of a second cr>'pto-
crystalline substance of strong birefringence, which gives off CO2 when treated with dilute
hydrochloric acid, and is probably calcite.

Preparation III: Precipitate from a medium composed of calcium acetate 5.0 grams,
sodium phosphate (Na2HP04i2H20) 0.25 gram, potassium nitrate 0.5 gram, sea-water
1,000.0 c.c, is again very fine-grained and consists (i) in large measure of minute grains of
a substance which agrees with calcite in its optical properties in so far as they could be
determined. On immersion in dilute HCl a distinct evolution of CO2 gas was observed.
(2) Of a substance whose grains are somewhat coarser than the calcite grains, their bire-
fringence being medium to weak; refractive index about 1.525; biaxial and apparently
optically positive; probably selenite, but not crystallized in the usual manner.

The small quantity of hydrated calcium sulphate present in these
precipitates is undoubtedly derived from that in solution in the sea-water
with which the media were made up, but the reason of its precipitation is
difficult to explain, since no such precipitation occurred in culture media
kept uninoculated under similar conditions as control experiments. It
would therefore appear that this deposition of calcium sulphate along with
the calcium carbonate must in some indirect way be the result of bacterial
action, and it would seem a possible suggestion that the odor of sulphureted
hydrogen noticeable in the deeper layers of the mud flats might be due to
the reduction of the calcium sulphate to a sulphide, and subsequent decom-
position of the sulphide.

These observations have shown that on the chalky mud fiats of the
Great Bahama Bank the B. calcis is found in enormous numbers, and also
that this bacterium is capable of precipitating calcium carbonate from
fluid media containing soluble calcium salts. It would seem a fair deduction
that these mud flats, consisting of fine unorganized particles of calcium
carbonate, have been precipitated by the action of the B. calcis on the
soluble calcium salts carried into the sea by drainage from the land, where
extensive and rapid weathering of the limestone rock is in progress.

44 Papers from the Marine Biological Laboratory at Tortugas.

CONCLUSION.

The observations so far available are too few, and the area they cover
too small, to attempt to make any broad generalization at present. How-
ever, it can be stated with a fair degree of certainty that the very extensive
chalky mud flats forming the Great Bahama Bank and those which are
found in places in the neighborhood of the Florida Keys are now being
precipitated by the action of the BacteriiLm calcis on the calcium salts
present in solution in sea-water. From this the suggestion is obvious that
the Bacterium calcis, or other bacteria having a similar action, may have
been an important factor in the formation of various chalk strata, in addi-
tion to the part played by the shells of foraminifera and other organisms in
the formation of these rocks. Dr. T. Wayland Vaughan has also suggested
that the Miami oolite and other oolitic rocks may owe their origin to the
occurrence of some diagenic change in the precipitate of very finely divided
particles of calcium carbonate produced in this way by bacterial action.
If this view as to the formation of chalk and oolite rocks is correct, it would
seem probable that these strata must have been deposited in comparatively
shallow seas whose temperature approximated to that of tropical seas at
the present time.

It has also been shown that bacterial denitrification is far more rapid
and complete in the tropical seas around Jamaica, Tortugas, and Andres
than in the temperate waters of the Bay of Biscay and the English Channel,
and hence an explanation is provided of the relative scarcity of plankton
and algal growth in the former localities, in accordance with the terms of
Brandt's hypothesis.

The distribution of the bacteria, both as to numbers and species, has
been shown to vary at different localities and at dififerent depths, but there
are at present too few observations to enable any conclusions or generaliza-
tions to be drawn.

As it now stands, the investigation can at most be considered to offer
a mere indication of the part played by bacterial growth in the metabolism
of the sea. To obtain a real insight into the question, it would be necessary
to make more extensive bacterial and chemical observations in tropical,
temperate, and arctic waters, to study the bacteriology of other areas where
calcium carbonate is being precipitated from the sea, and to make further
investigations in the laboratory into the chemistry of the reactions that can
be brought about by various species of marine bacteria.

On the Precipitation oj Calcium Carbonate. 45

LITERATUFIE.

1. Baur, E. 1901. Ueber zwei denitrifizierende Bakterien aus der Ostsee. Wiss.

Meeresunters., vol. vi. Kiel.

2. Brandt, K. 1905. On the production and the conditions of production in the sea.

Conseil Internat. Rapp. et Proc. V^erb., vol. ill.

3. 1904. Ueber die Bedeutung der Stickstoffverbindungen fiir die Produktion

im Meere. Botan. Centralblatt, vol. xvi.

4. Drew, G. Harold. 191 i. Report of preliminary investigations on the marine deni-

trifying bacteria. Year Book No. 10, Carnegie Institution of Washing-
ton.

5. 191 1. The action of some denitrifying bacteria in tropical and temperate

seas, and the bacterial precipitation of calcium carbonate in the sea.
Jour. Marine Biolog. Assoc, of the U. K., vol. ix, No. 2.

6. 1912. Report of investigations on marine bacteria carried on at Andros

Island, Bahamas, British West Indies, in April 1912. Year Book No. 11,
Carnegie Institution of Washington.

7. Feitel, R, 1903. Beitrage zur Kenntnis denitrifizierender Meeresbakterien. Wiss.

Meeresunters., vol. vil. Kiel.

8. Fischer, B. 1894. Die Bakterien des Meeres. German Plankton Expedition.

9. Gran, H. H. 1901. Studien iiber Meeresbakterien, I. Bergens Museums, Nr. 10.

Aarbog.

10. Keding, M. 1906. Weitere Untersuchungen uber stickstoffbindende Bakterien.

Wiss. Meeresunters., vol. ix. Kiel.

11. Kentner, J. 1905. Ueber das Vorkommen und Verbreitung stickstoflfbindende

Bakterien im Meere. Wiss. Meeresunters., vol. viii. Kiel.

12. Matthews, D. J. 1913. Deep-sea Bacteriological Water-bottle. Jour. Marine

Biolog. Assoc, of the U. K., vol. ix, No. 4.

13. Murray, J., and J. Hjort. 1912. The depths of the ocean. London,

14. Murray, J., and R. Irvine. 1889. On coral reefs and other carbonate of lime

formations in modern seas, Proc. Royal Soc. of Edinburgh, vol. xvii.

15. Raben, E. 1905. Ueber quantitative Bestimmung von Stickstoffverbindungen im

Meerwasser. Wiss. Meeresunters., vol. viii. Kiel.

16. Thomsen, p. 1910. Ueber das Vorkommen von Nitrobakterien im Meere. Wiss.
^H Meeresunters., vol. XI. Kiel.

17. Vaughan, T. Wayland. 1912. Carnegie Institution of Washington, Year Book

No. II, pp. 153-162,

PRELIMINARY REMARKS ON THE GEOLOGY OF THE

BAHAMAS, WITH SPECIAL REFERENCE TO THE

ORIGIN OF THE BAHAMAN AND

FLORIDIAN OOLITES.

BY THOMAS WAYLAND VAUGHAN,
Geologist in charge of Coastal Plain Investigations, U. S. Geological Survey.

47

PRELIMINARY REMARKS ON THE GEOLOGY OF THE BAHAMAS,

WITH SPECIAL REFERENCE TO THE ORIGIN OF THE

BAHAMAN AND FLORIDIAN OOLITES.^

By Thomas Wayland Vaughan.

At Doctor Mayer's request the following preliminary statement on
certain aspects of Bahaman geology and on the origin of the Bahaman and
Floridian oolites is submitted as a companion paper to that of Mr. Drew,
whose untimely death is both a source of great sorrow to his friends and a
heavy loss to science. As Drew's investigations have proven an essential
step in the solution of the problem of the origin of the extensive oolitic lime-
stone formations of Florida and the Bahamas, it is fitting that an account of
the processes by which they originate should accompany his posthumous
memoir. Although the investigation of the rock specimens and bottom
samples obtained in the Bahamas and Florida from April to July 191 2 is
not complete, certain definite and some tentative conclusions have been
reached which may here be appropriately stated.

The Bahama Islands and their accompanying shoals occupy a triangular
area which lies east of Florida and north of the islands of Cuba and Haiti.
The northwest corner of the triangle is in latitude 27° 35' N., longitude
79° 2' W., about 40 miles north of Jupiter Inlet on the Florida coast. The
southeast corner is Turks Islands, north of Haiti, in latitude 21° 20' N.
and longitude 71° 5' W. The western side is a nearly north-and-south line
from Little Bahama Bank to about latitude 25° 30', whence the boundary
bends into a south-of-east direction. The islands either occur on one of
two large banks, the Little Bahama and the Great Bahama Banks, or they
rise to the southeast of the latter bank as isolated eminences separated by
water as much as 1,000 fathoms in depth.

The shoalest water in the bottom of New Providence Channel between
Little and Great Bahama Banks is 253 fathoms in depth; that in the Straits
of Florida, between the former bank and the Florida coast, off Jupiter Inlet,
is 341 fathoms deep; between the latter bank and Florida the depth is 422
fathoms. The Great Bahama Bank is separated from Cuba by the Old
Bahama Channel, which ranges from 276 to 296 fathoms in depth. This
bank is indented by two bodies of deep water. The first of these, theTongue

1 The investigations, the results of which are here presented, were conducted under the joint auspices of
the Department of Marine Biology of the Carnegie Institution of Washington and the United States Geological
Survey. The former organization furnished opportunities for field work, while the latter authorized the prose-
cution of the investigation as a part of my official duties, and furnished faciUties for laboratorj' studies in
Washington.

4 49

50 Papers from the Marine Biological Laboratory at Tortugas.

of the Ocean, is sack-shaped and lies east of Andros Island and west of the
bank, on the northern end of which is New Providence Island. Its depth
ranges from 795 to 1,200 fathoms. The second indentation, Exuma Sound,
is eastward from the Tongue of the Ocean, across the bank and keys ex-
tending southward from New Providence Island. This sound is purse-
shaped and exceeds i ,000 fathoms in depth. Water i ,000 fathoms in depth
is close to the eastern shore of the Bahamas as far north as Elbow Cay
on Little Bahama Bank. Eastward of the i ,000-fathom curve the bottom
rapidly descends to a depth between 2,000 and 3,000 fathoms. The Bahama
Islands are subaerial protuberants above the nearly level, slightly sub-
merged surfaces of extensive plateaus which on one or more sides rise pre-
cipitously from oceanic depths.

The water from Gun Key to Northwest Passage, a distance of 67 knots,
is only 7 to 12 feet in depth, while the bottom southeast of New Providence
Island ranges from awash to 12 or 18 feet below water-level. New Provi-
dence Island is mostly a platform from sea-level to 20 feet in elevation, with
several east and west ridges standing above it; the highest of these, Nassau
Ridge, rises to an altitude of about 100 feet. Andros Island has a similar
physiography, a nearly level platform ranging in height from sea-level to
20 feet above that datum plane. Along its eastern front, lying near the
shore, is a series of interrupted hills that range in elevation from 20 to 60,
and perhaps in some instances to 100, feet. West of the coastal hills the
surface is low, usually less than 10 feet in elevation, while much of the areas
along the west front is affected by the tides. The Barrier Reef of Andros
stands on the seaward edge of a shallow, submarine platform, at a distance
of one-half mile to 2 miles off shore. The slope of the bottom seaward of
the reef is comparatively rapid ; between the reef and the shore the depth
ranges from i to 2 fathoms. The platform between the reef and the shore
is composed of hard lime rock, the surface of which is indented by numerous
submarine pits. The deepest of the holes, locally known as "blue holes,"
sounded was 36 feet deep. As the general aspect of these holes is similar
to the abundant subaerial pot-holes, locally known as "banana holes,"
formed by solution of the limestone, they are evidence of subsidence.

The general country rock of the Bahamas is oolite. The sea bottom
from Gun Key to Northwest Passage, New Providence Island, and Andros
Island is all or mostly composed of oolite, precisely similar to that of
southern Florida. The material forming the ridge at Nassau has been wind-
blown, and evidently is a dune accumulation. Many or all of the ridges
rising higher than 20 or 30 feet above sea-level have been formed of wind-
blown material, of which oolite is an important constituent, but the oolite
of most or all the lower areas is of marine origin and has not been subjected
to aeolian action.

The finely divided calcium carbonate oozes in the shoal water? of Florida
and the Bahamas have long attracted attention. Louis Agassiz, as early
as 1851, noted them, as did Captain Hunt, Alexander Agassiz, and others.

Preliminary Remarks on the Geology of the Bahamas. 5 1

Dall, so far as I am aware, was the first to suggest that "much of the limy
deposit of the area behind the reefs and defended by them is probably the
result of the deposition of lime originally in solution and precipitated
by chemical action rather than of mechanically transported sediment."^
Later, Samuel Sanford became convinced that the flocculent oozes in some
of the Florida bays and sounds were chemically precipitated, and so ex-
pressed himself.2 The most careful study as yet made of the shoal-water
bottom deposits of south Florida was the one initiated by me in 1908 and
conducted in conjunction with George C. Matson.^

This investigation led to the conviction that the fiocculent, so-called
amorphous, calcium carbonate is neither of detrital nor of organic origin, and
therefore must be a chemical precipitate. In 191 1 , G. H. Drew came to the
Tortugas Laboratory- for the purpose of studying denitrifying bacteria. As
he discovered that ammonia was a product of the metabolism of these organ-
isms he was led to conduct experiments to ascertain the effect they might
have in precipitating calcium carbonate. They were found to be active in
Florida; and investigations he subsequently made in the Bahamas showed
160,000,000 of these bacteria to the cubic centimeter of surface mud on the
west side of Andros Island. Drew says:^

A consideration of these observations shows firstly that B. calcis is found in enormous
numbers in the chalky mud fiats of the Great Bahama Bank, and secondly that this bac-
terium is capable of precipitating calcium carbonate from fluid media containing soluble
calcium salts. It would seem a fair deduction that these mud fiats have been precipitated
by the action of B. calcis on the soluble calcium salts carried into the sea by drainage from
the land, where extensive and rapid weathering of the limestone rock is in progress.

Drew has shown what we may be confident is the major agency in pro-
ducing the precipitation of the calcium carbonate ooze.

Every geologist who has studied the Florida oolites, except Alexander
x\gassiz, has considered them marine deposits; and Louis Agassiz, Sanford,
and myself considered them associated with or derived from the fine cal-
careous muds. My opinion^ was more definitely expressed than those of the
other investigators.

Rainey, as early as 1858, showed, in a remarkable book entitled "On
the mode of formation of shells of animals, of bone, and of several other
structures, by a process of molecular coalescence, demonstrated in certain
artificially formed products," that calcium carbonate precipitated from
water-soluble salts of calcium in a viscid solution form spherulites and not
crystals. In 1871, Harting, in a memoir, "Recherches de morphologic
synth6tique sur la production artificielle de quelques formations calcaires
organiques,"^ showed that calcium carbonate precipitated from a solution of
a water-soluble calcium salt by the addition of the carbonate of potassium

> U. S. Geological Survey Bull. No. 84, p. loi. 1892.

s Second Annual Report Florida Geological Survey, p. 228. 1910. e ^^r„^ui„^^„

« A contribution to the geologic history of the Floridian Plateau. Camegie Institution of Washington
Pub. No. 133. pp. 114-145. 1910.

* Camegie Institution of Washington, \ear Book No. 11, p. I44. 1912.
5 Carnegie Institution of Washington Pub. No. 133, PP- I73-I77. 1910.
sverh. Kon. Akad. Wetensk., Amsterdam, vol. 13. PP. 84, 4 Pls- 1871.

52 Papers from the Marine Biological Laboratory at Tortugas.

or sodium is at first of gelatinous consistency and later aggregates into
spherulites.

In 1903, G. Linck published, in an article entitled "Die Bildung der
Oolithe und Rogensteine,"^ accounts of a series of important experiments
he conducted on the precipitation of calcium carbonate from sea-water by
the addition of alkaline carbonates, and in 1909 another contribution from
him, "Ueber die Bildung der Oolithe und Rogensteine, " appeared in the
Jenaische Zeitschrift fur Wissenschaft, vol. 45, pp. 267-278. According
to Linck's experiments calcium carbonate precipitated from sea-water at
a temperature of 17° C. to 40° C, by adding an alkaline carbonate, is
aragonite and aggregates into spherulites, the diameters of which range
from 0.02 to 0.2 mm. Murray and Irvine have shown that calcium car-
bonate precipitated from sea-water by an alkaline carbonate at a tempera-
ture of 34° F. is calcite, at 47° F. is a mixture of calcite and aragonite, at
80° and over is aragonite.-

Drew in 1912 published the following statement regarding the calcium
carbonate precipitated^ in a culture from a sample of water from Marquesas
Lagoon :

To such a culture a trace of very finely powdered hydrated calcium sulphate was added.
This resulted in the formation of a precipitate, which, on microscopical examination, could
be seen to consist of finely laminated concretions, some of which appeared to have a particle
of calcium sulphate or sand as a nucleus. The concretions were soluble in dilute hydro-
chloric acid with evolution of carbon dioxide. These concretions bear a resemblance to
those of some oolitic limestones, and the experiment suggests the manner in which such
oolites may have been formed.

These references, although they lay no claim to completeness, are clear
evidence that calcium carbonate, particularly aragonite, when precipitated
from a water-soluble calcium salt by the addition of an alkaline carbonate,
is at first of gelatinous consistency and later aggregates into spherulites.^

As I did not observe oolite grains in the muds when collected in the
field, I was led to the opinion that oolitization was the result of secondary
changes after precipitation and directed my attention to ascertaining if
such supposed changes were taking place. ^

Bahaman shoal-water bottom muds were collected through South Bight
and off the west shore of Andros; and other samples were collected in
Florida. As has been stated, these flocculent oozes when collected were not
observed to contain oolite grains; howe\-er, such grains may have been
present and may have escaped attention. All the Bahaman muds when
examined at the end of November did contain oolite. Besides oolite grains,
finely divided particles less than o.ooi mm. in diameter, some fragmental

> Neues Jahrb. fur Min. Beilage. Bd. l6, pp. 49S-SI3-

2 Proc. Roy. Sci. Edin., vol. 17, pp. 79-109. 1890.

' Carnegie Institution of Washington, Year Book No. 10, p. 141. 1911.

* Without making a special diversion to go into the subject, it may be mentioned that spheruhtes or oohtes
of calcium malate and calcium phosphate are well known, and that the occurrence of spherulites of barium car-
bonate and strontium carbonate have been recorded. It is therefore established that a number of salts of the
calcium group of elements may assume a spherulitic form.

'" Carnegie Institution of Washington, Year Book No. 11, pp. IS5, I37, 158. 1913.

Preliminary Remarks on the Geology of the Bahamas. 53

calcium carbonate, alcyonarian spicules, shoal-water foraminifera, and other
micro-organisms were present.^

Dr. Fred. Eugene Wright has examined some of the bottom samples for
me and reports that the particles large enough for ascertaining their optical
properties are calcite, but the fine material can not be optically determined.
A series of the muds and several powdered specimens of Pleistocene oolites
from Florida and the Bahamas were tested with the cobalt nitrate reaction;
and as all of these upon healing gave the purple color characteristic of
aragonite, this form of calcium carbonate was shown also to be present.
The muds and the Pleistocene oolites, therefore, are composed of a mixture
of aragonite and calcite.

The grains of the Bahama and Florida oolites range in size from o.io
to 0.80 mm., and occasionally a grain perhaps exceeds i mm. in diameter.
The spherulites or spheroid aggregates in the muds range from 0.004 or
0.006 mm. to oolite grains of ordinary size.

In order to test the growth of the grains, samples of a number of muds
were strained through No. 10 bolting-cloth, which has a mesh of about
0.13 mm. in diameter, and the fine material that passed through the bolting-
cloth was put into bottles containing sea-water. These bottles were per-
mitted to stand from about November 25 until the first week in March,
over three months, when a small portion of each sample was again strained
through No. 10 bolting-cloth, and the portion retained on the cloth studied.
The formation of oolite grains was found to be in progress in every sample,
and numerous grains had apparently grown to such a size as to preclude
their passing through the mesh of bolting-cloth. The grains showed the
usual forms of oolite grains: spheroids, ovoids, and ellipsoids. The larger
grains had smaller diameters of 0.17 mm.; longer diameters up to 0.23 mm.
Those newly formed are soft and easily crushed by any kind of pressure.
The experiments indicate increase both in number and in size of grains.

The precipitated calcium carbonate may segregate around a variety of
nuclei, for instance, spherulites or round aggregates formed of the precipi-
tated material, small grains of sands, shells of foraminifera, and gas bubbles.^

Although there is need for additional study of the factors that accelerate,
retard, or inhibit the formation of spherulites and initial round aggregates
and the subsequent growth of the grains, the empirical facts in the process
of the formation of the Floridian and Bahaman oolites are clear. They
are as follows:

1. Denitrifying bacteria are very active in the shoal waters of both
regions and are precipitating enormous quantities of calcium carbonate
which is largely aragonite.

2. This chemically precipitated calcium carbonate may form spherulites
or small balls which by accretion may become oolite grains of the usual
size, or it may accumulate around a variety of nuclei to build such grains.

1 The investigation of these muds is still in progress [June 1913I and will not be finished for some months,
The completed results will appear in a subsequent publication.

2 See the papers by Linck, Drew, and Vaughan, already cited.

54 Papers from the Marine Biological Laboratory at Tortugas.

The deduction may be made that all marine oolites, originally composed
of calcium carbonate, of whatever geologic age, may confidently be attrib-
uted to this process.

Two other important deductions may be made from the knowledge of
this process, viz:

1. Neither the Bahamas nor the oolitic keys of southern Florida are
coral islands, but they have been formed by this other process. Elevated
coral rock is exceedingly scarce in the Bahamas and the Recent reef of
Andros is comparatively insignificant as a constructional geologic agent.
The material composing the land masses and much or most of the submarine
platforms of the Bahamas is thus removed from the category of "coral rock"
and the living reef reduced to a subordinate ratio as a builder of limestone.

2. Drew's studies (unfortunately incomplete) of the distribution of
denitrifying bacteria have shoM^n them to be most prevalent in the shoal
waters of the tropics. They therefore conform to the temperature relations
enunciated by Murray^ for the distribution of lime-secreting organisms. By
combining the results of Drew and Murray, the deduction seems warranted
that great limestone formations, whether they be composed of organic or of
chemically precipitated calcium carbonate, were laid down in waters of
which at least the surface temperatures were warm, if not actually tropical.

This brief paper may be closed with the following statement and tenta-
tive outline of the geologic history of the Bahamas, with special reference to
the eastern part of Andros Island :

The presence of oolite containing Pleistocene marine fossils above sea-
level indicates elevation in the Bahamas; the presence of submerged solution
pot-holes, not filled with sediment, indicates that the last movement was
downward.

TENTATIVE OUTLINE OF THE PLEISTOCENE AND RECENT GEOLOGIC HISTORY
OF THE BAHAMAS, AND ESPECIALLY OF ANDROS ISLAND.

(i) During Pleistocene, perhaps earlier Pleistocene time, numerous shal-
low submarine banks existed, on which calcium carbonate was chemically
precipitated mostly as aragonite, and this was converted into oolite.

(2) This submergence was followed by uplift with the production by
marine erosion of a platform and a margining shore cliff, and the formation
of shore dunes paralleling the windward sea-front.

(3) Apparently there wa^ further uplift and pot-holes were subaerially
formed by solution on the platform between the dunes and the sea-front.
At this time the land on the east side of Andros Island stood about 40 feet
higher than now.

(4) The last event was a subsidence of about 40 feet, admitting the sea
back to the edge of the previously cut low scarp, and the present barrier
reef developed on the off-shore margin of the submerged platform.

1 Natural Science, vol. 2, pp. 25-27. 1897.

IV.

THE BUILDING OF THE MARQUESAS AND TORTUGAS

ATOLLS AND A SKETCH OF THE GEOLOGIC

HISTORY OF THE FLORIDA REEF TRACT.

By THOMAS WAYLAND VAUGHAN,
Geologist in cliarge of Coastal Plain Investigations, U. S. Geological Survey,

55

IV.

THE BUILDING OF THE MARQUESAS AND TORTUGAS ATOLLS

AND A SKETCH OF THE GEOLOGIC HISTORY OF THE

FLORIDA REEF TRACT.

By Thomas Waylaxd Vaughak.

INTRODUCTION.

The following brief article is supplemen'ar>' to my paper "A contri-
bution to the geologic history of the Floridian Plateau,"^ in which it is
stated that "the geologic history of the Marquesas and Tortugas is reserved
for future consideration." The study of the geology and the geologic
processes of the Florida reef tract, and especially of the Tortugas and the
Marquesas, has been continued since 1910. During the field season of
1913 Mr. R. B. Dole, of the U. S. Geological Survey, undertook a special
investigation of the solvent effect, by virtue of the content of carbon dioxide,
of sea-water flowing into and out of Tortugas Lagoon. As the various lines
of investigation have given definite results, the principal factors in the
formation of the Marquesas and Tortugas atolls and their inclosed lagoons
now seem clear, although some details still require elucidation. It is now
also possible to outline with a feeling of confidence the salient geologic
episodes in the history of the entire Florida reef tract and to institute
comparisons with other coral-reef areas. As this paper is only a preliminary
statement, it is intended subsequently to publish a more comprehensive
discussion of the different problems here treated in a cursory manner.

BUILDING OF THE MARQUESAS AND TORTUGAS ATOLLS.
An inspection of the charts of the U. S. Coast and Geodetic Survey,
No. 1252 for the Marquesas and No. 471a for the Tortugas, will show a
striking similarity in the configuration of both, for in each banks or keys
surround a central depression and in each there are major entrances in the
southeast, southwest, and northwest quadrants. The Tortugas, however,
are of a more elliptical form than the Marquesas, the axis of the ellipse extend-
ing from northeast to southwest. The Marquesas lagoon is only i to 15
feet in depth, while that of the Tortugas is from 7 to 13 fathoms. In both
the principal key or bank slightly under water extends from the southeast

1 Camegie Institution of Washington, Pub. 133. PP- 99-i8s. IS pIs. 1910.

57

58 Papers from the Marine Biological Laboratory at Tortugas.

to the northwest entrance, or is on the east, northeast, and north sides.
The keys and banks between the southeast and southwest passages and
between the southwest and northwest passages are minor as compared with
the larger one mentioned. It may also be pointed out that the Marquesas
occur just west of Boca Grande Channel, and thac the Tortugas lie just
west of Rebecca Channel.

That the Marquesas were formed through the agency of waves and
currents operating under proper conditions has been evident for some time,
but the processes by which the Tortugas were built were less clear. It was
thought possible that submarine solution had been efficacious in the latter,
while it was apparent that it v/as ineffective in the former. In order to
answer the question for the Tortugas, two lines of inquiry were undertaken.
One of these was the examination of the bottom of the Tortugas Lagoon,
to discover whether or not sediment is there accumulating; the other was to
ascertain by chemical examination whether or not there was any excess of
carbon dioxide in the water flowing into the lagoon and whether there was
any difference in the quantity of carbonates in the outflowing and inflowing
water. The latter investigation was undertaken by Mr. R. B. Dole.
The bottom samples collected at numerous localities in the Tortugas
established that the bottom in the channel between Garden Key and Bush
Key, in Bird Key Harbor, and in the lagoon channels west and east of White
Shoal and north of that shoal is formed of soft calcareous mud, largely of
the kind Drew showed to be precipitated by denitrifying bacteria. The
examination showed conclusively that sedimentation predominates over
the removal of material, perhaps except where currents may corrade the
sides or bottoms of the channels along which they flow. The conditions
in the Tortugas lagoon are similar to those in the bays and sounds behind
the Florida keys and to those in Marquesas Lagoon, in all of which filling
by deposition is progressing.

The specimens examined by Mr. Dole were collected twice a day on the
principal flowing and ebbing tides for nearly a lunar period, from May 20
to June 15, near buoy No. C3, in Southwest Channel, ofT the southern end
of White Shoal. Mr. Dole's results are summarized on page 74 of this
publication. They show that none of the water, neither that flowing into
nor that flowing out of the lagoon, contains any dissolved free carbon
dioxide, that all the carbon dioxide in the water is present in the carbonate
or bicarbonate form, and that the inflowing water possesses no power to
dissolve calcium carbonate by virtue of its content of carbon dioxide, as
its capacity for the solution of lime has been exhausted before flowing into
the lagoon. The chemical investigations of the waters and the study of
the bottom deposits gave mutually confirmatory results, that the water
flowing through the Tortugas Lagoon does not dissolve calcium carbonate
and that the lagoon has not been formed by solution.

As two kinds of phenomena observed in the Tortugas might be con-
sidered to afford evidence in favor of solution, both will be briefly discussed.

Building of the Marquesas and Tortugas Atolls, etc. 59

One is the frequent occurrence of submerged coral heads, the interiors of
which ha^•e been removed, leaving a partially inclosing skeletal shell covered
by living polyps. Some of these heads have a suggestive resemblance to
atolls, but the resemblance is only superficial. The excavation of the
interior is to be explained by the upper surface of the envelope of living
tissue having been damaged in some way, thus giving an opportunity for
boring organisms of great variety to begin and prosecute their work of
disintegrating the skeleton. Waves and currents wash out the broken-up
material, forming a cavity which is bounded by a shell of substance whose
outer surface is covered by living polyps. The other phenomena that
might be attributed to solution are the overhanging projections along the
steep, submarine cliffs or rock pinnacles which in places occur on the sides
of channels. Since these stand against the sides of the currents it is probable
that submarine scour, corrasion, has been a factor in producing them;
and subaerial modification deserves consideration, as will later appear.
As the dynamic agencies have been more and more thoroughly studied, the
necessity for referring any phenomena observed in the Tortugas to sub-
marine solution has been eliminated, and the conclusion definitely reached
that such solution produces no geologic effects of importance.

As the production of the Marquesas and Tortugas atolls with their
inclosed lagoons by solution is eliminated, other factors, especially waves
and currents, will be considered. There are two current-shaped ph\sio-
graphic forms to which it is desired particularly to call attention — these
are current-shaped crescents and current-shaped linear ridges. With
reference to crescents it may be said that if an obstruction lies across the
line of direction of a constant current, the current shears to each side
and will drift detrital material in such directions as to form a crescentic
accumulation, the bow of the crescent facing the current, while the horns
of the crescent will curve before the current. An eddy may be produced
in the space between the horns of the crescent, and a deposit may be formed
there.^

Crescentic sand-dunes and crescentic keys are both well known. Should
there be no cross obstruction, currents both aerial and aqueous will form
linear ridges, under conditions which induce dropping material. Both
linear sand-dunes and linear submarine banks are well known. Of current-
transported detritus, crescentic accumulations lie across the main direction
of the current, while linear ridges lie along its direction.

The currents of the Florida keys are of three kinds : wind-formed currents,
which are accompanied by waves; the Florida counter current; and tidal
currents. The winds at Key West, where accurate records have been
kept, prevailingly range in direction from northeast to southeast, those
from the east predominating over those from any other direction. Both
the Marquesas and the Tortugas lie within the tract of the Florida counter
•current, which moves in general toward the west. The directions of the

1 Dr. A. G. Mayer called my attention to the occurrence of this eddy.

6o Papers from the Marine Biological Laboratory at Tortugas.

wind-formed currents and the counter current are concordant and they
therefore cooperate. The tidal impulse, however, is roughly from north
to south and the direction of tidal currents is therefore transverse to the
other currents.

The principal arcs of the rims of both the Marquesas and the Tortugas,
i. e., those from the southeast to the northwest entrance, are bowed toward
the east, against the prevailing direction of the wind and that of the counter
current, while the northern limb of the arc in each instance trails with
these currents. These arcs therefore satisfy the requirements of wave and
current-shaped arcs.

The southern sides of both the Marquesas and Tortugas are broken by
southeast and southwest passages. These breaks are probably to be
attributed to cross-tidal currents. Also both the Marquesas and Tortugas
have a northwest passage. The explanation of this is probably as follows:
In the Tortugas ; although the data on currents are inexact, the available
evidence indicates a stronger tidal inflow than outflow, much of the water
of the receding tides not passing through the lagoon but around the outside
of its rim. The water coming into the lagoon would tend to be driven
toward the west by the winds, and would consequently find its outlet in
the northwest quadrant.

In the Marquesas, the land area west of the southeast passage follows a
line that would be expected of the southeast horn of a crescent formed
under conditions of wind and currents known to prevail there. The
formation of the bank and the small keys on it on the west side of the
lagoon is more obscure. Perhaps they were formed by winds and accom-
panying waves that come, especially during hurricanes, from the west and
by a current eddy produced to the leeward of the bow of the crescent.

The main agencies whereby the principal arc of the rim of the Tortugas
was formed seem definitely known, but those building other parts of the
rim are still somewhat vague. Between what is known as Five-foot Channel
and Southwest Passage are holes, some of which are 6 fathoms or more in
depth. It is evident from the distribution of these holes that once there
were several channels trending more or less from southwest to northeast
and approximately following the course of tidal flow, the channels bounded
by banks elongate in the same direction. Detritus carried westward by
waves and currents has filled parts of the channels and broken them up
into holes, some series of which may still be recognized.

Loggerhead Key, as well as the bank on which it stands, is elongate
from southwest to northeast and has each of its ends somewhat curved
toward the west. The general direction of the bank is approximately along
that of tidal flow, but it is not an unmiodified linear ridge coinciding in
direction with that of one current, for it is evident that the prevailing
run of the waves before the wind, and perhaps the counter current acting
at its southern end, have curved its ends westward. The effect of waves
running before the wind in shaping a key was studied in detail on Logger-

Building of the Marquesas and Tortugas Atolls, etc. 6l

head Key, where both the north and south horns of the key curve before
the prevaiUng easterly winds. A reversal of a part of the horn of the
north end of the key was observed during a swell from the northwest.
During this reversal of the direction of the run of the waves, a reversed
crescent was formed at the end of the spit, the bow of the crescent being
toward the current while the horns curved before it. The more indurated
beach rock on each side of Loggerhead Key slopes toward the sea, showing
that the key is wave-built, on the surface at least. In this connection it
may be said that a current too weak to pick up and carry detritus will,
when agitation has lifted particles above the bottom, appreciably move
them in the direction of the current. Loggerhead Key stands on the lee
side of the lagoon, where waves from the east would be weaker than on the
windward side, and where waves from the west would exert their full force.
The following is ofTered as an hypothesis for the formation of the bank on
which this key stands: It is largely or mostly a detrital ridge, formed in a
partial eddy on the leeward side of the lagoon, mainly along the direction of
flow of the stronger tidal currents. Wave agitation from the east and west
has facilitated the movements of detritus and has built the sloping beach
rock on both the lagoon and the sea sides. The stronger winds from the
east have given a slight westward curvature to the extremicies of the bank.
Before leaving the discussion of Loggerhead Bank and Key, attention
will be directed to changes that have occurred since October iQio, v.hen the
last severe hurricane visited this region. Since that time the waves have
acted with only slight cessation on the eastern side of the key and have rut
away the eastern side of its northern end, while building has taken place
on the western side; thus the outline of the northern end of the key has
shifted westward. Accurate measurements of some of these changes have
been kept but will not be given here.

Besides the banks and keys forming the perimeter of the Tortugas,
there are also shoals and some small keys in the lagoon. Some of the
shoals, White Shoal for example, lie along the direction of flow of the major
tidal currents, although with their ends slightly curving toward the wQat;
others are isolated rocks. That some of the latter may be coral masses is
suggested by an experience on Garden Key, where all large rocks examined
proved to be dead coral heads.

The rocks composing both Marquesas and Tortugas are known only
from superficial examination. The foundation of the Marquesas may be
oolite, similar to that found on Boca Grande Key, as some of the bottom
samples from them were found to contain hard oolite grains. The keys,
however, are composed of calcareous detritus, of which Halimeda is an im-
portant constituent. There is no important coral growth around the
Marquesas. The keys of the Tortugas are mostly composed of the com-
minuted remains of a considerable variety of organisms that secrete calcium
carbonate, mollusca, corals, nullipores, echinoids, etc. The corals here, in
contradistinction to the Marquesas, are important. Coral reefs, patches,

62 Papers from the Marine Biological Laboratory at Tortugas.

or heads occur along the west side of Loggerhead bank throughout its
entire length and around both its southern and northern ends. Bush Key
Reef, southwest of Southeast Passage, is luxuriant, and outer reef patches
occur northward of Southeast Passage. Reef corals also occur at many
places along the sides of channels within the lagoon. Other species of corals,
not so large and massive as those on the reefs proper, live on the shallow
flats, where not overwhelmed with detritus. Although it has not been
practicable to make for the Tortugas a quantitative estimate of the pro-
portions of material contributed by the different kinds of organisms, corals
are rather surely the most important. The Marquesas, therefore, are
composed of calcareous detritus mostly of other than coral origin; while
the Tortugas are composed of calcareous material, of complex origin but
largely due to the activity of corals.

The foregoing remarks may be summarized as follows: (i) Submarine
solution has not been instrumental in the formation of the Marquesas or
the Tortugas lagoon; (2) the atoll rings are constructional phenomena and
were shaped by prevailing currents, those caused by winds, the Florida
counter current, and tidal currents. The detrital material on which theses
agencies have worked in the Marquesas is mostly not of coral origin; while
in the Tortugas, although of complex composition,^ it is largely composed of
coral debris.

The relative ages of the Marquesas and Tortugas atolls and their
relations to oscillation of water-level will now be briefly considered, and in
this connection further attention will be given to the probable relations of
corals to the formation of the latter atoll. The Marquesas rim is composed
of unconsolidated detrital material and evidently requires no oscillation
of water-level for its formation. No indication of change of level during or
since the building of the rim around the lagoon was observed.

Such a structure as the Tortugas would scarcely be built in the Florida
region through material derived from reef corals without change in depth,
as the depth of water in the lagoon, 13 fathoms, and outside the atoll rim,
13 fathoms or more, is greater than that in which the Florida reef species
are known to flourish .2 Other evidence, to be adduced in giving an account
of the coastal oscillacion, renders it reasonably certain that the Tortugas
were initially outlined during subsidence." Certain facts indicate that the
history of the Tortugas is more complicated than that of the Marquesas.

I The following literature is cited in tiiis connection:

Vauglian Cornish. On the formation of sand dunes. Geog. Jour. vol. 9, PP. 278-309. 1897.
C. Hedley and T. Griffith-Taylor. Corals of the Great Barrier, Queensland, a study of their structure,
life-distribution, and relation to mainland physiography. Austral. Assoc. Adv. Sci. Adelaide meeting, 17 PP-.

^ ^ ^' F.^Wood-Jones. Coral and atolls. London. Lovell Reeve & Co. Ltd. (especially chapters xii and xiii,

Tames JBryce. ' South America observations and impressions. New York, Macmillan Co. 1912. JPP;
58, 59 give a good description of crescentic sand-dunes, "the convex of the crescent always facing the wina,
on the great inner plateau of Peru.] . > .. u -a „„•• ^,<. „=ori

» In the foUowmg discussion of oscillation the terms uplift, 'depression, and subsidence are usea
with reference to land, but it is recognized that oscillation may be due to negative or positive movement ot
sea-level independent of crustal movement. , ,. .• <• Co,„,,ai

»The wells bored on Key Vaca by the Florida East Coast Railway, under the supervision of bamuel
Sanford, showed the thickness of the elevated coral reef rock to be 105 feet (= 17J fathoms), indicating tnat
the Pleistocene barrier reef was formed under conditions of subsidence, similar to those postulated lor tne
initial building of the Tortugas rim.

Building of the Marquesas and Tortugas Atolls, etc. 63

Most or all of the material above sea-level in the Tortugas is unconsolidated,
except an outer crust of some of the beach rock, as observed on Loggerhead
Key. Extending beneath the sea-level, however, to a depth of 8 feet or
more, is beach rock that appears more indurated. Harder rock was also
observed at a depth of 3 or 4 feet under water on Southwest and Northeast
keys, the subaerial portions of both have been cut away by wave action,
and in places aubmarine cliffs of hard rock, which has been undercut and
overhangs the sides of the channels, are evident at depths of (jver 20 feet.
The presence of an older formation for the living reef corals is clear. It
seems probable that this rock was subaerially indurated and then depressed.
The probability of this subsidence will be further discussed in a succeeding
paragraph, wherein the later movements of the Florida coast line will be
more fully considered. It appears that the Tortugas atoll with its inclosed
lagoon was outlined previous to the present development of the coral reefs
of that area, that it was elevated about 50 feet, and that there has been a
subsequent (the latest) depression; the land after the oscillation perhaps
lacked 10 or 15 feet of being depressed to its former level, thus leaving the
keys standing somewhat higher than before.

It has already been stated that the skeletal remains of corals are not
only an important constituent of the Tortugas rocks, but are rather surely
the most important contributor to its composition. Whether or not the
various keys, banks, and isolated rocks were outlined by coral reefs may not
be ascertained, but as coral reefs obey the laws of current-moved detritus,
in many instances being the principal single source of the detritus, a positive
answer to this question is not essential for the validity of the explanation
here presented for the building of the Tortugas. Hedley and Taylor say
in their paper on the Great Barrier, referred to in the footnote on page 62 :

The growth of an individual reef is shown to proceed in a regular cycle. If the reef
reaches the surface with its axis along the wind, then its shape endures; but if across the
wind, then its extremities are produced backward, forming first a crescent, later a horseshoe,
and lastly an oval, thus inclosing a lagoon.

SKETCH OF THE GEOLOGIC HISTORY OF THE FLORIDA REEF TRACT.
The probability of elevation and subsidence in the Tortugas area leads
to a consideration of evidence from which the later movements of the
Florida coast line have been inferred and of conditions antecedent to the
development of the living Florida barrier reef. That the last important
movement of the Florida east coast was downward from Jacksonville to
San Augustine is attested by the submerged channel of the mouth of St.
Johns River and the submarine fresh-water springs off San Augustine.
Late depression of the west coast is positively shown not only by the forms
of Tampa Bay and Charlotte Harbor, but also by submerged channels off
shore in that region. As the channels off this coast, however, appear not
to affect the lo-fathom curve the depression was probably not so great as
60 feet. The very ragged character of the coast line from Cape Romano to

64 Papers from the Marine Biological Laboratory at Toriiigas.

the head of Florida Bay, with its salients, reentrants, and almost numberless
small islands lying near shore, as Ten Thousand Islands, separated by an
interlocking maze of channels, suggests that the southern end of the penin-
sula to its very tip has undergone subsidence at no distant date. On keys
west of Bahia Honda Sanford noted holes, which, at a distance of a quarter
of a mile or more from shore, connect with underground passages containing
salt water; and on more easterly keys formed of elevated coral reef rock he
observed free openings or cavities that extend to a depth as great as 30 feet
below sea-level. The evidence is clear that the keys participated in the
uplift and subsequent depression that affected the mainland and that at
one time they stood more than 30 feet higher with reference to sea-level
than they now do. This uplift and the subsequent depression, according
to all available evidence, extended to the Tortugas.

The highest points to which the elevated reef extends are on Windleys
Island and Plantation Key, where elevations of 18 feet are attained. Sup-
posing these points to have been at sea-level, an uplift with reference to
sea-level, of somewhat more than 48 feet, must have taken place since the
elevated reef was formed.

The evidence presented shows that the platform, on the outer edge of
which the present barrier reef of Florida is principally growing, has geo-
logically, just antecedent to its present relations to sea-level, stood 30 feet
or more higher. It is on this shelf, the last oscillatory movement of which
was one of depression, that the living reefs of the Florida barrier have
established themselves and have grown. However, although the last earth
movement has been downward, the net result of the last oscillation, i. e.,
uplift followed by subsidence, has been to bring the surface of the sub-
marine shelf nearer to the surface of the sea. Although the building of the
platform seaward of the Pleistocene barrier reef needs further study, it is
known that the base of the Pleistocene reef is about 100 feet below sea-
level. As the reef would not have begun to grow in a marked depression,
it is probable that the water seaward of it was not shallower than that at
its base. The maximum depth of Hawk Channel is about 7 fathoms, while
in places the living barrier reaches the surface of the sea. Therefore, since
the initiation of the Pleistocene reef the bottom seaward of it has been
built up to a thickness ranging from 70 to 100 feet. As the question, how
much of this material may be Pleistocene, can not now be answered, it is
not known whether or not the living reef has Pleistocene corals beneath it.
However these questions may ultimately be answered, it is clear that both
addition of material (upbuilding) and uplift have contributed to raising
the surface of the basement of the Recent reef, notwithstanding that the
last oscillatory movement carried the land surface downward. The hydro-
graphic charts (see particularly U. S. Coast and Geodetic Chart No. 165,
Hillsboro Inlet and Fowey Rocks) clearly show the northward extension
of the reef platform beyond the limits of the living reefs, which are
merely superimposed on a platform whose existence antedates their own.

Building of the Marquesas and Tortugas Atolls, etc. 65

The conditions under which the Pleistocene barrier reef was formed
may be reconstructed by using the results obtained from the recent geologic
investigations on the Florida mainland and the records of the wells bored
by the Florida East Coast Railway on Key Vaca. Pliocene deposition
was followed by uplift, the ensuing Pleistocene deposits being laid down
on the eroded surface of those preceding. The Pleistocene reef, which is
105 feet thick, was therefore formed on a subsiding sea-bottom succeeding
the uplift at the end of Pliocene time.

The oscillations of southern Florida with reference to coral reefs may
be tabularly expressed as follows:

Oscillations of Southern Florida.

Recent Depression (modern) reefs,

r Uplift.
„, . ) Depression (Pleistocene reefs, parts of which now stand as

Pleistocene < ^^^^ ^3 j8 feet above sea-level).

I Uplift.
Pliocene Depression (some reef corals but no well-developed reefs).

The results of the recent investigations by Sanford, Drew, Dole, and
myself have revealed the salient facts in the geologic history of the tract
of the Florida keys and reefs, except the immediate foundation rock of the
living barrier reef. Great progress was made by clearly showing the
process by which the oolites were formed and by proving that the atoll rims
of the Marquesas and Tortugas are not due to the solution of an interior
mass of limestone but to constructional geologic processes. The hypoth-
esis that Biscayne Bay was formed by marine solution is also definitely
disproven.

This history may be summarized as follows: Pliocene deposition was
followed by uplift, which was succeeded by depression; during this Pleisto-
cene subsidence along a curve from the eastern side of Biscayne Bay, first
trending southward and then bending westward, a barrier coral reef flour-
ished, separated by a channel from the main bank on which the Miami
oolite was forming or had formed in strongly agitated waters. West of
the coral reef, on an extensive flat in shoal water, the Key West oolite
was formed, while still farther to the westward the Tortugas were out-
lined under the influence of winds and currents. This period of events
was succeeded by the elevation of the entire key region to more than 50
feet above its previous level. This uplift was succeeded by one of depres-
sion, lowering the surface 30 feet or more, establishing practically the
same relation of the sea-level to land that now prevails. Subsequent to
the beginning of this last depression the present barrier reef has developed
seaward of the keys on a platform already prepared for it, the Marquesas
have been formed by winds and currents, and coral reefs have reestab-
lished themselves in the Tortugas.
5

66 Papers from the Marine Biological Laboratory at Tortiigas.

COMPARISONS OF THE FLORIDA REEF TRACT WITH SOME OTHER
CORAL REEF AREAS.

In order to compare the relations of the barrier reef of Florida with
the last change of sea-level in Cuba and in Andros Island, Bahamas, the
following statements will be made: Hayes, Vaughan, and Spencer^ have
shown, as is evidenced by the pouch-shaped harbors of the Cuban coast
and filled channels, such as the submerged filled channel in Havana Harbor,
that the last movement of the Cuban coast has been downward with refer-
ence to sea-level. As an account of the Cuban reefs would be out of place
here, reference is made to that published by A. Agassiz.^ The platform on
which the Cuban reefs grow has been brought to its present position by
subsidence. The great barrier reef of Andros Island, Bahamas, as I have
shown,' occupies the outer edge of a depressed platform. Therefore the
Floridian, Andros, and Cuban reefs all have a similar relation to the oscilla-
tion of sea-level, as in each instance there has been elevation antecedent
to depression which has brought the platforms on which the reefs stand
into their present positions.

There is, however, in one important respect, a difference between the
relations of the Recent reefs of Florida, Andros Island, Bahamas, and of
those of Cuba to Pleistocene oscillation. The oscillations in Florida and
Andros Island have taken place without notable differential crustal move-
ments; while in Cuba there was notable deformation accompanying the
oscillations. The Pleistocene terraces rise in height toward the eastern
end of Cuba in Orlente Province, where altitudes of about 600 feet are
attained near Cape Maisi; while in Santa Clara, Matanzas, and Havana
provinces terraces on the north side are distinct and rise in several steps to
heights of about 200 feet, but are absent or indistinct on the southern side.'*
These facts show that there has been pronounced tilting in late Pleistocene
time. In Barbados Pleistocene reefs extend to 1,000 feet in elevation.

By working out the salient features in the development of the Florida
reefs and instituting comparisons with the West Indies a basis has been
supplied for explaining certain puzzling relations of the reefs in the Tropical
Pacific. A. Agassiz discovered that in the Paumotuan atolls the Recent
corals were growing as a thin crust on an older limestone foundation. His
explanation of the formation of the atolls by the destruction of the interior
of a limestone mass must be discarded, as they were certainly not formed
in that manner but by constructional agencies. There was evidently a
period of atoll formation in the Paumotus previous to the growth of the
Recent corals, which have established themselves on an atoll basement
already prepared for them. A great development of Pleistocene (and per-
haps late Tertiary) coral reefs in the tropical Pacific has been proven in the
most convincing way.

1 Report on a geological reconnaissance of Cuba, made under the direction of General Leonard Wood,
Military Governor, 1901. Report of tiie Military Governor for 1901.

• Bull. Mus. Comp. Zool., vol. 26, pp. 133-136. 1894.

• Cam. Inst. Wash. Year Book, No. 11, p. 154, 1912; Cam. Inst. Wash. Pub. 182, page 50, 1913.

• Hayes, Vaughan, and Spencer, Geological reconnaissance of Cuba, pp. 18-20.

Building of the Marquesas and Tortugas Atolls, etc. 67

There is abundant evidence of differential crustal movement in the tropi-
cal Pacific similar to that indicated for the West Indies. E. C. Andrews^
and C. Elschner^ have both described warping or tilting, the former for the
Fijis, the latter for the Pacific more or less in general. Agassiz described
an elevated atoll, Makatea, in the Paumotus, and Andrews has given a
detailed description of the elevated atoll of Mango, Fijis. The old reefs
have been subjected to differentiated earth movement, so that in certain
places old atolls, such as Rangiroa and others in the Paumotus, now stand
at or near sea-level, while other atolls have been uplifted to heights as great
as 230 feet in Makatea and 600 feet in Tuvuth^, Fijis. In other areas
there has been depression, Bora-Bora, for instance, as Dana, P. Marshall,
and more recently W. M. Davis, have shown.

In conclusion it will be said regarding the Pacific coral reefs: (i) Atolls
are not formed by solution, but by constructional geologic agencies; (2) there
was in the tropical Pacific a great development of Pleistocene and perhaps
late Tertiary reefs which have subsequent to their formation been sub-
jected to extensive differential crustal movement; (3) these deformed older
reefs are frequently the basement of the Recent reefs. Although important
contributions have already been made to the study of Pleistocene and
Recent oscillation and deformation in the Tropical Pacific, there is great
need for more extensive and more detailed investigations of this phase of
the coral-reef problem in order to ascertain more accurately the relations
of the older to che Recent reef series.

1 Notes on the limestones and general geology of the Fiji Islands; Bull. Mus. Comp. Zool., vol. 38, PP^
27-30. 1900.

* Corallogene Posphat-Inseln Austral-Oceaniens, pp. 1S-19. Lubeck. 1913.

V.

SOME CHEMICAL CHARACTERISTICS OF SEA-WATER

AT TORTUGAS AND AROUND BISCAYNE

BAY, FLORIDA.

By R. B. dole,

Chemist U, S. Geological Survey.

One Map.

69

SOME CHEMICAL CHARACTERISTICS OF SEA-WATER AT TORTUGAS
AND AROUND BISCAYNE BAY, FLORIDA.

By R. B. Dole.

INTRODUCTION.
The chemical tests at Tortugas were performed by the writer in June
1913, in the Marine Biological Laboratory, Tortugas, Florida, for the
primary purpose of ascertaining what soluble effect, if any, carbon dioxide
in sea-water might have on coral and other deposits of calcium carbonate.
The tests of waters from Biscayne Bay were made to ascertain the differ-
ences in concentration of sea-water in the bay and the diluting effect of
Miami River. The samples for the latter work were collected by the
writer June 23, 1913, and they were examined during the summer by E. C.
Bain, junior chemist, U. S. Geological Survey.

DETERMINATION OF CARBON DIOXIDE.

To 100 c.c. of the sample, in a Nessler tube measuring 17.5 cm. to the
graduating mark, 10 drops of i per cent phenol phthalein was added and
the solution was titrated to an acid reaction by means of N/20 sulphuric
acid. This is the usual procedure for estimation of normal carbonates
in water. The result of the titration has been calculated to "carbon dioxide
{CO2) present as carbonate (CO3)" by means of the formula A = 22Q/d,
in which A represents milligrams per liter of carbon dioxide, Q the number of
cubic centimeters of N/20 sulphuric acid required, and d the density of the
water.

A sufificient excess of N/20 sulphuric acid was added to 200 c.c. of the
sample in a Jena flask and the solution was gently boiled long enough to
drive off carbon dioxide, after which the total quantity of acid consumed
was determined by titrating the excess with barium hydrate in presence
of phenolphthalein. This procedure was followed by Fox^ and others in
determining the alkalinity of sea-water. The "carbon dioxide (CO2)
present as bicarbonate (HCO3) " was computed by means of the formula
B=ii(Q' — /^Q)ld, in which B represents milligrams per liter of carbon diox-
ide, and Q' the total quantity of acid required. The result of the second
titration has also been expressed as "total alkalinity in equivalent OH"
in order to afford comparison with the results of certain other analysts.
This calculation has been made by the formula C = ^.2-j^Q'ld, in which
C represents the equivalent alkalinity in milligrams per liter of the hydroxy!
radicle (OH).

» Fox, C. J. J., On the coefficients of absorption of the atmospheric gases in distilled water and sea-
■ivater. Conseil permanent hiteraational de la mer, pub. de circonstance 44. February 1909.

71

72 Papers from the Marine Biological Laboratory at Tortugas.

DETERMINATION OF CHLORINE.
Chlorine was estimated by means of the salinity outfit suppHed by the
Copenhagen laboratory of the Conseil permanent international de la mer.
This apparatus was obtained through the courtesy of the United States
Bureau of Fisheries. The procedure is an adaptation of the usual method
of estimating chlorine by titrating with silver nitrate in presence of potas-
sium chromate. An essential feature is a sealed tube of standard sea-
water whose content of chlorine has been very carefully determined. This
water is used for comparison and the apparatus is so constructed and cali-
brated as to insure maximum accuracy. Standardized burette No. 8,
measuring about 1.5 mm. between graduations, and pipette No. 3, having
a capacity of 15.04 c.c, were used. Standard sea-water P7 2/2, 1912, with
a chlorine content of 19.386 grams per kilogram, was titrated frequently
during the tests at Tortugas. In the later tests comparison was made
indirectly with the same standard by means of a large sample of sea-water
from Tortugas that had been titrated sexeral times. Knudsen's correction
{k) has been applied to the titrations^ and salinity (5) has been computed
by means of his formula,^ in which CI represents grams per kilogram of
chlorine: 5 = 0.030 + 1.805 CI.

SPECIFIC GRAVITY.

Density was determined by comparison in 10 c.c. weighing flasks with
distilled water at 25° C. Knudsen's values^ for pvj,^ were used for com-
puting density {S^^ as follows S^l = i + p~, where pn.s = 1. 00129
(0.1245 + o-o - 0.05950-0 + O.oooi55(ro2) and ao = — 0.069 + 1.4708
CI - 0.001570CI2 + 0.0000398 CP. Part of the relatively slight difTerence
between determined and computed figures for sea-water around Biscayne
Bay may be attributed to difference of standard; that is, by the deter-
mined value the sea-water is compared with distilled water at 25° C,
while by the calculated value the specific gravity of the sea-water at
17.5° is referred to distilled water of the same temperature.

SEA-WATERS AT TORTUGAS.
It is believed that solution of calcium carbonate by carbon dioxide in
sea-water, if such action takes place, would be shown by regular differences
in condition of carbon dioxide in the ebb and flood waters passing out of
and into the lagoon surrounded by the shoals and keys of the Tortugas.
Accordingly, samples of sea-water were collected twice a day from the
middle of Southwest Channel about a mile southeast of Loggerhead Key;
these were taken in bottles provided with washered caps that could be
clamped down to prevent escape of gases and the samples were kept on
ice until they were examined. Those collected from June 11 to 16, inclu-
sive, were tested immediately, and though those taken before the former
date had been stored for several days the tests do not indicate that the
delay in examination had any appreciable effect on their composition.

* Knudsen, Martin, Hydrographical tables. Copenhagen, ipor.

Some Chemical Characteristics of Sea-water at Tortugas.

73

Table i. — Salinity and Carbon- Dioxide Content of Sea-Water from Southwest Channel,

Tortugas, Florida.

May 20

31

June

Ebbing

Flowing

Nearly slack ..

Nearly full

Ebbing

Flowing

Ebbing

Flowing

Ebbing'

Flowing

Ebbing

Flowing

Ebbing

Nearly slack...

Ebbing

Flowing

Ebbing

Ebbing

Flowing

Ebbing

Flowing

Ebbing

Flowing ,

Ebbing

Flowing

Ebbing

Flowing

Ebbing

Flowing

Ebbing

Flowing

Ebbing

Flowing

Flowing

I Ebbing

7| Flowing

I Ebbing

8 Flowing

I Ebbing

9 Ebbing

Flowing

Flowing

Ebbing

Flowing

Ebbing

Nearly slack..

Flowing

Ebbing

Flowing

Ebbing

Flowing

IS| Ebbing

: Flowing

i6l Ebbing

•O o -do

•|§gi2g9;

Z 4)2 u jj cs

Grams per

19.91 ]
19.93
19.94
19.94

19.92 1
19.93
19.91
19.88
19.8s
19.89
19-93

19.92
19.93
19.92
19-93
19.93
19.94
19.93
19.93
19-96
19-94
19-94
19-95
19-97
I9-9S
19-94
19-94
19.96
19.96
19.95
19.97
'19.50
19.9s
19.95
19-93
19.94
19.9s
19.9s
19-95
19.96
19.94
19.94
19.98
19.98
19.96
19.97
19-92
19.92
19-93
19-93
19-8S
19-86
19.91

Kgm.
35-97
36.00
36.02
36.02
35.99
36.00
35.97
35.91
3S.86
35-93
36.00

35.99
36.00
35.99
36.00
36.00
36.02
36.00
36.00
36.06
36.02
36.02
36.04
36.08
36.04
36.02
36.02
36.06
36-06
36.04
36.08
=35-23
36.04
36.04
36.00
36.02
36.04
36.04
36.04
36.06
36.02
36.02
36.09
36.09
36.06
36.08
35.99
35.99
36.00
36.00
35.86
35.88
35.97

^ Carbon dioxide (COs).

CO.

0.80
.92
.68
.79
.89
.96

c.c.

9.66
9.58
9.54
9.63
9.88
9.70

■93 9.80

.76 10.07

.51 1 9-48

.80 10.32

.51 10.26

.58
.62
.72
.70
.75
.75
-83
.82
.92
.79
.86
.99
.94
.82
.88
.80
.70
.89
1.04
.77
.80
.75
.85
.79
.77
.79
.81
.77
.73
.61
.54
.58
-94
.80
.80
.74
.77
.88
.77
.71
.80
.68

Highest ebbing

Lowest ebbing

Average ebbing

Average deviation ebb-
ing

Highest flowing

Lowest flowing"

Average flowing

Average deviation flow-
ing

Highest entire series

Lowest entire series

Average entire series

Average deviation entire
series

Composite flowing

Composite ebbing

19.98
19.85
19.932

=t.02

19.98
19-86
19.936

±.02
19.98
19.8S
19-934

±.02

36.09
35.86
36.01

36.09
35.88
36.01

36.09
35-86
36-01

9.46
9.49
9-50
9.75
9-36
9-72
9.68
9.28
9-53
9-57
9-51
9-41
9.44
9-37
9-43

10.21
9-45
9-34
9.72
9-54
9.93
9-4S
9.59
9-45
9.45
9-64
9-55
9-46
9-46
9-SS
9-52
9-43
9-54
9.86
9.44
9-49
9-23
9.41

10.04
9.33
9.56
9.42

00

SB

01 rt

2- 2

<« CO

*J Or )
Si ^^

c-Sq

— Ml'

^"1

0.018
.020

.ois
.017
.020
.020
.020
.017
.011
.018
.011

Grams per kilogram.

.013

.014
.016
.015
.016
.016
.018
.018
.020
.017
.019
.021
.020
.018
.019
.018
.015
.020
.022
.017
.018
.016
.019
.017
.017
.017
.018
.017
.016
.013
.012
.013
.020
.018
.018
.016
.017
.019
.017
.016
.018
.015

9-53
9-60

.021
.011
.017

=4= .002
.022
.013
.018

±.002
.022
.Oil
.017

=t.002

.0x6
.014

0.069
.063
.073
.069
.068

.062
.065
-075
.080
.076
.088

0.087
.083
.088
.086
.088
.082
.08s
.092
.091

■094
.099

.076
.075
.071
.074
.068
.072
.068
.064
.062
.069
.065
.059
.060
.065
.063
.075
.071
.062
.060
.069
.072
.069
.066
.067
.068
.069
.067
.070
.068
.076
.079
.076
.062
.071
.067
.070
.066
.063
.075
.069
.068
.072

.089
.089
-087
.089
.084
.088
.086
.082
.082
.086
.084
.080
.080
.083
.082
.093
.086
.082
.082
.086
.090
.085
.08s
.084
.08s
.086
.085
.087
.084
.089
.091
.089
.082
.089
.08s
.086
.083
.082
.092
.08s
.086
.087

.088
-059
.070

= .004
.076
.060
.068

= .004
.088
-059
.069

= .004
.071
.075

.099
.080
.087

±.003
.094
.080
.086

±.002
.099
.080
.086

±.003
.087
.089

0.0402
.0399
.0398
.0402
.0411
.0404
.0408
.0419
.0395
-0430
-0427

.0394
.0396
-0396
.0406
.0390
.0405
.0403
.0387
.0396
.0399
.0396
.0392
.0394
.0391
-0393
-042 s
.0394
.0389
.0405
.0398
.0413
.0394
.0400
.0394
•0394
.0402
.0398
.0394
.0394
.0398
.0397
.0393
.0398
.0410
.0394
.0396
.0385
.0392
.0418
.0389
.0399
.0392

1.02748
1.02750
1.02752

I.027S2
1.02749
1.02750
1.02748
1.02743
1.02739
1.02745
I.027SO

1.02749
1.02750
1.02749
I.027S0
I-O27S0
I.027S2
1.02750
1.02750
1.02754
1-02752
1-02752
1-02753
1.02756

I.027S3

1.02752

1.02752

I.027S4

1.02754

1.02753

1.02756

1. 02691

1.02753

1.02753

1.02750

1.02752

1.027S3

1.02753

1.02753

1.02754

1.02752

1.02752

1.02757

1.02757

I 1.02754

i 1.02756

j 1.02749

1.02749

I 1.02750

! 1.02750

1.02739

I 1. 02741

! 1.02748

.0427
-0387
.0399

.0430
-038s
.0401

-0430
.0385
.0400

.0008
.0399
.0400

I.027S7
1.02739
1.02750

1.02757
1. 0269 1
1.02749

1.02757
1.02691
1.02750

' After heavy shower lasting one hour.

■ Two determinations; omitted from summary.

74

Papers from the Marine Biological Laboratory at Tortugas.

SALINITY.

The data in table i show that there is no significant or regular difference
in chemical composition of the incoming and the outgoing tide. The
average chlorine content of the flood water is 19.936 grams per kilogram
and the maximum range is only 0.12 gram per kilogram, while the content
of the slack water is 19.932 grams per kilogram with a maximum range of
0.13 gram per kilogram. The close agreement of these averages, the narrow
range of the values, and the insignificant differences between daily tide
contents indicate that there was no essential difference in concentration
of the waters during the period in which the samples were collected. These
results, with those in table 2, show, however, that the salinity of the sea-
water at Tortugas is slightly greater than that of average ocean-water. This
greater concentration is doubtless due to high rate of evaporation in the
comparatively shallow water.

Table 2. — Chlorine Content and Salinity of Sea-Water at Tortugas, Florida.
[Grams per kilogram.]

Source.

Off Loggerhead Keyi ....
Reef, Loggerhead Key^ .
Southwest Channel

D0.2
Wharf, Fort Jefferson^...

Off Garden Key^

D0.3

Moat, Fort Jefferson^....

D0.3

Do.'

D0.3

D0.3

D0.3

Do.'

D0.3

Date.

1910

June 191 2

Average for May and June 1913.

June 1912

Do

May 25, 1913

June 14, 1913

June 191 2

During rain, Jan. 27, 1913

After rain, Jan. 27, 1913

During rain, Feb. 27, 1913

After rain, Feb. 27, 1913

May 25, 1913

June 14, 1913

June 15, 1913

Chlorine. Salinity

Content of ocean water* 19.386

19.60
19-95
19.93
19.99
19.99
19.84
19.96
20.09
19.61
19.60
ip.83
17.79
19.67
19.83
19.19

35.41
36.04
36.01
36.11
36.11
35.84
36.06
36.29
35.43
35.41
35.82
32.14
35.53
35.82
34-67

35-02

1 Computed from report of analysis by G. Steiger, of the U. S. Geological Survey Laboratory; pub-

lished in The Data of Geochemistry, by F. W. Clarke, Bull. U. S. Geological Survey, 491, p. 113, 1911.

2 Tested at the Plymouth (England) laboratory.
' Tested by R. B. Dole.

* Standard P; 2/2, 1912.

ALKALINITY.

The average alkalinity of water from the Southwest Channel, 0.0400
OH gram per kilogram, is that given by Allen and others for normal sea-
water.^ The average alkalinity of the flood water, 0.0401, and that of the
slack water, 0.0399, differ from each other much less than the average
deviation of individual observations (="=0.0008), and agree with the esti-
mates of alkalinity of the composite flood- and slack-tide samples recorded
in the last two lines of table i. The alkalinity of these waters seems to be
practically the same as that of normal sea-water.

The average total content of carbon dioxide of the incoming water is
•essentially that of the outgoing water, and though the recorded values differ
somewhat from day to day the range is not very great or very significant;
there is no regular excess of the values of one set over the other. Though
the probable error of individual determinations, about 0.0020 gram per

• Allen, E. J., and E. W. Nelson. On the artificial culture of marine plankton organisms.
Joum. Micros. Science, vol. 55, pt. 2, June 1910.

Quart.

Some Chemical Characteristics of Sea-water at Tortugas. 75

kilogram, is less than some of the differences, the variations are about what
might be expected in a series of examinations of a flowing water.

The quantity of acid required to neutralize the cold sea-water in presence
of phenolphthalein proves conclusively that "/ree" carbon dioxide, in the
ordinary acceptance of that term in America, is not present. Consequently,
the inflowing waters at Tortugas would have no soluble action on deposits
of calcium carbonate by virtue of the carbon dioxide they contain. One-
half the carbon dioxide in equilibrium as the bicarbonate radicle (HCO3)
is commonly termed the "half-bound" carbon dioxide and that quantity is
probably the carbon dioxide referred to as "free" by some writers on the
composition of sea-water. So far as can be ascertained from the literature
the carbonate radicle (CO3), but not the free acid, is always present in sea-
water; or, as Dittmar expressed it in his computation of hypothetical com-
binations, the carbon dioxide is more than sufficient to complete the forma-
tion of the carbonate but considerably less than is required to form the
bicarbonate.^ Dittmar reports^ the average alkalinity of 130 samples of
ocean water as 54.70 milligrams per liter of CO2, meaning thereby the com-
bined carbon dioxide (total minus the half-bound acid) ; the corresponding
figure computed from the results at Tortugas is 52.5 milligrams per liter.

Briefly, it may be stated that the condition and the quantity of carbon
dioxide in the water entering Tortugas Harbor do not appear to be essen-
tially different from those of the escaping water or normal sea-water. The
inflowing water has no apparent action on deposits of calcium carbonate by
virtue of its content of carbon dioxide.

SEA-WATER AROUND BISCAYNE BAY.
HYDROGRAPHY.

Biscayne Bay ranges from 3 to 8 miles in width and extends north and
south of the mouth of Miami River, though lying mostly south of it. It is
bounded on the east by low keys and reefs, through which there are numer-
ous narrow channels. It is very shallow, being for the greater part only 8
to 12 feet deep. Featherbed Bank, a shoal, cuts off the northern part of
the bay, and coral banks co^•ered with sand extend from Ragged Keys to
Key Biscayne. Miami River is a small stream, discharging only about 400
second-feet at the time of sampling.

COMPOSITION OF THE WATER OF MIAMI RIVER.

No analyses of water from Miami River above tidal influence are avail-
able, but as it receives drainage from the Everglades its water is doubtless
similar in concentration and composition to that of Lake Okeechobee and
Fort Lauderdale Canal, analyses of which are quoted in Table i. These
are analyses of calcium-carbonate waters of low mineral content carrying
considerable organic matter. Normally Miami River may be expected to
contain a small amount of free carbon dioxide and no carbonate (CO3).

' Quoted by Fox, C. J. J. Conseil permanent international de la mer, pub. de circonstance 44. Feb-
ruary 1909.

* Dittmar, William. The alkalinity of ocean water. Report of the voyage of H. M. S. Challenger
1873-76, vol. I, p. 124.

76 Papers from the Marine Biological Laboratory at Tortugas.

Table 3. — Composition of Drainage from the Florida Everglades}
[Milligrams per liter.]

Total solids at 180° C

Total solids at dull red heat.
Organic and volatile matter.

Silica (SiOj)

Iron (Fe)

Calcium (Ca)

Magnesium (Mg)

Sodium (Na)

Potassium (K)

Carbonate radicle (CO3)

Bicarbonate radicle (HCO3).

Sulphate radicle (SOi)

Chlorine (CI)

Nitrate radicle (NO3)

Color

North New
River Canal,

at lock 6
miles above

Fort
Lauderdale.

130
126

4

8.2
•03
26

6.7
19

2.0
.0
94

4.1
26

Lake Okee-
chobee, 3

miles north
of Ritta
Island.

183
143
40
8.2
.03
31
7.0
16
2.0
4.8
104
7.3
28

.1

41

I Samples collected in June 1913, by R. B. Dole; analyzed by W. T. Read.
TESTS OF BAY WATERS.

Samples were collected in or just outside Biscayne Bay at the points
shown on the accompanying map and the chemical tests of them are re-
corded in table 4.

Table 4. — Some Chemical Characteristics of Sea-Water Around Biscayne Bay, Florida.
[Samples collected June 23, 1913. by R. B. Dole; analyses by E. C. Bain.]

No.

Location.

Mouth Miami River inside bar near
Royal Palm dock

i.S miles east of Miami at 1 1 ft. depth
in government channel north of
Virginia Key

Off Cormorant Point (tide ebbing) ...

I mile west of west point of Key Bis-
cayne near inside P & O marker ...

0.2S mile west of Old Man Beacon ...

3 miles northeast of Black Ledge

0.5 mile west of Soldier Key

In channel through Featherbed Bank

1.5 miles west of Fowey Rocks Light,
outside bay

In channel south of Key Biscayne and
o.s mile south of Old Florida Cape
Light, outside bay

At red buoy outside bar across en-
trance to Bear Cut, a mile south-
east of Virginia Key; outside bay...

Temp.

of
water.

Time.

SO-P 5

'C.

26.3 t 4'»30"' a.m.

26.6

27.2
26.0
27.3
27.0
26.4

26.9

2 00 p.m.

4 45 a.m.

5 30

6 00

7 00
9 00

8 00

II 00

to i

•3 '3

(J CO

Grams per kilogram.

27.2 IX 45

I

27.7 I 15 p.m.

Tr.

0.232

0.078

1.42

0.022
.018

.068
.091

.043
.049

14.73

13.3s

.013
.023
.026
.018
.018

.083
.064
.047
.064
.037

.043
.043
.038
.039
.028

16.80
18.79
20.331
20.28I
20.28

.026

.083

.053

19.93

.026

.033

.032

19.83

.026

.055

.042

19.99

26.62
24.13

30.35

33.95
36.73
36.64

35.82

>>r"

>>o^

V ti

a M

S"-"

&:;;

fe-S

«l

laz

•0J3

•og

S3

%.^

u.-g

1.0020

1.0023

1.0203

I.OI97

1.0184

I.OI84

1.0232

1.0228

1.0259

1.0256

1.0281
1.0280

1.0279

1.0280

1.0278

1.0275

1.0270

1.0274
1.0276

1.0274

DETERMINED AND CALCULATED DENSITIES.

No particular importance can be given to the relatively slight differences
between the determined and calculated densities, especially in view of the
absence of several duplicate determinations on each sample. On the other
hand, the fairly close agreement of the estimates indicates that the density
of sea-water, even when largely diluted by river water, may be used in

Some Chemical Characteristics of Sea-water at Tortugas.

77

25f
50''

80°eo'

80° 15'

80°I0'

PORTION OF BISCAYNE BAY, FLORIDA

SHOWING POINTS AT WHICH SAMPLKS OF WATER WERE |)
COLLECTED

25

45

25

40'

25
35

25
30'"

NAUTICAL MILES

04 SAMPLING STAT/ON J^'^ . „

Base traced from U. SCS Chart No. 166 i " °

o7

80°05
]25,

25

45'

25

SoioSoldierKey"^ Fowe^
I -^ T Rocks

80 20

80°i5'

80°io'

80 OS

78 Papers from the Marine Biological Laboratory at Tortugas.

place of the directly determined density in computation of the chemical
analysis. Obviously, however, the relation could not be depended upon
for inland saline solutions, which vary greatly in the percentage composition
of their anhydrous residues.

DIFFERENCE IN CONCENTRATION.

The salinities of the three samples taken outside the reefs (2, 3, and 4)
agree closely with each other and with the salinity of gulf water at Tortugas.
Florida (36.01), which is somewhat greater than that of standard ocean
water (35.02). The water in the south part of the bay is somewhat more
concentrated, samples Nos. 9, 5, and 8 having salinities of 36.73, 36.64, and
36.64, respectively. This evidence that the water in this part of the bay
is concentrated by evaporation during its retention in the shallows serves
further to indicate that circulation there is not very rapid and that the
greater bulk of the water inside the keys is not thoroughly mixed or shifted
by the tides. The inside samples only as far south as Old Man Beacon
give evidence of dilution by fresh water; therefore it may be concluded
that, at least at the time these samples were collected, the effect of Miami
River on the water of the bay did not extend south of Soldier Key nor
outside the keys. Sample l has a salinity obviously higher than the pure
water of Miami River alone may be expected to have, and represents
admixture with bay water; carbonates are absent from it, but bicarbonates
are much higher than in the normal drainage from the Everglades and may
be attributed to reaction of the carbon dioxide that the river water carries.
In all other samples normal carbonates are present as in sea-water and
free carbon dioxide is entirely absent, but there are wide differences in the
alkalinity from point to point.

VI.

OBSERVATIONS UPON THE GROWTH-RATE AND
OECOLOGY OF GORGONIANS.

By lewis R. GARY,

Instructor in Biology, Princeton University.

Two plates.

79

OBSERVATIONS UPON THE GROWTH-RATE AND OECOLOGY OF

GORGONIANS.

By Lewis R, Cary.

The following account includes the record of observations, extending
over a period of three years, upon the growth-rate of three species of gor-
gonians under the natural environmental conditions found on the reefs about
Dry Tortugas, Florida. The oecological data have been secured in the
same periods of observation and are supplemented by observations made
at Montego Bay, Jamaica.

RATE OF GROWTH OF GORGONIANS.

The records of the growth-rate of Gorgonia flabellum and Plexaura
flexuosa are for specimens occurring on the reefs immediately surrounding
Loggerhead Key, the location of the Tortugas laboratory. All of the
specimens have been allowed to remain in their original place of attach-
ment except those shown in the accompanying plates, the measurements
of which are given in table 2. These were cemented to tiles in order that
they could be more easily photographed from time to time to afford a visual
as well as a linear record of their growth.

Table i.

Specimen
No.

Gorgonia

flabellum.

Height.

Increase from —

June 1910.

Jan. 191 I.

June 191 1.

Aug. 1912.

1910-11.

1911-12.

mm.

mm.

mm.

mm.

p.a.

p.ct.

I

»I00

135

150

gone

so

2

'100

138

150

155

so 3.33

3

170

100

145

183

107.14 26.20

4

300

300

374

10 10.30

S

ISO

219

245

46.38 12.38

Specimen
No.

Plexaura flexuosa.

Height.

Increase from —

June 1910.

Jan. 1911. July 1911.

Aug. 1912.

1910-11.

1911-12.

I
2
3
4
S

mm.

80
ISO

SO
24s
IIS

mm. mm. mm.

105 135 175

.... 155 210

6S 83 loi

.... 269 287

ISO 163 183

p.ct.
68.77

3.3
66.66

9.7
41-39

p.ct.
29.63

35.48

21.68
6.69
12.27

' These specimens had been exposed to the air for a long time during an exceptionally low tide on June 6,
1910, so that the height given in the table does not represent the actual growth since attachment, but the
amount living at that time.

6 81

82 Papers from the Marine Biological Laboratory at Tortugas.

The most complete records of the growth of specimens of G. flahellum
and P. flexuosa among my data are those for some specimens growing upon
a single coral head {Orbicella annularis) some 12 feet in diameter, situated
about 200 yards west from the laboratory upon a shallow reef. Table i
shows the growth-record for the period since June 1910, and gives the
original size and the change in height of the same specimens of gorgonians
after 7 months, i year, and 2 years from the time of first measurement; with
the percentage of increase in height for the 2 years.

Similar records of individuals of the same species growing upon tiles,
photographs of which constitute a part of this report, are shown in table 2.

Table 2. — Heights of gorgonians attached to tiles to show groivth in a year.

Tile and species.

July 191 1. Aug. 191 2.

Per cent
increase.

mm. mm. |

T-i» r / G. flabellum 6S 148 [ 127.69

^"^ M P. flexuosa 130 iSO ! 15-38

T-i^ o 1 P- flexuosa (a) / I45 I7S i 20.62

lile 2 1 p_ flexuosa (6) , 70 1 120 70.43

Tile 3. G. flabellum 41 124 202.43

All of the specimens recorded in table 2 were selected without any
special reference to size, as at that time no data were at hand to afford the
basis for a correlation between size and age. Small (young) individuals
of either of the above-listed species were so rarely found on the reefs during
the period of observation in June 1910 or July 1912 that, although the
specimens on any of the reefs studied could be readily arranged into a
series of groups, each containing individuals of about the same size, the
relation in respect to age among the several groups could not be determined
with certainty.

The study of the same areas of reef in August 1912 showed the presence
of a considerable number of small specimens of both species. On the coral
head west of the laboratory there was found on August 21, 1912, seven
specimens of P. flexuosa, which had become attached, or at least had become
noticeable, since July 23, 191 1. The length of these specimens, in order of
magnitude, was 25 mm., 26 mm., 31 mm., 32 mm., 42 mm., 42 mm., and
46 mm., respectively. Beside these there was found a single young specimen
3 mm. in length, which consisted of only two calyces and polyps. The
identification of this specimen as P. flexuosa, which is based upon the
character of the spicules and that of the orifice of the calyx, is at best
uncertain. The specific identity of the specimen is of considerable interest
in connection with other data bearing on the time and duration of the
breeding season, as will be pointed out in the section of this paper dealing
with oecology.

On a shallow reef east of Loggerhead Key, where measurements of all
gorgonians had been made yearly since 1910, there were found 15 specimens
of P. flexuosa less than 50 mm. in length. These were not of a recognizable

Observations upon the Growth-rate a fid Oecology of Gorgonians. 83

size, if present, when the measurements were made over the same area in
July 191 1.

Upon the last-mentioned reef 26 specimens of G, flabellum, less than 65
mm. in height, were also found, which were not recognizable when the
measurements were made in July 191 1.

On the basis of these observations, in conjunction with the records given
in tables i and 2, it becomes possible to make the records complete for at
least two years in the case of G. flabellum (fig. 8, plate 2). Observations
after another year's growth will, moreover, fill in the stages to complete the
record for six years.

All specimens of G. flabellum and of P. flexuosa, on a reef east of Logger-
head Key, were measured in July 191 1 and again in August 1912. Most of
the individuals, as already mentioned, fall into natural "size groups," which
for convenience have been designated as follows: Group i, o to 75 mm.;
group 2, 75 to 150 mm.; group 3, 150 to 300 mm.; group 4, 300 to 450 mm.;
group 5, 450 to 600 mm.; group 6, 600 + mm. As shown in table 3, the
percentage of individuals in groups i and 2 varied markedly in the two
seasons, while the number of the larger specimens was proportionately
similar for the two years.

Table 3. — Size groupings as shown by measurements of all individuals on a single sliallotv reef.

Species and date.

Group I.

Group 2. 1 Group 3.

Group 4. Group 5. | Group 6.

p.ct.

0

15.4

p.ct. p.ct. p.ct.
7.35 36.76 30.88
10.61 20.70 '• 22.02

p. ct. p. ct.
16.17 8.82
18.4s 3.59

6.54 0.62
5.00 o.ss

0.62 8.06 45.16 1 37-09

8.33 17.77 1 Ai..AA \ 2.^.88

P. ilexuosa, 19 12

The observations upon the breeding season and the attachment of the
planulae are intimately connected with those in relation to the absence of
small (young) specimens of gorgonians during the seasons of 1910 and 191 1.
In June 1910 almost every colony of P. flexuosa examined was carrying
mature eggs which could be stimulated to development by the use of chemi-
cals which are ordinarily efficacious in bringing about artificial partheno-
genesis. None of the planulae among those to be obtained in the tow net
could with certainty be identified as those of this species, nor could any of
the colonies be kept long enough in aquaria to secure planula?. The evi-
dence from artificial parthenogenesis is, however, enough to establish the
fact that the breeding season was in progress at that time.

In July 191 1 some of the colonies of P. flexuosa contained ripe eggs, but
the greater number were without apparent gonads. Late in August 1912
not a single specimen was found in which gonads were recognizable, although
more individuals were examined to determine this point than in either of
the other years. The breeding-season for P. flexuosa is, then, apparently
at its height in June. The number of individuals with ripe sexual products
diminishes rapidly through July, while by the latter part of August the

84 Papers from the Marine Biological Laboratory at Tortugas.

breeding season is over for any given year. On August 21, 1912, a young
specimen, probably of P. flexuosa, was found, the size of which (3 mm. in
height) was such that it would have been overlooked had it not been growing
in a location where its color was in striking contrast to the background of
white coral.

To apply these facts to the observed scarcity of yearling specimens in
191 1, which was a general occurrence over all of the reefs visited in the
regions about Tortugas, it is necessary to take into account a severe hurri-
cane which occurred on October 16, 1910. The destruction of adult gor-
gonians by this storm is known to have been very severe upon certain
reefs and the amount of debris moved over the bottom in shallow water
was very great, which in itself would account for the destruction of the
very young specimens at that time. In July 191 1 there would be no easily
recognizable young which had set during the same season, and only very
rarely had any of those from the set of the year before been able to survive
the destructive effects of the hurricane. In August 1912, on the other hand,
young specimens of several gorgonians, besides the two above discussed,
were found upon every considerable reef examined. This would seem to
show that the setting of a considerable nuniber of young was the usual
occurrence. That the number of young found during the past season on the
reefs, which had been carefully examined each season for three years, may
be greater than would ordinarily be found might be inferred from the
breeding habits of some other sedentary invertebrates, especially the
Mollusca, where after a season, when for any cause the set is below normal,
an especially hea\y set is to be expected the following year.

ATTACHMENT OF PLANUL^.

Vaughan's observations on the attachment of coral planulae^ show that
for effective attachment they must settle in a place protected from too rapid
currents and wave-action, where the bottom is rough, and where other more
rapidly growing organisms are least likely to obtain a foothold.

For the effective attachment of the planulffi of the gorgonians studied
the roughness of the bottom, i. e., the presence of small depressions or
cracks into which the planulae could settle, appears to be the most important
factor. In every instance the one-year old specimen had its base in a
depression. Frequently the depth of the point of attachment below the
surface of the rock was more than half the length of the colony. The single
(very small) specim.en found, when it consisted of only one or two calyces,
was entirely below the level of the surface of the coral head to which it was
attached. Many of the adult colonies have their point of attachment at
the bottom of a crack of considerable depth, so that, on the basis of the
observed growth for a year, they must have remained below the general
level of the bottom until more than one year old. In the later growth of
the gorgonian colony its non-polyp-bearing basal portion spreads beyond the

« Carnegie Institution of Washington Year Book No. lo, pp. 151-152.

Observations upon the Growth-rale and Oecology of Gorgonians. 85

limits of the crack over the bottom, so that its apparent height is not
measured from its original point of attachment.

On the shallow reefs, where the gorgonians are most abundant, the
amount of algae, bryozoons, and other incrusting organisms appears to vary
only slightly during the different seasons of the year. In no observed
instance did these organisms seem to form a mass dense enough to exclude
a sufficient amount of water for the well-being of the young gorgonian
colony. Frequently the fronds of algae extend beyond the free ends of the
gorgonian colony, but the specimens found in such locations appear, to
judge from comparative measurements, to have suffered no harm from being
so closely surrounded.

In comparison with the young coral polyp the gorgonian colony has an
obvious advantage in that its most rapid growth is perpendicular to the
surface of the substratum, which would keep the most rapidly growing
part in a position favorable for the securing of food and oxygen.

On the sea-bottom, away from the shallow reefs, gorgonians are gener-
ally found as scattered specimens or in groups of at most only a few indi-
viduals. These colonies will almost invariably be found to be attached to
some irregular mass of rock which affords the protecting cavity necessary
for the permanent attachment of the young gorgonian.

DESTRUCTIVE AGENTS.

Of the destructive agents to which the gorgonians are subjected the
wave-action during severe storms is apparently by far the most important.
The wave-action developed by the ordinary northeast trade-winds during
the winter months is of sufficient force to cause the destruction of many
colonies which have a comparatively weak attachment. The tendency for
any colony to be torn away will increase from year to year as its surface
area becomes increasingly larger.

An opportunity to study wave-action at its maximum severity was
afforded the writer in January 191 1, and later, in July of the same year,
when the observations made during the first visit were confirmed and
extended to a much greater area. The hasty observation made in January
shows that, as the result of the hurricane of the previous October, there had
been a great destruction of gorgonians on the reefs about Tortugas. On the
east shore of Loggerhead Key, the only one visited at that time, many
specimens of some five or six species of gorgonians were found thrown up on
the beach along its entire length. At that time no estimate of the number
cast up was made nor was the area visited sufficient to give conclusive
evidence of the proportion of colonies carried from their natural location.
In July 191 1 the observations on these points were extended to cover a
large area of shore-line and submerged reef.

At one point on the inside of the east reef, near Bush Key, the gorgonian
skeletons were counted over a strip of beach 112 yards in length, where

86 Papers from the Marine Biological Laboratory at Tortugas.

there was a windrow, perhaps a yard wide, made up of these skeletons.
The number of skeletons in a linear yard (it was practically a square yard)
averages 75.7 for ten counts made at approximately equal distances through
the above-mentioned distance, or nearly 8,500 colonies for the whole area.
This area showed very clearly that the most destructive part of the storm
came from nearly northeast. On the outside of the reef, in the direction
indicated by the wash of the storm, only two living colonies of gorgonians
were found for as far off-shore as the water was sufficiently shallow to allow
one to wade about over the reef. Gorgonians were growing abundantly
over this area when it was visited in 1910, so the destruction had been almost
complete. That the great number of the colonies found in the windrow
on the beach had come from this shallow-water area was shown by the
fact that in deeper water on the outer portion of the reef there was little
evidence of the lessening of the number of colonies below the normal number
for such locations; the colonies here are always comparatively scattered,
never forming dense "thickets," as they do on the reefs in shallow water.
The very shallow water on the outside of the reefs contained a considerable
number of skeletons of Gorgonia flahellum and Plexaura flexjiosa, which
had been broken off from their supports and carried away from their original
location while still attached to a good-sized piece of coral rock. Apparently
these specim.ens had reached their present location at a time when the
wave-action had become insufficient to carry them over the crest of the reef
onto the beach on the lagoon side. By far the larger number of specimens
on the beach, on the lagoon side of the reef, were still attached to a piece of
coral rock, usually of small size. In almost every instance the skeleton
shows that the colony was complete when washed on shore. Any physical
injury undergone had not been to the extent of having branches broken ofif
or, in the case of G. flahelUim, having suffered any tearing of the blade-like
portion of the colony.

On the east side of Loggerhead Key, where the greatest force of the storm
came across comparatively shallow water, the reef just mentioned being
about 3 miles distant, most of the specimens, when examined in January
191 1, had the usual spicule-bearing tissues present, although considerably
macerated in many instances. None of these colonies showed any con-
siderable amount of injury, such as the loss of branches or the tearing of the
Hving tissue from the skeleton.

None of the common gorgonians of the Tortugas region can be kept
alive for any considerable time after they have been broken ofif from their
natural support and allowed to fall over into a horizontal position. When
such a colony is put into a live-car, where most of the other marine inverte-
brates and practically all of the sedentary Coelenterates can be kept alive
for an indefinite period, it will be only two or three days before maceration
sets in. It seems apparent, therefore, that the greatest destruction by
storms comes from the tearing of the gorgonian colonies from their natural
supports rather than from any laceration of the tissues.

Observations upon the Growth-rate and Oecology of Gorgonians. 87

On the reef where the measurements previously mentioned were made,
a considerable number of colonies were carried away during the storm. Of
the remaining colonies, many were found in January 191 1 which had suffered
laceration to a considerable extent. The most common injury observed
was a destruction of the living tissues of the colony, such as would result
from twisting a specimen in one's hand while holding it firmly by each end.
The loss of branches in the branched forms, or the tearing of the skeleton
of the leaf-like portion of G. flabellum, was a very unusual occurrence. The
injuries caused by the twisting from the wave-action were quite evenly
distributed among the different species growing on this reef. Such injuries
often involve as much as half the total surface-area of each colony. In all
large colonies, of whatever species, the injury was greatest over the basal
portion, while the outer end or branches were usually uninjured. At the
time of the examination in January 191 1 there had been comparatively
little ingrowth of new tissue over the denuded areas, so that the extent of
the injury to any colony could be readily determined.

When the same specimens were re-examined in July 191 1, in at least
50 per cent of the specimens noted as injured at the time of the earlier obser-
vations, the reparation had been so complete that there was no longer any
evidence of injury. In all of the colonies where the injury consisted in the
removal of living tissue from about the base of the skeleton, there was yet
an area where the skeleton was exposed. In all cases where the injury was
of this nature there was no evidence of any growth of the living tissue
down over the naked skeleton. The exposed skeleton was usually covered
with a dense growth of algae, bryozoa, or hydroids, so that there seemed to
be no probability that it would again be covered with the normal tissues.

The examination of the same individuals in August 191 2 showed com-
paratively little change since the previous summer. In some instances the
coenenchyma had grown down over the naked skeleton for some little dis-
tance, but usually the basal portions of the skeleton were, just as at the last
examination, covered with algae, incrusting bryozoa, etc.

The leaving exposed of areas of the skeleton near the base of the colony
may have a causal relation to the fact that, on all of the shallow reefs,
gorgonians are found which are to a greater or less extent covered by an
o\ergrowth of MiUepora alcicornis. In most cases the Millepora surrounds
what was really the basal portion only of a gorgonian skeleton. Not at
all uncommonly, however, a colony of Millepora is found which covers a
skeleton of Gorgonia flabellum, in which the mesh-work of the skeleton of
the "leaf" could be made out in all its details. Frequently the relation
between the height and width of the colony, as well as the characters of its
outline, showed that the Gorgonia had not suffered any disintegration before
being covered by the Millepora. As in a small number of instances the
Millepora was observed growing up about the naked skeleton at the base
of the Gorgonia colony, and as the complete naked skeleton of such a colony
is very rarely found in its natural attachment on the reef, it seems almost

88 Papers from the Marine Biological Laboratory at Tortugas.

certain that the Millepora must have, on account of its more rapid growth,
covered over and destroyed the Uving tissue of the colony.

A skeleton thus incrusted with Millepora, which was complete in struc-
tural detail in July 191 1, had by August 1912 been reduced to about one-
third of its original height. Nothing of the Gorgonia skeleton beyond the
stouter branches remained . The whole of the ' ' leaf ' ' had been broken down ,
presumably by the weight of the incrusting Millepora.

The previously mentioned hurricane was, as had been pointed out in
dealing with the growth-records, sufificiently prolonged to destroy almost
every young gorgonian which had become attached during the breeding-
season of that year. No evidence could be secured to determine whether
the destruction took place by an actual washing of the colonies from their
attachment or by the shifting of the easily movable material — algae, sedi-
ment, shells, and even quite large fragments of coral rock — which by filling
up all cracks in the surface of the reef would have been sufficient to smother
the small individuals.

Gorgcnians growing upon the shallow reefs are, with more or less regu-
larity, subjected to exposure to the air. In summer the uncommonly low
spring-tides (when the low full-moon tide comes in the late afternoon) most
frequently come on calm days, when the exposure is most destructive. On
June 6 and 8, 1910, and again on July 23, 191 1, all of the shallow reefs were
exposed for a space of time sufficient to destroy the superficial tissues of most
corals and gorgonians which grew on the highest part of the reefs. While
all of the species of gorgonians studied suffered to a greater or less extent,
the injuries were most severe in the case of G. flahellum and G. acerosa.
Frequently there resulted the death of the distal half of the colony which,
within a few days after its exposure to the air, would slough off from the
uninjured proximal part. Plexaura and Eunecia were much less frequently
injured, the latter in only two observed instances, and then only after pro-
longed exposure. The ability of the two last-mentioned forms to withstand
the exposure is at least correlated with, if not dependent upon, the thickness
and toughness of the coenenchyma into which the polyps may be retracted
when the colony is exposed.

In all observed instances of injury from exposure to the air the uninjured
portions of the colony showed no ill effects from the injury of the distal
portion and their growth was at least as rapid as in an uninjured specimen.

REGENERATION EXPERIMENTS.
The injuries occurring in nature are so variable and are besides so entirely
uncontrollable that a series of experiments was undertaken in order to secure
more definite data concerning the capacity for regeneration in these animals.
All of the experiments here described were extensive in nature, usually
involving skeleton and coenenchyma as well as the polyps. In short, they
were made to simulate the injuries to which the gorgonians are subjected
under natural conditions on the reefs.

Observations upon the Growth-rate and Oecology of Gorgonians. 89

Table 4 shows the nature and extent of the operations performed upon
the gorgonians used in the regeneration experiments and the extent of
recovery at different times after the operations.

Table 4.

Species and
specimen Nos.

Recovery.

Operation June 191 1.

January 19 n.

July 191 1.

August 1912.

No change.

o.S inch lateral

growth.
Spaces nearly filled

with new tissue.
12.25 inches long.

No change.

13 inches high.
No change.

G. flabellum:
No. I....

No. 2

No. 3....

No. 4.---

P. flexuosa:
No. I....

No. 2.. ..
G. acerosa . . .

One-half cut from each

leaf (longitudinally).
Split lengthwise into S

pieces.
Circular pieces cut

from leaf.
Cut back one-half (6

in. remaining).

Side branches cut off,
central ones slabbed
on one side, exposing
the skeleton.

Cut back from 11 inch-
es to 6 inches.

One side of all branch-
es trimmed and
slabbed.

Healed, no regenera-
tion.
Edges healed

Edges healed

Healed over; had
grown to 7.5 inches.

Had healed over the
cut places; new
polyps appearing in
the scar-tissue.

Ends of branches
healed over. Now
7.2s in. high.

No new branches.
Scar-tissue covers all
wounds. No polyps
on scar-tissue of
larger branches.

No lateral growth ....

0.25 inch lateral

growth.
0.25 inch in growth. . .

9.5 inches long, distal
end of leaf rounded.

Polyps now cover all
stripped areas. New
branches (8) have
come in on stumps
of cut branches.

Now ID inches high. . .

Polyps have appeared
in scar-tissue on all
smaller branches but
not on any of the
larger branches.

Pieces of the regenerating branches of G. acerosa were preserved when
the reefs were visited in January 191 1 and sections of this material cut and
studied. In this species there is always found a complete layer of coenen-
chyma over a denuded area which remains free from polyps for some time.
The entodermal canals keep pace with the formation of the other tissues, but
it is only after a considerable time that the bud-like swellings of the canal,
which mark the points of origin of new polyps, make their appearance.
They always follow the same sequence as that shown in the formation of new
tissue over a denuded area, appearing first at the periphery of the wound
nearest to the uninjured polyps. The formation of new skeletal tissue at
the cut ends of the branches takes place very slowly at first. After the new
rod of skeletal tissue has reached a diameter equal to that of the older
portion the elongation of the branch takes place rapidly. If the living
tissue be removed from about the base of the colony, there is, so far as my
observations go, little down-growth of the tissues from the cut surface over
the naked skeleton.

EXPLANATION OF PLATES.

The disks to which the specimens photographed are attached are 8 inches in diameter. While the reduction
Is not exact in every instance, the diameter of most of the disks in the reproductions is about 2 inches, making
the specimens approximately 0.25 natural size.

Plate i.

1. Tile I. A, Plexaura flexuosa; B, Gorgonia flahelliim; C, Eunecia sp. Photographed

August 191 1.

2. Tile I. Specimens same as above. Photographed August 1912.

3. Tile 2. A and B, two specimens of Plexaura flexuosa. Photographed August 191 1.

4. Tile 2. Same as in figure 3. Photographed August 1912.

5. Tile 3. \, Eunecia sp.; B, Gorgonia flabellum. Photographed August 191 1.

6. Tile 3. Same. Photographed August 1912.

Plate 2.

7. Tile 8. A, Eunecia sp.; B, Plexaura flexuosa; C, Eunecia sp. Photographed August

1911.

8. Tile 8. Specimens as in figure I. Photographed August 191 2.

9. Tile 7. A, Plexaura flexuosa; B, Eunecia sp.; C, Gorgonia flabellum. Photographed

August 1912. See figure 13.
ID. Tile 15. A, Eunecia sp.; one year old, B, Gorgonia flahellum, one year old. Photo-
graphed August 191 2.

11. Specimen of Plexaura flexuosa, one year old; 0.5 natural size. Photographed August

1912.

12. Tile 16. One-year old specimens of P/ea:aMm /e:>£-M05a. Photographed August 1912.

13. Tile 7. Specimens as in figure 9. Photographed August 19 II.

14. Tile 18. A, One-year old specimen, and B, a two-year old specimen of Gorge «ia^&e//MW.

Photographed August 1912.

90

PLATE 1

GARY

PLATE 2

^M

1

^^

Hk^^^I

H

'

I ]

VII.
GROWTH-CHANGES IN BRITTLE STARS.

BY HUBERT LYMAN CLARK.
Three plates.

91

GROWTH-CHANGES IN BRITTLE-STARS.

By Hubert Lyman Clark.

INTRODUCTION.

Although the systematic work of Ljungman, Lutken, Lyman, and
Koehler, which has brought to light hundreds of species of brittle-stars, has
been of a very high quality, the real interrelationships of the genera and
families within the group are still virtually unknown. Indeed the defini-
tions of most of the families and of many of the genera are so hazy as to be
exasperating, and the more one studies the system at present in use the more
assured one becomes that it is artificial and unnatural. The first step
towards a rational classification was that taken by Bell (1892), when he
divided Recent ophiurans into three orders, based on the character of the
arm-vertebrae. Unfortunately his subdivision of these orders is not so
satisfactory, yet no one has attempted any essential improvement on it.
Whether the orders proposed by Bell represent natural groups which have a
definite place in the phylogenetic history of the brittle-stars has not yet
been critically determined; we can only say that they seem to be such.

Although the chief reason why our classification of the brittle-stars is
so unsatisfactory is undoubtedly because no sustained attention has been
given to the problems involved, there are subsidiary reasons which should
not be overlooked. One of these is that the class shows a rather remarkable
morphological homogeneity so that the differential characters which un-
doubtedly exist are very considerably overshadowed by the extraordinary
development of species in the group, and as a consequence the systematic
work hitherto done has been almost exclusively concerned with the descrip-
tion of new forms.

A second reason why so little progress has been made towards the
natural classification of the group is because the development of so few
species is known and in nearly all the known cases attention has been
centered on the larval structures or on the metamorphosis. Until very
recently the wealth of important data to be found in post-larval stages has
been almost wholly overlooked, largely, no doubt, because suitable material
is not easily obtained. But Ludwig (1899) and, more recently, Mortensen
(1912) have made notable contributions to our knowledge of these most
interesting stages.

A third reason why our progress has been so slow is the comparatively
small amount of paleontological evidence available, together with the

93

94 Papers from the Marine Biological Laboratory at Tortugas.

tendency of zoologists to ignore such evidence as there is. Recent work by
the Sollases (1912) is an encouraging step along the right lines.

Finally, no attention whatever has been paid to localized stages, the
importance of which, in both plants and animals, was pointed out by
Jackson (1899) some years ago. They are remarkably well shown in the
ophiuran arm, and the attempt is made in the present paper to point out
some of them; the chief difficulty is our lack of knowledge, save in a very
few species, of the ontogenetic stages with which to compare them.

Brittle-stars are conspicuously common on the reefs and shores of
Jamaica, and when I was invited to make use of the marine laboratory of
the Carnegie Institution of Washington, located, during February and
March 1912, at Montego Bay, Jamaica, I very gladly availed myself of
the opportunity, determined to secure, if possible, material which would
throw light on the post-larval development of some brittle-star.

I wish to express here my thanks to the authorities of the Museum of
Comparative Zoology for the necessary leave of absence from my duties
there.

Material was easily secured at Montego Bay and young stages were
obtained of three different brittle-stars. The purpose of the present paper
is to report the results of my studies on these three species, to sum up our
present knowledge of the post-larval development of brittle-stars, and to
suggest a few conclusions that may be drawn from this knowledge which
will be of service in future work.

MATERIAL.

Although species of Ophioderma and of Ophiocoma are the most common
brittle-stars at Montego Bay, no really youthful stages of species of these
genera were found. As Ophiocoma is known to breed in the early summer,
and Ophioderma is also a summer breeder, at least farther north, it is quite
probable that the season of the year was the reason why the young of these
genera were not found. This was to be regretted, since these genera repre-
sent widely different families and in neither is there any species whose
development is known beyond the metamorphosis. Grave (1900) has
worked out the early development of Ophioderma brevispina from material
collected near Woods Hole, Massachusetts, and has published a preliminary
account (1898) of the early stages in Ophiocoma echinata. In neither case,
however, were post-larval stages available, and I had therefore hoped it
might be possible to supplement Grave's work from material obtained at
Montego Bay. Although disappointed in this respect, I had the good
fortune to find that in the seaweed and fixed animal life (sponges, hydroids,
ascidians, etc.) which form great masses on the mangrove roots, at the
Bogue Islands, west of Montego Bay, brittle-stars literally swarmed. Of
these the commonest was Ophiactis savignyi (Muller and Troschel), which
was exceedingly abundant in the cavities of a red sponge, individuals of
all sizes, from young with disk barely a millimeter across up to adults

Growth-changes in Brittle Stars. 95

whose disk exceeded 5 mm., occurring by the hundred. In the same sponge
there also lived, although not in anything like the same abundance, the
widely distributed and well-known Amphipholis squamata (Delle Chiaje),
whose early development and metamorphosis have been studied by both
Fewkes (1887) and Ludwig (1881). I secured a good series of this species
and, owing to its viviparous habit, was able to get some very young material.
The third species of brittle-star of which I secured young also inhabited
this red sponge and the adults and half-grown specimens were very abun-
dant. It is Ophiothrix angulata (Say), a well-known West Indian species
of a cosmopolitan genus, particularly interesting because it is the type genus
of one of the very few well-characterized families of brittle-stars. Appar-
ently Ophiothrix was not breeding, as Ophiactis and Amphipholis appeared
to be, and the young were very scarce. But enough were obtained to throw
considerable light on the post-larval development of the species.

METHODS.

When the abundance of Ophiactis in the red sponge was discovered, it
was hoped that they could be kept in aquaria and studied as living material,
but this proved to be impracticable. If they were removed from the
sponge and put in clean water by themselves, they soon became sluggish
and died in a comparatively short time; few lived more than 24 hours. On
the other hand, if pieces of the sponge were placed in aquaria, the water
became foul with astonishing rapidity and the death of all animals speedily
ensued. The observations made on the brittle-stars in life were therefore
of little importance. Material was preserved for further study by the very
simple method of killing with alcohol, or with formalin in which corrosive-
sublimate was dissolved, after the animals had been narcotized with mag-
nesium sulphate (Mayer's method). The best material proved to be that
which had been killed and preserved in alcohol. As the calcareous plates
were the parts particularly desired for study, of course any acid reagent
was out of the question. Work on the preserved material has been greatly
facilitated by the use of sodium hypochlorite. This alkaline reagent is so
powerful a solvent of organic matter that when used in full strength it will
reduce a small brittle-star to a heap of calcareous particles in a few minutes.
But as it may be diluted with water and mixes readily with glycerine
without affecting the "clearing" properties of the latter, it can be perfectly
controlled, and the solution of the organic matter and the separation of the
calcareous plates can thus be accomplished as rapidly or as slowly as one
wishes. Usually glycerine served as a clearing agent, but occasionally
better results were obtained from the use of xylol. The latter, however,
will not mix with the hypochlorite.

Attention has been centered on the skeleton, since it is well-known that
in all echinoderms, except holothurians, modifications of the skeleton almost
always accompany changes in any of the soft parts. This is particularly

96 Papers from the Marine Biological Laboratory at Tortugas.

true apparently of ophiurans, the only soft parts which show any evident
modification without corresponding skeletal changes being the podia (tube-
feet), and even they show relatively little diversity.

The investigations have been strictly confined to growth-changes, as
indicative of relationships within the class. I have intentionally avoided,
on the one hand, any discussion of histological changes, partly because this
was aside from the ends in view and partly because such changes have been
so ably and fully described for Ophiactis by Simroth (1876). On the other
hand, I have not entered into any discussion of the homologies of the
skeleton in ophiurans as compared with starfishes, nor of the homologies
of the jaws, as compared with the arms, since Ludwig's (1878, 1879, 1880,
1 881) work has rendered any such contribution on my part quite super-
fluous. Suffice it to say that in all of my work I have had the homologies
suggested by our Nestor of echinoderm morphology constantly in mind,
and I have seen nothing whatever to cast any doubt upon his interpretations
of ophiuran structure, with the possible exception of the homology of the
peristomal plates. On this question I do not care at present to express an
opinion.

RESULTS.

The facts which I have been able to ascertain will undoubtedly be more
plainly set forth if each of the three species is treated by itself. Ophiactis
and Amphipholis are unfortunately rather closely related to each other,
the morphological differences between them being of little importance.
Ophiothrix, on the contrary, is quite different and shows some interesting
peculiarities. As Ophiactis seems to be undoubtedly less speciaHzed than
Amphipholis, it may well be the first genus discussed.

Ophiactis savignyi (M tiller and Troschel).

The large number of localities from which this species has been recorded,
led me to give special attention to the identification of the specimens
from Montego Bay. I have compared them with specimens, most of them
identified by Lyman, from Zanzibar, Mauritius, Singapore, the Philippine
Islands, the Pelew Islands, Fiji, Hawaii, Lower California, Panama, Florida,
the Bahamas, the Bermudas, and Brazil. Absolutely indistinguishable
in every essential particular, these numerous specimens prove that this
species is truly tropicopolitan. It is one of the small species of the genus,
the disk-diameter of adults being usually about 5 mm. and very rarely
exceeding 7 mm. There are usually 6 arms, but about one specimen in
ten is perfectly pentamerous. In some lots the proportion of pentamerous
specimens runs considerably higher than this, while in others the hexam-
erous symmetry is almost unvaried. I have failed to find any correlation
between the pentamerous symmetry and any other character of the indi-
vidual, nor is it any more frequent in adult specimens than in the young.
In addition to its hexamerous symmetry and small size, savignyi is char-

Growth-changes in Brittle Stars. 97

acterized by its color (above green, more or less variegated with whitish
or yellowish; below whitish or yellowish), its rather short arms (5 to 6 times
the disk-diameter), long and large radial shields, spinelets on disk especially
around the margin, 5 or 6 arm-spines, and 2 (sometimes i or 3) oral papillae.
The genital slits are well-developed and bursas seem to be present, so that
in these particulars as well as in the presence of conspicuous radial shields,
savignyi is easily distinguished from the European Ophiactis virens. Like
virens, howe\'er, savignyi is characterized by the remarkable extent to which
schizogony is carried as a means of reproduction. The process is apparently
identical in the two species and has been very fully described by Simroth
(1876), whose observations were based on the European form. The process
seems to begin, in savignyi at least, almost as soon as the adult form is
assumed, for the youngest specimen found has already divided at least once
(pi. I, fig. 2).

No evidence was secured on the important points of egg-laying and
larval form in Ophiactis. None of the adults contained either embryos or
eggs large enough to be visible under a lens. The impression gained from
the examination of the living material was that the eggs are laid and fertili-
zation and development take place outside the body of the mother. There
W'as no indication whatever that the species is viviparous. It was not
possible to decide whether egg-laying was completed for the season by
March first, but the appearance of the adults and the great scarcity of very
young specimens led me to the opinion that breeding was over.

My observations confirm Simroth in his statements that there is no special
plane of division in schizogony, and that the process shows more or less indi-
vidual diversity. Thus in fifteen young specimens selected at random, show-
ing evidence that division had recently occurred, fourteen had 3 arms, the
other only 2 ; the latter had 3 jaws and 3 oral shields, however. Of the four-
teen 3-armed specimens three have 4 jaws and four have only 2. The plane
of division, which passes through two opposite interradii, does not then divide
a jaw, but passes sometimes on one side and sometimes on the other. Only
about half the time does it divide the animal into perfectly complem.ental
halves. The process of healing and the simultaneous rapid growth of the
water-vascular system have been well described by Simroth. Almost, if
not actually, before the circular canal is completely healed buds which will
give rise to new radial canals have pushed out, one at each side of the
healing margin of the now semi-circular disk, and before there is any
external evidence of the new half-disk the buds of two new arms, one at
each side, become visible. Before their growth has proceeded very far,
however, a third arm appears between them, associated with the third
outgrowth of the now completely healed circular canal. The failure of this
third radial canal to start development probably accounts for some of the
5-armed individuals met with, but as the plane of division occasionally
leaves 4 arms on one disk-half and only 2 on the other, it is evident that
these 2-armed halves will form 5-armed adults even with a normal develop-

7

98 Papers from the Marine Biological Laboratory at Tortugas.

ment of 3 new radial canals. As no 4-armed or 7-armed specimens were
seen (excepting of course the 4-armed complements of 2-armed individuals),
it would seem to be true that the presence of 4 arms on a regenerating half
always prevents the growth of the third new radial canal, while the presence
of only 2 arms always prevents the suppression of the third new arm.
In very young regenerating individuals, the difference in size between the
third new arm and its two fellows is usually quite marked (pi. i, fig. l).

As Simroth, Fewkes, and Ludwig have all pointed out, the development
of the skeletal pieces of the new arms takes place rapidly at the tip of the
growing radial canal, forming a covering about it and, as it gives rise to
pairs of podia, forming the vertebrae and arm-plates of successive joints.
In Ophiactis the podia are not peculiar in any way, but are simple, slightly
tapering, blunt tubes with thin, smooth walls. We may therefore turn
our attention to the skeletal plates.

DISK-COVERING.

In the youngest specimen found the disk measured a millimeter across,
but it was only semi-circular, since division had already occurred at least
once (pi. I, fig. 2). That not more than one division had occurred is indi-
cated by the arrangement of the plates on the disk. The older half of the
animal is easily distinguished from the new one developing, the size of the
plates in the two areas contrasting sharply. The older half is covered by
3 pairs of radial shields, a central plate, 3 radial plates, and 6 interradials.
On the developing half-disk it is less easy to identify the plates, but there
seem to be 3 pairs of radial shields, separated from each other by interradial
plates. The arrangement of the plates on the older half corresponds well
with that which Ludwig (1899) figures for specimens of Ophiactis asperuJa
with the disk only 0.68 to 0.87 mm. across. The only difference is in the
relative sizes of the different plates; in asperula the central and radial
plates are conspicuously bigger than the radial shields, but in savignyi this

is not so.

There seems no reason for doubting that in its earliest stages the disk-
covering of savignyi is identical with that of asperula, and Ludwig (1899)
has shown that the same is true of Ophiactis kroyeri. In savignyi, however,
the occurrence of schizogonous reproduction in very early life interferes with
the orderly and uniform development of the disk-covering, shown by
asperula and kroyeri, so that after division has occurred twice it is difficult,
if not impossible, to distinguish the primary plates. This is well shown by
figure I of plate i, which represents a specimen with a disk rather more
than I mm. in diameter. Division has apparently occurred twice and it
will be seen how irregular is the resulting disk-covering. The irregularity
seems to be due to the fact that whereas, in asperula and kroyeri, the new
plates are formed near the distal margins of the radial and interradial
plates, in the regenerating savignyi they seem to be formed at the proximal
edge of marginal plates. In other words, the typical disk-covering develops

Growth-changes in Brittle Stars. 99

centrifugally, while the covering of the regenerated disk arises centripetally.
As division appears to take place repeatedly, by the time savignyi is full
grown there is not the least trace of the original primary plates nor of any
very definite arrangement in the numerous small plates covering the center
of the disk. The radial shields are always conspicuous and cover a large
proportion of the disk. Soon after the disk is a millimeter across, spinelets
arise in connection with the marginal disk-plates. These spinelets are not
outgrowths of the plates, but arise from separate centers of calcification in
the tissue covering their outer surface. Not often is more than one associ-
ated with a single plate, and many plates, especially near the center of the
disk, have no spinelets. When fully formed the spinelets are attached to
the underlying plates, but no actual union takes place and the spinelets are
always movable to some extent. The number of these spinelets varies in
different individuals, but they seem to be relatively most numerous in young
specimens. Adult, dry specimens commonly show them plainly, but they
are often difificult to detect in fresh or alcoholic adults.

MOUTH-PARTS.

As has been stated above, the plane of division in schizogony does not
cut through the jaw, but passes to one side, so that the resulting animals
may have either 2, 3, or 4 jaws. The formation of new jaws takes place
rapidly, but follows of course the formation of the new rays. Consequently
the same factors which control the new-ray formation govern the develop-
ment of new jaws. In case there are, after fission, only 2 jaws on one result-
ing half (pi. I, fig. 3), the outer side of each of the lateral rays must assist
in the formation of new jaws, while in case there are 4 jaws each of the
new-formed lateral rays will give rise to jaw-elements only on that side
which is next the youngest ray. The formation of the jaws takes place as
already described by Ludwig (1881) for Amphipholis and I have nothing of
importance to add to his account. Ludwig has not discussed, however, the
formation of teeth and oral papillae, nor did Fewkes (1887) pay any special
attention to these points. I found it possible to trace the development of
teeth and oral papillse in both Ophiactis and Amphipholis, but as my material
was more satisfactory in the latter genus, the account may for the present
be deferred. I need only say here that there is no difference between the
two genera save in the number and form of the teeth and papillae. As is
well known, the torus (the plate which bears the teeth) forms at the tip
of the jaw where the two jaw-plates (adambulacrals) meet. It is at first
about as wide as high, but grows rapidly vertically and soon is twice as
high as wide or even more. As will be shown under Amphipholis, the oldest
tooth is at the top of the torus, the new ones forming below it. As soon as
a tooth is well formed, the part of the torus against which it rests is resorbed
and thus a socket is formed for it. The process of resorption goes so far
that the torus becomes perforated (pi. i, fig. 5). On an adult jaw of
Ophiactis savignyi there are 5 or 6 teeth of approximately uniform size (pi. i

100 Papers from the Marine Biological Laboratory at Tortugas.

fig. 4), but on a growing jaw the youngest tooth, the lowest, is distinctly the
smallest.

ARJM-BONES OR VERTEBRA.

Although Simroth (1876) has figured the vertebrae of Ophiactis virens,
it has seemed desirable to give a series of illustrations of the vertebrae in
savignyi, not merely because they are obviously different from those of
virens, as shown by Simroth, but because the series in each arm, taken as a
whole, reveals stages in developmient which are of particular interest. As
the origin and growth of the vertebrae are in all essentials exactly as Ludwig
and Fewkes have described in Amphipholis, there is no occasion to go into
great detail here. The growing tip of the radial water-tube is protected by
a cylindrical plate or, better, a calcareous cylinder, back (proximal) of which
arise all the calcareous structures of the arm. The first plates to appear
are the side arm-plates, but no sooner are they well under way than the
rudiment of the under arm-plate appears followed by the pair of calcareous
rods which are to form the vertebra, lying side by side, above (dorsal) the
water-tube. They develop very rapidly and soon the proximal ends are in
contact with each other (pi. i, fig. 6). New joints are arising distal to them,
and by the time half a dozen joints have appeared the two halves of the
vertebra are in very close contact; when 10 joints have formed it will be
found that the first-formed vertebra is now a single, well-formed, bilateral
structure, which can hardly be separated into its component halves. At
this time the vertebra is much longer than wide, a little wider than high,
and narrower distally than proximally. It soon begins to broaden at the
distal end, a sort of knob or projection forming on each side (pi. i, fig. 8).

Further development is chiefly along three lines: rapidly increased pro-
portional breadth, greatly increased height, and growth of spurs, knobs,
ridges, and hollows. The increased breadth is easily seen by comparing
the vertebra of the thirtieth segment of an adult arm (pi. i, fig. 11) with that
of the fifty-fifth (pi. i , fig. 8) ; while the length has only increased 25 per cent,
the breadth has increased 180 per cent; or compare a fully developed
vertebra, as in the tenth segment (pi. i, fig. 14), with the young one of the
fifty-fifth segment, and it will be seen that while it is only twice as long, it
is more than six times as wide. The increased height is easily shown in the
same way; thus although in segment 10 the vertebra is about 8 per cent
higher than long (pi. i, fig. 18), in segment 55 it is more than three times
as long as high. The development of spurs, knobs, ridges, and hollows is
equally remarkable and is easily understood by an examination of the
figures given. Unfortunately no detailed description of an ophiuran verte-
bra, with technical names for the different parts, has ever been published,
Ludwig, Lyman, and others making use of descriptive phrases for the
various parts. Since the structure of the vertebrae will undoubtedly play
a more important part in the classification of ophiurans in the future than
it has in the past, it is desirable to have a uniform nomenclature for the
various parts.

Growth-changes in Brittle Stars. loi

Taking the fully formed vertebra of Ophiactis as a type (pi. i, figs.
14 to 18), we may distinguish the following parts, when the vertebra is
seen from above (fig. 14): the alee form the bulk of the adoral end, their
upper margins forming the upper alar ridges; these alar ridges are widely
expanded at their inner end, but do not meet in the mid-line; the space
between is of little importance in Ophiactis, but is very conspicuous in some
genera; it m.ay be called the zygantriim; the expanded inner end of each alar
ridge is more or less clearly divided into three parts which may be called the
epapophyses; on each side there is an aboral, median, and adoral epapophysis
and in Ophiactis they are all distinguishable, although the aboral is the most
conspicuous. Directly in front of the aboral epapophysis is a broad, nearly
flat platform, the protapophysis, the aboral corners of which are higher
than the median portion. Projecting from beneath this median portion
may be seen the zygosphene, a shining, glassy knob which still shows evidence
of its paired origin. At either side, but still lower, are the aboral hypapo-
physes; back of these and about half-way between them and the alae are the
parapophyses, not very conspicuous or important ridges closely associated
with the median hypapophyses; back of and below the zygantrum is a
conspicuous median, unpaired process, the epanapophysis, which is particu-
larly noticeable in the vertebrae within the disk (pi. i, fig. 19) ; on either side
of the epanapophysis, but in a much lower plane, lies a zygapophysis.

These are all the parts which are visible from above, but if we look at
the adoral surface of the vertebra, (pi. i, fig. 17) we shall see below the
zygapophyses, the adoral pair of hypapophyses, while between the zygapo-
physes is the important space, the zygotreme. Looking at the lower surface
of the vertebra (pi. i, fig. 16) a deep furrow, the taphrus, is seen running from
between the zygapophyses to the zygosphene; it is bounded on either side
by the three hypapophyses and the thin, vertical wall which connects their
bases; the taphrus contains the radial water- vessel and nerve; the lower
alar ridges form the ventral margin of the alae.

Making use of these terms, then, we may describe the vertebra of
Ophiactis as short and high, with a small zygantrum, a broad, low prota-
pophysis, a well-marked zygosphene, a large epanopophysis, well-developed
zygapophyses, and small, adoral hypapophyses. The basal segments of
the arm are so crowded and compressed that they do not fairly show the
appearance of a typical vertebra, but from the tenth to the fifteenth seg-
ment the vertebrse show clearly their characteristic features. Beyond the
twentieth segment youthful characters become evident and these become
more and more marked as one approaches the tip of the arm. This is well
brought out by comparing a basal joint of a young arm with a corresponding
joint of an adult arm. For example, the resemblance is very noticeable
between the tenth vertebra (counting from the base of the arm outward)
of an arm having 18 joints (pi. i, fig. 12) and the thirtieth vertebra of an arm
with more than 60 joints (pi. i, fig. 11), but the difTerence between this
young tenth vertebra and the tenth vertebra of the adult arm (pi. i, fig. 14)

102 Papers from the Marine Biological Laboratory at Tortugas.

is most striking. No doubt the existence of localized stages in the series of
vertebra of the ophiuran arm will be easily demonstrated when the details
of arm-structure are worked out for a large number of species and genera.

The development of the young vertebra into the adult deserves a further
word. It has already been pointed out that during this development growth
longitudinally is very slow as compared with the lateral and vertical growth.
Attention may now be given to the important question as to what processes
of the adult vertebra are the first to develop and which are the last. In a
very young vertebra the alae are the only evident projections, but in one a
little older (pi. i, fig. 8), the aboral hypapophyses are well marked, the
zygosphene is defined, and the aboral end of the protapophysis can be
distinguished; there is also an indication of the epanapophysis. In the
vertebra of the segment, 5 joints nearer the disk (1. e., the fiftieth), the upper
alar ridge is distinct and the protapophysis is fully defined. In the vertebra
of the fortieth segment the parapophyses show and the aboral hypapophyses
are very prominent. In the vertebra of the thirtieth segment the first
indications of the epapophyses are seen in the line of distinction between the
protapophysis and the alar ridges. In the vertebra of the twentieth seg-
ment this distinction is so far completed that the aboral epapophyses and
the zygantrum are well defined.

Before leaving the subject of the vertebrae, a word may be said as to
their articulation with each other. The ala of course furnish the chief
surface for muscular attachment, but the parapophyses and hypapophyses
are important points of attachment. The zygosphene fits into the zygo-
treme of the next distal vertebra and is supported on each side by a zyga-
pophysis. The aboral margin of the protapophysis fits into the space below
the epanapophysis and above the zygapophyses, and the aboral hypa-
pophyses fit into the space below and outside the adoral pair. So while the
zygosphene and zygotreme furnish what might be called the axial joint,
which is thus of the ball-and-socket type, too great vertical freedom is pre-
vented by the protapophysis and epanapophysis, while excessive lateral
movement is impeded by the aboral and adoral hypapophyses.

ARM-PLATES.

The sequence of formation of the arm-plates has been correctly stated
by Ludwig and by Fewkes, their observations on Amphipholis being fully
confirmed by mine on that genus and on Ophiactis. The 2 side arm-plates
appear simultaneously, one on each side, directly back of the terminal
plate, and are the first indication of the formation of an arm-segment.
They very quickly meet in the mid-ventral line and a little later in the mid-
dorsal line. They are followed at once by the rudiment of the under
arm-plate, which arises directly in front of, or distal to, their ventral line
of division. The under arm-plate is soon followed by the upper one, which
is formed in a corresponding position on the dorsal side of the arm. In a
normally developing young arm, the fourth segment is generally the first

Growth-changes in Brittle Stars. 103

one that can be considered complete (pi. i, fig. 6); the third has only a
rudiment of the upper arm-plate; the second lacks the upper arm-plate
altogether; the first lacks both upper and under arm-plates. It is true that
the rapidity with which new segments are formed shows more or less
diversity in different individuals. Apparently, in the regenerating tip of
an adult arm the segments are formed more rapidly and develop much more
slowly than in a young arm. As a consequence, the sixth, tenth, and
fifteenth segments of such an arm are all in about the same stage of develop-
ment. Simroth (1876) has correctly described how the new "arm-buds"
arise on the regenerating half of the disk. As soon as they are visible, the
terminal plate appears at the tip just above the water-tube. It grows
down and around the tube very rapidly and soon the two sides meet and
fuse in the mid-ventral line. By the time this process is complete the
first pair of side arm-plates Is fully formed (pi. I, fig. 7). The first side arm-
plates meet broadly both above and below, the under and upper plates
being relatively very small on the young joints; but both grow more rapidly
than the side arm-plates and slowly but surely force the latter apart. By
the time 20 segments are formed, the side arm-plates are nearly separated
dorsally and the separation is completed before 25 segments have developed.
Ventrally the process is a trifle slower, but it is completed before 30 segments
are formed. On the basal part of the arm, therefore, the side arm-plates
are completely separated above and below, while the under arm-plates are
in contact with their fellows and the upper plates with theirs.

As regards their relation to the vertebra which they inclose, in the adult
arm-segment it can be easily seen that the side arm-plates are in very close
conjunction with the alae, fitting more or less neatly into the area between the
alar ridge and the parapophysis. The under arm-plate is in contact with
and apparently somewhat attached to the aboral and median hypapophyses.
The upper arm-plate rests on and is perhaps attached to the epapophyses
and even to the protapophysis.

ARM-SPINES AND TENTACLE-SCALES.

The first-formed arm-spine is the lowest and no additional spine ever
appears normally between it and the tentacle-scale. It ordinarily appears
as soon as the side arm-plate is well formed, but is usually lacking from the
first segment (i. e., the segment directly adoral to the terminal plate).
On the second segment the lowest arm-spines are well marked; they each
originate from a separate center of calcification close to the aboral margin
of the side arm-plate near its lower end. The second arm-spine is the
next dorsal to the lowest; it may appear almost simultaneously with the
lowest or it may be wholly lacking on the second segment. Although
appearing later, the second arm-spines grow more rapidly than those first
developed and are soon distinctly larger. The third arm-spine of each
vertical series generally appears on the sixth or seventh segment from the
tip ; it lies just dorsal to the second, of course. The fourth spine, just dorsal

104 Papers from the Marine Biological Laboratory at Tortiigas.

to the third, arises very much later; in some individuals it may be found on
the fifteenth segment back of the terminal plate, but in other specimens as
many as 25 segments may have only 3 spines on each side. In large speci-
mens, a fifth, sixth, seventh, and very rarely an eighth spine appear in each
series dorsal to the fourth, but the position of the segments on which these
spines first appear difters much in different specimens. In the largest
specimen seen, the basal arm-segment has 2 spines, the next 4, the next 5,
the next 6, the next 7, and the next 8; then follow 13 joints, each with 7 arm-
spines on each side, 23 joints with 6, 24 with 5, and 18 with 4; the rest of
the arm is unfortunately missing. The species character (5 or 6 arm-spines)
is thus shown from the tenth to the fifty-seventh segment. This is a much
more extended series than is shown by the vertebrae, which are typical from
about the eighth to the twentieth segment only. The next to the lowest
arm-spine is the largest (longest and stoutest), as long as there are only 4,
but as the number increases the lower spines grow very little, while the upper
ones grow rapidly, so that near the base of the arm the next to the upper-
most, or the uppermost, becomes the largest. At the tip of the arm, the
spines are smooth, but they soon develop thorn-like projections on the
lower, adoral margin, and generally on all sides, especially near the tip.
Excepting on the lowest spine, these projections are neither large enough nor
sufficiently definitely arranged to have any particular significance, but on
the lowest spine they are more characteristic. The lowest spines of the
third segment (pi. 3, fig. 12) have an enlarged base and a smooth, nearly
cylindrical tip, but those of the sixth (pi. 3, fig. 13) have the thorn-like pro-
jections or "teeth" of the lower, adoral margin very conspicuous, especially
the terminal one. There are half a dozen or more of these teeth at this
stage, but as the spine grows older they increase in number and occupy the
other angles of the spine, so that at the middle of the arm the lowest spine
(pi. 3, fig. 14), although much more thorny than any of the others, does not
differ from them as evidently as it does near the tip of the arm. It can not
be said, then, that the lowest arm-spine of Ophiactis is, in any sense, char-
acteristic.

The tentacle-scales in Ophiactis savignyi are large and conspicuous,
although there is but one on each pore. It arises almost as soon as the
lowest arm-spine, but it is associated with the ventral surface of the side
arm-plate. It arises from a separate center of calcification and its rudiment
appears simultaneously with that of the under arm-plate. There is nothing
in the development of the tentacle-scales to determine whether they are
homologous with the arm-spines or not.

SUMMARY OF GROWTH CHANGES.

1. Fission occurs as soon as adult form is assumed; it usually results in

complemental halves, but may leave /owr rays on one half and only
two on the other.

2. The 5-rayed individuals arise in either one of two ways: by the forma-

tion of a normal 3-rayed half on a regenerating 2-rayed individual.

Growth-changes in Brittle Stars. 105

or by the failure of the median ray to develop on the regenerating
side of a normal 3-rayed individual. But only about one specimen
in ten has 5 rays.

3. In fission the dividing plane does not cut through a jaw, but passes on

one side or the other; accordingly a half-disk may have 2, 3, or 4
jaws; normally there are 3.

4. The disk-covering, when the adult form is first assumed, consists of

the typical primary-plates, shown by Ludwig to be characteristic
of Ophiactis, but owing to repeated fission, the covering of the disk
is ultimately made up (aside from the large radial shields) of small
plates with no regular arrangement.

5. The torus bears at first none, then i, 2, 3, 4, 5, and sometimes 6 teeth,

of which the oldest (first-formed) is uppermost and the youngest
is lowest (most ventral).

6. Young vertebrae are very much longer than high or wide; the alas are

the first outgrowths to appear, soon followed by the aboral hypa-
pophyses; the zygosphene, protapophysis, and epanapophysis
appear almost immediately thereafter.

7. Typical vertebrae occur in the adult arm from about the eighth to the

twentieth segment.

8. Of the plates covering the vertebrae, the side arm-plates appear first,

followed by the under arm-plates and lastly by the upper. The
side arm-plates, of any one segment, are broadly in contact v.ith
each other both above and below on the young joints, but are ulti-
mately separated dorsally by the growth of the upper arm-plates
and ventrally by the lower.

9. The lowest arm-spine is the first formed; all succeeding spines arise

serially dorsal to it.

10. The typical number of arm-spines is to be found on about the tenth

segment of an adult arm and thence distally for a varying number
of segments, rarely as many as 40.

11. The lowest arm-spine has no characteristic form.

12. The tentacle-scales appear with the first under arm-plate; they may

not be homologous with the arm-spines.

Amphipholis squamata (Delle Chiaje).
The peculiar habitat of the specimens of Amphipholis collected at
Montego Bay led me to question whether the species was really identical
with that found on shelly bottoms in New England waters. But critical
comparison of these Jamaican specimens with others from Woods Hole
(Massachusetts), Casco Bay (Maine), Grand Manan (New Brunswick),
Norway, and Naples has forced me to the conclusion that they can not
be distinguished and are undoubtedly entitled to the name squamata. At
Montego Bay, the species was found only in the red sponge in which Ophi-
actis occurred so abundantly, but in that situation it was fairly common.
The smallest free-living specimen seen had the disk a millimeter across and
of an orange-red color, a peculiarity of the young noted by Fewkes. Most
of the specimens with disks more than 2 mm. across contained young in a
more or less advanced stage of development. The youngest embryo
examined had the disk 0.35 mm. across, and the arms, each with 3 segments,
projected only 0.20 mm. beyond the disk margin. The largest adult did
not exceed 3 mm. in diameter.

io6 Papers from the Marine Biological Laboratory at Tortugas.

In two particulars the Amphipholis from Montego Bay showed rather
striking differences from those studied at Newport by Fewkes (1887), who
says that 4- and 6-armed young "were repeatedly found" and that "adults
with 6 arms were common." I have not found a specimen in my Jamaican
material, either adult or young, which varies in any way from the typical
pentamerous condition. Again, Fewkes says that "ordinarily a gravid adult
would have from 10 to 15 (generally 10) free young," In the Jamaican
form the number of young ranged from i to 3, but in no case were more
than 3 seen. There were no 2 of the 3 at the same stage of development;
for example, in one typical case one young was 0.50 mm. across with 5-
jointed arms, a second was 0.63 mm. across with 8-jointed arms, and a third
was 0.80 mm. across with ii-jointed arms. The young are not born until
the disk is nearly a millimeter across; one mother only 2.2 mm. across the
disk contained a young one 0.90 mm. across. Fewkes does not mention the
size of his adults, and it is quite possible that they were much larger than
the Montego Bay specimens. If so, that fact might account for the physi-
ological differences just noted, in the frequency of variation, and the number
of young. Or it is possible that the small number of young may have been
due to the breeding-season being nearly over at Montego Bay. No special
attempt was made to settle the matter, as it was not until after my return
to Cambridge that I had any doubts as to the breeding-season being at its
height.

Unlike Ophiactis, Amphipholis shows no evidence of schizogonous repro-
duction, but ordinary reproduction by ova begins very early in life. The
development has been so carefully worked out by Fewkes (1887) and Lud-
wig (1881), as well as by several other students, that there is little that is
new for me to set forth, except in certain details omitted by these earlier
writers. As in the case of Ophiactis, attention has been centered on the
skeletal parts, as neither the podia nor any other soft parts reveal any
features of interest, from the present point of view.

DISK-COVERING.

In the youngest specimens seen the disk was covered by 16 plates (a
central, 5 large radials, and 10 interradials) as figured by Ludwig, but the
5 pairs of radial shields were already beginning to appear. By the time
the disk is 0.80 mm. across, there are 5 more radials and 5 more interradials,
while the radial shields are now easily recognized as such. The new radial
and interradial plates arise between the central and the primary plates of
radii and interradii. With increasing size of disk the conspicuousness of the
primary plates disappears, and while they can sometimes be made out in
the adult, they are not as a rule easily seen. The adult disk is covered
by 150 to 300 plates. No spinelets develop in connection with any disk-
scales.

Growth-changes in Brittle Stars. 107

MOUTH-PARTS.

My observations on the mouth-parts confirm fully the work of Fewkes
and Ludwig. I am, however, able to add certain details in regard to the
teeth and oral papillae which were either overlooked by those writers or
considered of too little importance to note. As mentioned under Ophiactis,
the torus forms at the tip of the jaw as a small plate about as wide as high,
but rapidly becomes vertically elongated. Before it is twice as high as long,
the first tooth is formed from a triradiate spicule lying close to the torus
on its inner side, near its upper end. This spicule rapidly forms a triangular
pyramid, the base of which rests very closely against the torus (pi. 2, fig. 5).
Fewkes says that "the teeth are not separate centres of calcification," but
appear "to grow out directly from the adoral region of the torus." My
first impressions were that this statement was correct, so closely are the
teeth associated with the torus, but careful maceration with sodium hypo-
chlorite made it possible to determine the truth. So closely does the young
tooth press against the torus, that resorption (as stated under Ophiactis)
takes place in the latter and a socket is formed in which the tooth sets very
snugly, but probably not immovably. The first-formed tooth attains some
size (pi. 2, fig. 6) before the second is formed. As the torus elongates
vertically the second tooth arises below the first and subsequently a third
forms beneath that, and finally a fourth below that; the maximum number
in Amphipholis from Montego Bay seems to be 4; probably larger specimens
will be found to have 5. The growth of the torus seems to be chiefly at the
lower end, and that of the teeth is mainly distal. In the adult torus the
sockets, in which the teeth are set, completely perforate its skeletal tissue,
with the occasional or perhaps usual exception of the lowest. The torus
is widest and thickest at the top.

Fewkes has laid considerable stress on the formation of a spinous pro-
jection at the distal end of the second adambulacral plate, during the
formation of the jaw, which he considers homologous with an arm-spine.
There is no doubt of the existence of this process. Ludwig and others
refer to it and I have had no difficulty in finding it (pi. 2, fig. i). But in
following its development we find it never becomes a spine or any other
special organ; it simply forms the distal, concealed end of the adoral plate,
to which the second adambulacral plate gives rise. We have already seen
under Ophiactis that the arm-spines arise from separate centers of calci-
fication, but neither Fewkes nor Ludwig noted this fact. Had it been
observed by Fewkes, he could hardly have homologized the distal process
of the adoral plate with an arm-spine; he seems to have been misled by a
superficial resemblance and to have overlooked the important fact that
spines do not arise by budding or breaking off from previously formed
plates.

The origin and development of the oral papillae in Amphipholis is of
especial importance, since their appearance in the adult is the character upon
which the genus is based. They arise very early in the development of the

io8 Papers from the Marine Biological Laboratory at Tortugas.

jaw, for while the adambulacral plates which form the latter are still quite
distinct, their rudimentary beginnings may be seen (pi. 2, fig. 2). At this
stage there are two papillae on each side of each jaw; one is at the outer
(adradial) proximal corner of the first adambulacral, while the other is on
the inner (proximal) side of the second adambulacral, just distal to the oral
tentacle. Later a third papilla arises on the adradial side of the first
adambulacral, close behind the first papilla. The time of appearance of this
third papilla shows considerable diversity ; usually it forms very soon after
the other two, but occasionally it is delayed, and even a specimen a milli-
meter across may lack it (pi. 2, fig. 3). The typical generic character
(pi. 2, fig. 4) is commonly recognizable by the time the disk is 1.30 mm.
across. The existence of the stage with only the proximal and distal
papillae (pi. 2, figs. 2 and 3) is of extraordinary interest, for that is the con-
dition characteristic of the genus Amphiura and we are therefore justified
in saying that Amphipholis passes through an Amphiura stage. Evidently
light is thus thrown on the phylogeny of Amphipholis. The appearance of
the rudimentary oral papilla? and their relation to the first and second adam-
bulacral plates, with their accompanying podia, suggest the possibility that
the oral papillae are homologous with tentacle-scales. Investigations on
other genera will be necessary in order to determine the point.

ARM-BONES OR VERTEBRAE.

Ludwig (1881) has worked out so completely the development of the
vertebrae that there is really very little to be added. My observations do
not disagree with his in a single essential feature, although in details of
secondary hollows and ridges my impressions of an adult vertebra do not
agree with his figures. The largest vertebra I have examined is relatively
longer than his largest, but, as I have already stated, the Montego Bay
specimens of Amphipholis were all small, and their arm-segments are
undoubtedly relatively longer than those of larger specimens. The adult
vertebra may be described as somewhat elongated, with a small zygantrum,
a low, long protapophysis, a well-developed zygosphene, a large epana-
pophysis, well-developed zygapophyses, and very small adoral hypa-
pophyses. In all essentials it is like that of Ophiactis, but relatively much
longer and somewhat lower; the protapophysis is narrower, the para-
pophyses more conspicuous, and the epapophyses less well marked. In all
these particulars the vertebra of Amphipholis is more youthful than that
of Ophiactis.

ARM-PLATES.

In no particulars does the development of the arm-plates differ from
what has been shown for Ophiactis. The side arm-plates appear first and
are followed by the under arm-plate, while the upper arm-plate appears last
of all. In one point, however, development does not go so far in Amphi-
pholis as in Ophiactis. The side arm-plates are never fully separated, either
above or below; at least, in the largest specimen I have examined, one

Growth-changes in Brittle Stars. 109

which has 40 arm-segments, the side arm-plates are still distinctly in contact
even in the first basal segment. In this particular character, then, Amphi-
pholis is more primitive or youthful than Ophiactis.

ARM-SPINES AND TENTACLE-SCALES.

One of the most peculiar features of the development of Amphipholis
is the fact that it is not the lowest arm-spine which appears first, but the
next to the lowest. So unwilling was I to believe this that it was only after
repeated observations, on many specimens, that I satisfied myself there
is no doubt about the fact. Development at the tip of the arm is so rapid
that each side arm-plate of the newest segment usually bears two spines
before a new segment begins to form; this first segment may also have the
rudiment of an under arm-plate. If one examines this distalmost segment
when the side arm-plates have just met in the midventral line, the rudiment
of an arm-spine will be found on the aboral margin of each side arm-plate.
Almost immediately a second rudiment appears below (ventral to) the first,
so that there are two rudimentary spines on each side. They remain of
about equal length for some time, but the upper is the stouter and more
thorny (pi. 3, fig. ii). On the third segment they are of about equal
length, but on the fourth the lower is longer. By the time 8 segments are
formed a third spine appears on each side arm-plate above the other two.
This rapidly becomes the longest of the three, while the middle one is the
stoutest and roughest. By the time 25 segments are complete a fourth spine
dorsal to the third may arise, but it does not become noticeably longer than
the third and is generally shorter. The development of the "second"
spine before the lowest is so contrary to a priori expectations that some
explanation is necessary. It is true that in its early appearance (pi. 3,
fig. 11), it resembles the lowest arm-spine of Ophiactis (pi. 3, fig. 13) and it is
possible to consider the two homologous, since each is the first spine formed
on the segment. But in that case the lowest arm-spine of Amphipholis
requires explanation, for there is nothing homologous to it in Ophiactis.
A preferable explanation seems to me to be the suggestion that we have
here an extraordinary case of acceleration in development. We have
already seen in Ophiactis that after the "second" spine is formed it grows
more rapidly than the lowest and soon exceeds it in size. Now, the whole
appearance of young Amphipholis shows that its development within the
mother's body has led to an acceleration which is most noticeable at the tips
of the arms, although marked on the disk. Is it not possible that accelera-
tion in the development of the arm-spines has led the more-rapidly growing
"second" spine to appear before the lowest? Such a suggestion seems to
me well within the bounds of possibility. In any case, howe\"er, neither the
lowest nor the second arm-spine develops any unusual hooks or teeth or
assumes any particularly characteristic form.

Although the basal joints of the arm show two tentacle-scales on each
side, one associated with the under arm-plate and one with the side arm-

1 10 Papers from the Marine Biological Laboratory at Tortugas.

plate, it is a remarkable fact that at the tip of the arm there are no tentacle-
scales at all. It is not until a dozen or more segments are formed that the
first tentacle-scale appears and it is associated with the side arm-plate,
arising near its outer margin, at the base, from a separate center of calci-
fication. The origin and appearance suggest homolog>' with the arm-
spines. Almost immediately after the formation of the first tentacle-scale,
a second one associated with the under arm-plate begins to form and is
usually visible after one or two segments have given rise to the first one.
The exact place of appearance of the tentacle-scales is subject to a good
deal of individual variation; thus in a specimen with 21 or 22 segments in
each arm, the terminal 15 or 16 have no tentacle-scales, but the sixteenth or
seventeenth has one associated with the base of the side arm-plate, while
the second scale may be seen on the eighteenth or nineteenth segment in
4 arms, but not until the twentieth in the fifth; on the other hand, in older
specimens, with 30 or 40 arm-joints, the first tentacle-scale may appear only
12 segments from the tip of the arm and the second on the fourteenth. The
first tentacle-scale has every appearance of being the homologue of an
arm-spine, but the second is apparently quite a different organ. It must
be considered as a possibility that this second tentacle-scale in Amphipholis
is homologous with the single tentacle-scale of Ophiactis, but the question
involves the examination and study of many related genera before it can be
answered.

SUMMARY OF GROWTH CHANGES IN AMPHIPHOLIS SQUAMATA.

1. The disk-covering when the adult form is first assumed consists of the

typical primary plates, figured by Fewkes and by Ludwig. Rapid
growth, however, soon decreases the conspicuousness of these
plates and in the adult they are seldom recognizable.

2. The torus of the adult bears 4 teeth, which do not arise as outgrowths

from it, but originate from separate centers of calcification. The
uppermost tooth is oldest, the lowest youngest.

3. The second adambulacral plate, which becomes the adoral plate in

the adult, does not bear a spine at its distal end at any stage of
its development, nor does a spine arise at any time by a separation
of a projection of the original plate.

4. The oral papilla arise from separate centers of calcification associated

with the first two adambulacral plates. They may be homologous
with tentacle-scales.

5. At an early stage of development, Amphipholis has only 2 oral papillae

on each side of the jaw, as in Amphiura. It may therefore be
said to pass through an Amphiura stage.

6. The vertebra of Amphipholis are essentially like those of Ophiactis, but

have retained certain youthful features, greater relative length,
less relative height, narrower protapophysis, more conspicuous
parapophyses, and less well-marked epapophyses.

7. Arm-plates arise in the same sequence as in Ophiactis, but the side

arm-plates of any one segment are never separated from each other,
either above or below. The arm is thus more youthful than that
of Ophiactis.

Growth-changes in Brittle Stars. ill

8. The next to the lowest arm-spine and not the lowest is the first one

formed, but the first segment usually has both, on each side, before
the second segment arises.

9. The lowest arm-spine never develops any special or characteristic

form.
10. The distal arm-segments have no tentacle-scales, but at the base of
the arm each pore has two; one, associated with the side arm-
plate, arises after 12 to 18 segments have been formed; the other,
associated with the under arm-plate, arises one or two segments
later. The former may be homologous with an arm-spine; the
latter may be homologous with the tentacle-scale of Ophiactis.

Ophiothrix axgulata (Say).

Two species of Ophiothrix, angulata and cerstedii, are very common in
shallow water throughout the West Indian region. Both are exceedingly
variable in coloration and the specific differences between them are often
hard to make out. The commoner and more variable species is angulata,
and at Montego Bay a dozen specimens could be gathered to every one of
cerstedii. The latter has very characteristic lines of white or cream-color
cross-banding the upper surface of the arms, and these are rarely, if ever,
lacking, though they are occasionally faint. In angulata such cross-lines
are rarely indicated, but there is present a longitudinal stripe of white, or
some light shade, running the length of the arm. Although such a dis-
tinction seems trivial, especially when the rest of the coloration is very
variable, it is really remarkably constant and specimens can generally be
assigned to one species or the other at a glance.

As is generally the case when a species is very common, angulata occurs
in very diverse habitats. It is quite common at Montego Bay among
coral fragments on the reefs, it is very common among the eel-grass on the
flats, and it is exceedingly common among the bryozoa, sponges, mollusks,
hydroids, and ascidians growing in masses on the mangrove roots among
the Bogue Islands. Most of the young specimens obtained were found in
the last-named habitat, especially in the cavities of the same red sponge in
which Ophiactis and Amphipholis occur. No very young specimens were
found, the smallest having the disk 2 mm. across. While this is greatly to
be regretted, much interesting information has been secured from the young
ones that were obtained. No indication of schizogony nor of variation
from the typical pentamerous form was seen.

COLORATION.

Neither Ophiactis nor Amphipholis has a color pattern sufficiently
distinctive to make its development a matter of any significance, but as
already stated, Ophiothrix angulata is most easily distinguished from its
nearest ally by a characteristic color pattern. The development of this
pattern, therefore, is of no little interest. The youngest joints at the tip
of the arm are colorless and contain no pigment. In the third, fourth, or
fifth segment, however, a little pigment begins to appear and gives a tinge

112

Papers from the Marine Biological Laboratory at Tortugas.

of color, deepening with the increasing age of each successive segment.
This pigment is bright purple of a deep shade. It first appears in the side
arm-plates near their upper ends and gradually spreads distally and to
some extent ventrally. It soon invades the margins of the upper arm-
plates and follows them to their distal end. In a typical case it is very
sharply confined to the sides of the upper arm-plates and fails to cross from
one side arm-plate to its fellow of the same segment. Consequently there
is an unbroken, unpigmented {i. e., apparently white) stripe along the dorsal
surface of the arm. In some specimens pigment fails to invade the distal
corners of the upper arm-plates and thus there appears to be a row of white
spots along each side of the stripe; or the pigment may fail in the spine-
bearing ridge of the side arm-plates and then the dorsal white area of the
arms is markedly increased. On the lower side of the arms the pigment
also spreads longitudinally in the side arm-plates and in the margins of the
lower arm-plates, but it rarely if ever crosses the latter. In other specimens
the pigment spreading in the margin of the upper arm-plates follows back
along the converging sides to the proximal end and there is confluent with
its fellow of the other side and with the pigment of the side arm-plates.
In such cases the longitudinal white line is broken into a series of spots,
formed by the mid-anterior and central parts of the upper arm-plates.
When pigment is formed very abundantly the white stripe, both dorsally
and ventrally, is very narrow and the general coloration is uniform deep
purple. Many specimens, however, are not at all deep purple, but are
very light colored, or are brownish, reddish, or greenish. This diversity of
color is very easily explained. Light-colored specimens are simply those
which have developed little pigment. Green specimens are those in which
the purple pigment is masked by a green coloring-matter acquired after
the formation of the purple ; whether this green matter is another pigment
or consists of symbiotic algse I am not prepared to state, as no study of the
point was made on living animals. Brown and red shades are produced by
the masking of purple by other substances. In all cases, however, purple
pigment is the primitive coloring-matter.

The coloration of Ophiothrix mrstedii begins, like that of angidata, by
the development of pigment in the upper part of the side arm-plates. But
the pigment spreads laterally and not longitudinally and the whole develop-
ment of the color-pattern is thus entirely different from that of angidata
and there is never, and naturally there can not be, an uninterrupted white
stripe either above or below. The first bit of pigment to appear may
generally be seen in the third or fourth arm-segmxent in the upper part of
the adoral (proximal) portion of the side arm-plate. This rapidly spreads
across to meet its fellow from the opposite side and also descends along the
side arm-plate ventrally, where it also ultimately meets its fellow from the
opposite side and thus completes a narrow purple ring around the arm.
Meanwhile another patch of pigment arises just above the upper arm-spine
on the side arm-plate and spreads on to the upper arm-plate and down

Growth-changes in Brittle Stars. 113

along both sides of the spine-bearing ridge of the side arm-plate. The
later behavior of this pigmented area varies more or less in different speci-
mens. In what may perhaps be considered typical osrstedii, the two
margins (adoral and aboral) of the area accumulate all the pigment and
the nearly or quite unpigmented area between them separates them sharply
from each other. In such specimens the upper arm-plate is crossed by two
heavily pigmented lines, one of which passes down the side arm-plate adoral
and the other aboral to the spine-bearing ridge. These lines extend across
the under arm-plates and thus form rings encircling the arms. Each arm-
segment, in an individual colored like this, has therefore three pigmented
rings, one near its aboral margin, distal to the spines, one near its middle
just proximal to the spines, and the third at its adoral margin. In many
specimens, however, the ring either just distal or just proximal to the arm-
spines is indistinct or lacking and the one which is present is very broad on
both the upper and under arm-plates. The same obscuring of purple
pigment by green and other coloring matters, referred to under angulata,
occurs in this species also.

It is obvious from the above account that the color pattern of neither of
these species of Ophiothrix can be derived from that of the other very easily,
since one is due to a longitudinal and the other to a transverse distribution
of the pigment. While it is true that some specimens of angulata have the
dorsal white stripe so much broken up, and so much white present elsewhere
on each segment, that they are easily confused with cerstedii, such speci-
mens do not show definite transverse pigmented rings. Possibly further
rearrangement of their pigment might give rise to rings and we could thus
say the transverse pattern was evolved from the longitudinal. Evidence,
however, that such a development actually occurred is wanting, for cer-
stedii certainly goes through no stage of development wherein it shows any
longitudinal distribution of the pigment.

PODIA.

As is well known, the podia of adult Ophiothrix are remarkable for their
highly papillose surface. It is a matter of some interest, therefore, to note
that the podia of the 3 or 4 terminal segments of each arm are perfectly
smooth, those of the next 5 or 6 older ones are rough with small papillae,
while it is not until some 9 or 10 segments are developed that the character-
istic papillose podia are present.

DISK-COVERING.

One of the most interesting facts in the development of Ophiothrix was
discovered in the study of the disk-covering. Although no specimens seen
were small enough to show simply the primary plates, the specimen only
2 mm. across the disk shows certain primary plates, among which the
central one is conspicuous (pi. 3, fig. 18). In a specimen 3.5 mm. across
the disk (pi. 3, fig. 19), it will be noticed that the central plate is not only

8

114 Papers from the Marine Biological Laboratory at Tortugas.

relatively but actually smaller than when younger. Evidently a remarkable
process of resorption is going on. The extent of this resorption is well shown
by the following table :

Diameter of

disk in
millimeters.

Diameter of
central plate.

Length of
radial shields.

Ratio of

radial shields

to central

plate.

2

3

3-5

S

o.6o
•50
.36

.22

0.55

1. 00
1.20

I.8S

0.91
2.00
3-33
8.40

Besides the change in size, another change is going on in this plate,
during the growth of the disk that is no less remarkable. In the youngest
specimen examined the plate is a smooth, thin sheet of lime, entirely
unperforated except for a number of minute holes around the margin. As
resorption goes on these holes encroach more and more on the plate, not by
increase of size so much as by increase of number. When the plate is only
0.25 mm. across, the unperforated area has practically disappeared, and after
that stage is reached it is no longer practicable to distinguish the plate from
the numerous other small plates at the center of the disk. It can not be said
whether the plate is resorbed entirely or whether a minute portion persists.
It will be noticed that this plate is utterly unlike anything hitherto reported
among echinoderms. Instead of beginning as a triradiate spicule, and
steadily increasing in size by dichotomous branching and fusing of the
branches, as an echinoderm plate typically does, this plate starts as a large,
homogeneous disk which by a process of resorption gradually becomes like
its associated plates in both size and structure. In the present state of our
knowledge it is futile to speculate on the significance or homologies of this
peculiar plate.

Very early in the formation of the disk-covering, but not while the
primary plates (excepting indeed the central) are still conspicuous, there
begin to arise (in association with the plates) vertical rods trifurcate at
their free end. These trifurcate spinelets are usually found one to a plate,
but on the radial shields there may be a number, arranged in series of 3
or 4, parallel to the radial margin of each plate. At first there are no
spinelets associated with the central plate, but when the disk^ is about
4 mm. across there may be as many as 5 spinelets upon it. During youth
the spinelets are all of approximately uniform height and terminate in 3
sharp teeth of moderate length; but as maturity is reached some of these
spinelets grow out into spines, not unlike the arm-spines in length and
appearance. They are two or three times as long as the spinelets. The
number and distribution of these disk-spines is very diversified and the
appearance of the disk is greatly modified by their appearance. As a rule,
it may be assumed that the presence of these spines in abundance is indica-
tive of age and their absence is indicative of youth, but the rule is by no
means unvarying. The number of spines and spinelets on the radial shields
is another variable factor in the appearance of the species. When the

Growth-changes in Brittle Stars. 115

radial shields are small (i. e., in the young individuals) a few spinelets
(3 to 6) will give them the appearance of being well covered, but as they
increase in size, if the number of spinelets does not increase they soon have
the appearance of being quite bare. In adults the shields may have very
few spinelets, possibly none, and no spines, but it is just as common to find
each shield with 10 or 12 spinelets and 3 or 4 spines. The bareness of the
radial shields is therefore not a reliable specific character in angulata.

MOUTH-PARTS.

The mouth-parts of Ophiothrix are essentially different from those of
Ophiactis and AmphiphoUs, since there are no oral papillae, but a group of
dental papillae are borne on the lower (ventral) end of the torus, below the
teeth. None of my specimens are young enough to show a stage of develop-
ment in which teeth are present but no dental papillae, though there can
be no possible doubt of the existence of such a stage. In the youngest
specimen I have examined, the torus is already expanded at its lower end
(pi. 3, fig. 15) and bears a group of 6 or 7 dental papillae, above which are
4 teeth (pi. 3, fig. 17). Resorption back of the teeth goes on as in the genera
already described, but in Ophiothrix the foramina thus formed are often
bisected by a more or less complete bar at right angles to the long axis of
the foramen (pi. 3, fig. 16). This indicates that the fully developed teeth
have two slightly separated roots and shows greater specialization than in
Ophiactis and AmphiphoUs. The number of teeth in angulata is 4 or 5,
while the number of dental papillae increases throughout growth, and a
large specimen may have as many as 17 (in the big East Indian Ophiothrix
longipeda there may be more than 30). The first dental papillae arise in a
pair, side by side, below the lowest tooth ; beneath them the growing torus
expands a little and a new papilla forms near the margin on each side ; these
two are thus a little separated from each other. A third pair arises below
them at the sides of the still expanding torus and the space thus left near
the center of the expanded portion of the torus is filled by the formation of a
seventh papilla. Subsequently, although the torus may increase much
in length, it does not expand greatly and the dental papillae are thus arranged
in two slightly curved marginal rows (one on each side) and an irregular
median series. In the adult, resorption has formed foramina back of the
upper dental papillae just as it has back of the teeth (pi. 3, fig. 16). It is
perfectly clear, therefore, from all these facts, that the dental papillae are
strictly homologous with the teeth and are not in any way to be homologized
with the oral papillae of other genera. As already stated, there are no oral
papillae in Ophiothrix^ the oral tentacles being unprotected and conspicU'
ously exposed on the hollowed sides of the jaw.

ARM-BONES OR VERTEBRA.

The vertebrae of Ophiothrix are as different from those of Ophiactis and
AmphiphoUs as are the jaws, yet their development is so similar as to call

Ii6 Papers from the Marine Biological Laboratory at Tortugas.

for no especial comment. In the young vertebra of angulata (pi. 2, fig. 7),
besides the excessive length, we note the very low proximal (adoral) and the
very high distal (aboral) ends, the slight development of the alae, the well-
developed aboral and median hypapophyses, the conspicuous protapophysis,
and a pair of hollowed projections above the zygosphene, which we may
call the zygocceles. In the later development of the vertebra the length
only increases about one-half, while the height anteriorly increases 3.5
times, and posteriorly more than 6 times. The adult vertebra is about as
high as it is long, while the width of the alae exceeds the total length. Seen
from above (pi. 2, fig. 9), the most striking feature is the very narrow
protapophysis, the immense zygantrum, the very small and insignificant
epanapophysis, the reduced epapophyses (particularly the median one),
the broad parapophyses and hypapophyses. Seen from the side (pi. 2,
fig. 8) the great height of the protapophysis, the bulging zygocceles, the
long zygosphene, and the prominent hypapophyses, all attract attention.
Looking at the vertebra from below (pi. 2, fig. 1 1), the breadth of the taphrus
and zygosphene are the most notable features. Looking at the aboral end
(pi. 2, fig. 10) we are again struck by the high, narrow protapophysis, the
conspicuous zygosphene and the zygocceles, while the broad alae furnish
a fitting background. At the adoral end (pi. 2, fig. 12) the reduced area of
the alee caused by the huge zygantrum, the small zygotreme, within which
is a small knob that may be called the zygophore, and the large adoral zyga-
pophyses are the principal features. We may then say that the adult
vertebra is short and high, as in Ophiactis, but, in contrast to that genus,
the zygantrum is large, the protapophysis is narrow, high, and projecting,
the epanapophysis is insignificant, and the adoral hypapophysis is large.
The joint formed by two vertebrae is more complicated than that in
Ophiactis. The protapophysis of each vertebra fits into the zygantrum of
the next distal segment and a compact mass of muscles, attached at their
outer ends to the sides of the cavity, hold the protapophysis in position;
the inner ends of these muscles are attached on either side of the prota-
pophysis. The zygocceles serve as sockets for the zygapophyses, while the
epanapophysis of the distal vertebra rests lightly on them, the zygosphene
meanwhile fitting into the zygotreme and resting against the zygophore.
The aboral hypapophyses fit outside and beneath the adoral pair of the
distal segment. This articulation is, therefore, considerably more complex
than the one described on p. 102 and yet it may permit just as much freedom
of movement. The lateral flexures of the arm could certainly be made
more vigorously and a curved position held more rigidly than in forms with
less-developed zygantra and protapophyses. Possibly the reduced epana-
pophysis permits a greater freedom in vertical movements, but it may be
that the larger adoral hypapophyses counteract this tendency.

Although the development of the vertebra is along the same lines as in
Ophiactis, there is a constantly increasing difi'erence between that genus and
Ophiothrix. After the two halves of the vertebra have come to lie close

Growth-changes in Brittle Stars. 117

beside each other, the progress of development in the two genera is quite
unlike. There is at no time an "Ophiactis stage" in the development of
the Ophiothrix vertebra. No other fact noted has so strongly impressed
me with the radical difference between Ophiothrix and the other two genera
studied.

ARM-PLATES.

In the development of the arm-plates, Ophiothrix shows a striking
difference from Ophiactis and Amphipholis. As in those genera, the side
arm-plates are the first to appear and are almost immediately followed by
the under arm-plate, but in Ophiothrix the growth of the under arm-plate
is so rapid and vigorous that the side arm-plates never meet ventrally, but
are from the first kept apart by the under arm-plates. The latter are in
contact with each other throughout the entire length of the arm. The
upper arm-plates, on the contrary, are very small near the tip of the arm
and it is not until 15 or 20 segments have been formed that they are suffi-
ciently large to separate wholly the side arm-plates. Even at the base of
the arm, the basal (adoral) end of the upper arm-plates is so narrow that
the side arm-plates of any one segment are not widely separated from each
other as they are ventrally. The spine-bearing ridge of the side arm-plates
is very well developed, so that the arms appear to be constricted between
them, and this appearance is accentuated by the fact that the broader,
distal end of each upper arm-plate is somewhat swollen. The side arm-
plates, like the vertebrae, are relatively very much longer distally than near
the base of the arm.

ARM-SPINES AND TENTACLE-SCALES.

In ophiothrix, as in Ophiactis, the lowest arm-spine is the first to appear.
It is, on the first segment, a simple, smooth, straight rod with an enlarged
base (pi. 3, fig. i), but by the time a second segment has formed it begins
to curve at the tip and becomes broader and somewhat flattened (pi. 3,
fig. 2). On the third segment it is distinctly curved at the tip and a tooth
is evident on the concave side (pi. 3, fig. 3). On the fourth segment this
tooth has become conspicuous and another one has arisen below it {i, e.,
nearer the base of the spine), while the spine-base is notably large and
projecting (pi. 3, fig. 4). On the sixth segment these various features are
all accentuated (pi. 3, fig. 5) and on succeeding segments the spine may be
said to show its characteristic form (pi. 3, fig. 6). As the animal matures,
these lowest arm-spines begin to undergo a sort of retrograde change which
well illustrates Hyatt's principle of senescence. Thus if the lowest arm-
spine on the twentieth or thirtieth segment from the mouth be examined,
on an adult arm, it will be found essentially like that on the twentieth
segment from the tip. But if one from somewhere about the twelfth basal
segment be looked at (pi. 3, fig. 7), it will be found to be obviously different
from any of the lowest arm-spines of the middle portion of the arm ; except
that it is immensely stouter, it is more like that of the fourth segment

ii8 Papers from the Marine Biological Laboratory at Tortugas.

from the arm-tip. It has one notable peculiarity in that foramina are
beginning to open through it. Examination of the lowest spine of segment
nine, counting from the mouth, shows (pi. 3, fig. 8) a still further reduction
of teeth, particularly the terminal one, while there is a corresponding
development of the foramina. The lowest spines of the segments still
nearer the mouth (pi. 3, figs. 9 and lo), show these tendencies carried still
further. Indeed, except for their smaller size, these basal lowest spines are
not essentially different from the ordinary spines of the arm.

The development of the other arm-spines offers no such interesting
features as those shown by the lowest. The next to the lowest is the second
one to appear and is generally present on the second segment. The others
appear in regularly vertical sequence dorsally. The third spine is to be
found on the fourth segment from the tip and the fourth on the ninth. It is
not until about 40 segments are developed that a fifth spine appears, while
the arm has no less than 60 joints when the sixth spine is to be seen. A
seventh spine will arise if about 75 joints are formed, and an eighth if the
number is about 90. I have not found more than 8 spines in any specimen,
and think that 6 or 7 is the characteristic number for the species.

The tentacle-scales are reduced to very narrow, sharp spinelets,, which
are quite lacking at the tip of the arm. The first indication of them is on
the sixth or seventh segment, where a minute one is to be found on each
side, connected with the lowest part of the side arm-plate. They never
reach a large size, but even on the basal segments give the impression of
being functionless remnants. There is no indication whatever of homology
with an arm-spine.

SUMMARY OF GROWTH CHANGES IN OPHIOTHRIX ANGULATA.

1. Although the coloration is distinctive, it is acquired very early without

passing through any recognizable stages. It is due to the deposit
of purple pigment in the arm-segments longitudinally and the
usually complete failure of any pigment to cross the median line
either dorsally or ventrally. When the color is other than some
shade of purple, it is due to the masking of the purple pigment
by a secondarily added color.

2. The first-formed podia are smooth, as is usual in ophiurans, and the

papillose condition characteristic of the genus Ophiothrix is only
acquired after about 10 segments are complete.

3. The central plate of the youngest specimens examined is a dispro-

portionately large, homogeneous, thin disk with a few minute per-
forations around the margin. During the growth of the disk this
plate is gradually resorbed, becoming not only much smaller
(actually as well as relatively) but more and more reticulate from
the increasingly numerous foramina.

4. The disk-plates carry trifid spinelets, which not only become more

numerous with growth of the disk, but have a tendency to elongate
into slender spines. The number of these disk-spines is very much
subject to individual diversity.

5. There are no oral papillse in Ophiothrix, but after 4 or 5 teeth are

formed, dental papillae (homologous with the teeth, but much

Growth-changes in Brittle Stars. 119

smaller) arise in double and then in triple series ventral to the
teeth. The torus expands correspondingly, and in maturity the
lower dental papillae may be in irregular quadruple or quintuple
series, though the lowest series are triple or only double, like the
highest. Resorption of the torus occurs back of the largest dental
papillae just as it does back of the teeth.

6. The vertebrae are remarkable for the high, narrow protapophysis

anteriorly and the large zygantrum and very small epanapophysis
posteriorly. The chief articulation between two vertebrae is
formed by the protapophysis of one and the zygantrum of its distal
fellow. Other peculiarities of the mature vertebrae are a pair of
zygocceles above the zygosphene and a small, round knob, the
zygophore, in the zygotreme. The zygapophyses and adoral
hypapophyses are well developed.

7. The side arm-plates are separated throughout the whole length of the

arm, even in the first segment, by the under arm-plates, so that
ventrally the two side arm-plates of any given segment are never
in contact. Dorsally those of the first 15 or 20 segments do meet
each other in the mid-dorsal line, but thereafter the upper arm-
plates keep them apart. The spine-bearing ridges of the side arm-
plates are quite prominent, especially on the older segments.

8. The lowest arm-spine, the first to arise, begins as a straight, slender

rod, but rapidly develops into a characteristic 3-toothed hook.
At the base of the arm, by a process of senescence, it is gradually
transformed into a nearly straight, rough, fenestrated spine like
those dorsal to it.

9. There are only 4 (distally fewer) arm-spines on the first 40 segments and

the full number for the species (6 or 7) does not appear until 60
or 75 segments are complete.
10. The tentacle-scales are reduced to mere spinules associated with the
basal part of the side arm-plate. They do not appear until after
6 or 7 segments are formed and they never seem to be of any
functional importance.

SUMMARY AND CONCLUSIONS.

Although references to changes in appearance, or in details of structure,
during the transition from young to adult brittle-star are to be found in
many of the writings of Lyman, Lutken, Verrill, Grieg, Koehler, de Loriol,
and other less well-known workers, no attention to the sequence or signifi-
cance of these changes has been paid, up to the present time, save by
Ludwig and Mortensen. And although it has been a generally accepted
and well-known fact that the terminal segments of the adult ophiuran arm
are the youngest, no use has been made of the important corollary that these
segments reveal the developmental stages through which the arm has
passed. The present report has endeavored to make full use of this corollary
in studying the development of Ophiactis, Amphipholis, and Ophiothrix. Its
great value lies, of course, in the fact that even where only a single specimen
of a species is available, and that specimen adult, if a single arm is complete
to its terminal plate the development of that species throughout its post-
larval growth, so far as the arms are concerned, can be read almost as

120 Papers from the Marine Biological Laboratory at Tortugas.

plainly and satisfactorily as if a large series of young specimens were avail-
able. Obviously this is not true of the disk. I would emphasize, therefore,
as the first conclusion of these studies, this truth of the very great importance
of localized stages in the distal part of the ophiuran arm. The other con-
clusions which follow are taken directly from Ludwig's very valuable paper
"Jugendformen von Ophiuren" (1899), supplemented by Mortensen's
observations on Asteronyx (1912) and by my own work. The following list
shows the species of brittle-stars, the growth-changes of which have been
more or less fully studied :

OphiacHs asperula (Philippi). Amphiura magellanica Ljungman.

kroyeri Liitken. Ophiacantha vivipara Ljungman.

savignyi (Miiller & Troschel). Ophiothrix angulata (Say).

Amphipholis patagonica Ljungman. Ophiomyxa vivipara Studer.

squamata (Delle Chiaje). Asteronyx loveni Muller & Troschel.

This list strikingly demonstrates the paucity of facts on which to base
a classification of brittle-stars. We have a fair knowledge concerning the
growth-changes of 10 of the species (out of approximately 1,100!), but these
represent only 7 genera and perhaps 5 families. It may be added that a
considerable number of facts are recorded concerning changes during
development from young to adult of several euryalids {Euryale, Gorgono-
cephalus), but in no one species has any series of changes been followed.

The following conclusions, a summary of our knowledge to date, will
serve as a basis for further work and may prove useful in future attempts to
properly locate doubtful genera and difficult species. Continued study of
growth-changes will ultimately lead to a natural classification of ophiurans.

1. The terminal segments of the arm are the youngest; passing from the

terminal plate towards the disk, each succeeding segment is older.
The segments in their individual development pass through the
same stages that the arm has passed through in its development and
therefore the terminal portion of the arm is made up of localized
stages, to use Jackson's (1899) expressive term, which reveal the
developmental or ontogenetic stages of the arm as a whole. The
study of the terminal portion of the arm in great detail is necessary
therefore to the proper understanding of the relationships of any
brittle-star.

2. The original disk-covering in ophiurans consists of a central^ and

5 radial plates; interradial plates are a somewhat later addition;
the radial shields still later; further development of both radial
and interradial plates between the primary plates is a more ad-
vanced stage; the development of spinules and spines in connection
with the disk-plates is further evidence of specialization ; and the
resorption of the disk-plates or their concealment by thick skin
or by a coat of granules must be regarded as still more advanced
stages in the development of disk-coverings.

3. The radial shields are originally small and comparable to any of the

other secondary disk-plates. Their development to a large size
is evidence of specialization. The larger the radial shields in
proportion to the other disk-plates, the more specialized is the disk-
covering, other factors (such as granule formation) being equal.

Growth-changes in Brittle Stars. I2I

4. The oral shields lie originally on the dorsal surface and only by subse-

quent developmental changes come to lie on the ventral side.
The apparent exception shown by Asteronyx seems to me an un-
doubtedly secondary condition ; this is indicated by their late and
unequal appearance.

5. The adoral plates develop from the second pair of adambulacral

(= side arm) plates. A large size relatively, as compared with
the oral shields, is a primitive condition, but a large size actually,
so that they are equal to or exceed the basal side arm-plates, is a
secondary and specialized condition. On the other hand, their
great reduction or entire absence is also a very specialized con-
dition.^

6. The jaws (oral plates) arise, in large part at least, from the first pair of

adambulacral plates. Obviously the more clearly they show their
origin the more primitive is their character, while the moi-e com-
pletely they are fused and overshadowed by either the torus or
the adoral plates, the more specialized is their condition,

7. The tonis arises, at the tip of the jaw, as a small plate not muchhigher

than wide. A low, wide torus, therefore, indicates a primitive
condition, while a high, narrow one or a high one expanded at the
lower end indicates specialization.

8. The teeth arise in association with the torus, but not as a part of it.

The uppermost tooth of the adult was the first one developed, the
later teeth appearing below it successively. Obviously, a small
number of teeth is a primitive condition, and the same is true
where the lower teeth are evidently smaller than the uppermost.
Young teeth are sharply pointed ; therefore pointed teeth are evi-
dently more primitive than those which are rounded or truncate.
A large number of truncate teeth of equal size shows a decidedly
specialized condition.

9. The dental papilla, where they occur, are homologous with the teeth,

but are a modified form of the latter. Their occurrence is there-
fore undoubtedly a mark of specialization, the degree of which
may be indicated by their number and arrangement.

10. The oral papillce arise in connection with the first and second adambu-

lacral plates {i. e., the orals and the adorals) and appear to be
homologous with tentacle-scales. There is as yet no evidence to
determine whether their entire absence is a primitive or highly
specialized condition. Possibly it may be either. There is evi-
dence to show that at least a relatively primitive condition is
indicated when there is one papilla in connection with each oral
and each adoral plate. There can be scarcely any doubt that the
presence of more than one papilla in association with each of these
plates indicates specialization, the degree of which is suggested
by the number of such papillse. The primitive form of oral-
papilla appears to be scale-like, flat with rounded tip; pointed and
spine-like papillse are specialized forms.

11. The genital-slits or bursal openings appear very early, one on each side

of the base of the arm. The evidence is still lacking as to whether

> In Mortensen's (igi2) very valuable and interesting account of growth-changes in Asteronyx, bespeaks
of the "seitenmundschilder" being resorbed and later as being greatly reduced. But examination of rus
admirable figures shows there has been no resorption; the fact is that having reached a relatively large size
quickly, their later growth has been slower than that of the rest of the disk; thus in his specimen 1.9 mm. across
the disk, the adoral plates are each o.ss mm. long, in the specimen 3-S mm. across the disk, they are 0.60 mm.
long, and in the specimen 2S mm. across they are 3 mm. long. Surely there is no resorption here.

122 Papers from the Marine Biological Laboratory at Tortugas.

their absence is a primitive or specialized condition. Perhaps it
may be either. In any case, the fusion of the margins of the slit
so as to form two openings on each side of the arm-base is un-
doubtedly a highly specialized condition.

12. The arm-bones or vertebrcs arise by the fusion of two skeletal pieces

lying side by side above the radial water-tube. They are at first
relatively elongated and low, with the two component halves
distinguishable and the articulation resembling a sirnple ball-
and-socket joint. Specialization is indicated by loss of indication
of the bilateral origin, by greatly increased height, and hence a
relatively short longitudinal axis, a greatly increased number of
elevations, projections, and ridges for the attachment of muscles,
and an articulation complicated by development of zygapophyses,
protapophysis, and epanapophysis. Owing to changes from com-
pression and other causes the vertebrae of the basal arm-segments
do not reveal the typical form for the species. At about 8 or lo
segments from the mouth the vertebrae of an adult ophiuran with
40 or more arm-joints are fairly typical, but after a dozen, or even
fewer (in less mature individuals) segments further out, youthful
characters begin to show and the vertebrae become less and less
typical as we approach the terminal segments of the arm.

13. The terminal plate of the arm is the first formed. It is originally a flat

or concave plate on the upper side of the tip of the water-tube.
It grows down and around the tube and by ventral union becomes
a cylinder. The distal margin of this cylinder may bear one to
several teeth or spinelets. It may become enlarged proxirnally
and even conspicuously cup-shaped, but such changes indicate
specialization of some kind. The most primitive condition is
undoubtedly a simple plate or a half-formed cylinder; a completed
cylinder shows at least some specialization.

14. The side arm-plates are the first plates of an arm-segment to appear.

They grow rapidly and soon meet each other in the mid-dorsal
and mid-ventral lines, where they are in close contact, although
they do not fuse. They inclose the rudiments of the vertebrae,
which do not arise until they are well begun. In later develop-
ment they may be more or less separated from each other ventrally
by the development of the under arm-plates and dorsally by the
upper arm-plates. At first they are low and longitudinally elon-
gated, the whole segment being long and flat. In their usual
development the height increases much more rapidly than length
and they consequently tend to become short and high. The distal
margin is somewhat thickened and bears one or more arm-spines.
It may become notably enlarged and a conspicuous spine-bearing
ridge may be developed, which in the course of growth comes to
lie in the median vertical region of the plate. The distal margin
of the lower end of each side arm-plate may be more or less deeply
grooved for the passage of the podium (tube-foot) and usually a
tentacle-scale arises at the same point. Evidently then, long,
low side arm-plates, meeting above and below and with no spine-
bearing ridge or tentacle-groove, show a primitive condition of
the arm-segment, while short, high side arm-plates separated above
and below, with a conspicuous spine-bearing ridge and a notice-
able tentacle-groove, indicate considerable specialization. All con-
ditions between these two extremes may be found.

Growth-changes in Brittle Stars. 123

15. The under arm-plate arises very soon after the formation of the side

arm-plates and may even precede the formation of the vertebral
rudiments. It is at first small and lies on the oral side of the arm,
in the median line just distal to the side arm-plates. It may grow
backward (adorally) sufficiently to completely separate the side
arm-plates and it then comes in contact with its fellow on the
succeeding segment. In such cases the under arm-plates are more
or less clearly in contact throughout the entire length of the arm.
There is no doubt that small, widely separated under arm-plates
are the simple, primitive arrangement and large under arm-plates
in contact with each other indicate specialization. The absence
of under arm-plates or their occurrence in irregular series and of
diverse sizes are certainly indicative of specialized conditions.

16. The upper arm-plates are the last of the arm-plates to arise, but they

appear usually not long after the under arm-plates and follow a
similar line of development. Small, widely separated upper arm-
plates are undoubtedly a primitive condition, while large upper
arm-plates, in contact with each other and separating the side arm-
plates, are certainly specialized. On the other hand, absence of
upper arm-plates, if associated with large side arm-plates broadly in
contact in the mid-dorsal line, shows a primitive type of structure,
while absence of upper arm-plates if associated with widely sepa-
rated side arm-plates is probably a highly specialized condition.

17. The lowest arm-spine is the first to appear, as a rule, though the one

just above it may accompany or, in very rare cases, slightly pre-
cede it. Succeeding spines arise dorsal to the earlier-formed and
the number increases slowly with the age of the segment until the
number characteristic of the species is attained. This number is
commonly to be found on about the tenth segment from the mouth
and thence distally for a variable number of segments. On the
basal arm-segments the number is always fewer than normal, but
whether the influence of the disk has promoted resorption or pre-
vented formation of the full number has not been ascertained.

18. The hook-shaped form of the young lowest arm-spine is not character-

istic of any particular family of brittle-stars, but its persistence
and perfecting on the mature segments of the middle arm may
be a family or at least a generic character. Young spines arise
as straight rods, but soon become hooked or give rise to tooth-
like projections. They usually acquire a secondary straightness.
Young spines are solid, but may become hollow with increasing
size. Hollow spines undoubtedly indicate a specialized condition.

19. The tentacle-scales may arise as early as the under arm-plate and

before the lowest arm-spine, or their appearance may be delayed
until after several spines have been developed. Each tentacle-
scale is at first flat and rounded and is associated with the ventral,
distal margin of the side arm-plate. It may be homologous with
an arm-spine, but its position ventral to the lowest arm-spine counts
against that view. The evidence is as yet inconclusive whether
the absence of tentacle-scales is a primitive condition or not.
When two tentacle-scales are present, the second one arises in con-
nection with the under arm-plate a little later than the first one.
Its occurrence is almost certainly a specialized condition and it
can not be homologized with an arm-spine. A spine-like tentacle-
scale is certainly a modified form.

124 Papers from the Marine Biological Laboratory at Tortugas.

20. The podia arise as smooth, paired outgrowths of the radial water-

vessel just proximal to the terminal plate. Each one is associated
with a side arm-plate. They ordinarily remain smooth throughout
life, but in some brittle-stars they are rough, minutely or conspicu-
ously papillose. Such changes are obviously specializations and
such podia are often less retractile than usual.

2 1 . The number of arm-segments included in the disk increases with age

and such segments are commonly more or less modified by their
position.

22. The number of arms is determined when the metamorphosis occurs or

when the typical ophiuran form is assumed. New arms never
arise subsequently, except in schizogonous species, when they
develop in connection with the new disk formation. A 5-rayed
individual never becomes 6-rayed after its 5 arms are fully formed
nor does a 6-rayed individual ever become 5-rayed by autotomy
and resorption, as has sometimes been supposed. A 5-rayed or
6-rayed young brittle-star will remain such throughout life unless
schizogony occurs.

Museum of Comparative Zoology,
Cambridge, Mass., March, 1913.

LITERATURE CITED.

Bell, F. J. 1892. A contribution to the classification of Ophiuroids, etc. Proc. Zool.

Soc, London, pp. 175-183.
Fewkes, J. W. 1887. On the development of the calcareous plates of Amphitira.

Bull. M. C. Z., vol. 13, pp. 107-150.
Grave, C. 1898. Embryology of Ophiocoma echinata Agassiz. Prelim. Note. Johns

Hopkins Univ. Circ, vol. 18, pp. 6, 7.

1900. Ophiura brevispina. Mem. Biol. Lab., Johns Hopkins Univ., vol. 4>

No. 5, and also in Mem. Nat. Acad., Washington, vol. 8, pp. 83-100.
Jackson, R. T. 1899. Localized stages in development in plants and animals. Mem.

Boston Soc. Nat. Hist., vol. 5, pp. 89-153.
LuDWiG, H. 1878. Beitrage zur Anatomie der Ophiuren. Zeits. f. w. Zool., vol. 31,

pp. 346-394.

1879. Das Mundskelet der Asterien und Ophiuren, etc. Zeits. f. w. Zool., vol.

32, pp. 672-688.

1880. Neue Beitrage zur Anatomie der Ophiuren. Zeits. f. w. Zool., vol. 34,

pp. 333-365.

1881. Zur Entwicklungsgeschichte des Ophiurenskelettes. Zeits. f. w. Zool.,

vol. 36, pp. 181-200.

1899. Jugendformen von Ophiuren. Sitz. d. K. Preuss. Akad. der Wiss. Berlin,

vol. 14, pp. 210-235.

Lyman, T. 1882. Challenger Ophiuroidea. Reports on the Zoology of H. M. S. Chal-
lenger, vol. 5, pt. 14, pp. 1-386.

Mortensen, T. 1912. ijber Asteronyx loveni M. Tr. Zeits. f. w. Zool., vol. loi, pp.
264-289.

SiMROTH, H. 1876-77. Anatomie und Schizogonie der 0/>Aiac/« weM5 Sars. Zeits. f. w.
Zool., vol. 27, pp. 417-485; vol. 28, pp. 419-526.

SOLLAS, L B. J., and W. J. 1912. Lapworthura: A typical brittle-star of the Silurian Age,
etc. Phil. Trans. (B), vol. 202, pp. 213-232.

Growth-changes in Brittle Stars. 125

LIST OF PLATES.
EXPLANATION OF LETTERING.

A = Ala. P = Protapophysis. Zc = Zygocoele.

AH = Anterior hypapophysis. Pa = Parapophysis. Zg = Zygapophysis.

E = Epanapophysis. PH = Posterior hypapophysis. Zp = Zygophore.

Ep = Epapophyses. T = Taphrus. Zt = Zygotreme.

LAR = Lower alar ridge. UpAR = Upper alar ridge. Zy = Zygantrum.
MH = Median hypapophysis. Z = Zygosphene.

Plate i.
Ophiaciis savignyi (M. & T.).

Fig. I. A young individual, showing bases of 3 arms, and 3 very youthful arms, of which
the median is evidently the youngest. X40.

2. Disk of smallest individual seen, showing primary plates on older portion. X40.

3. Mouth of a young specimen, showing 2 old jaws and 3 pairs of old oral tentacles,

and 4 young jaws. X90.

4. Side view of a young jaw with 4 teeth and an old jaw with 5 ; 2 oral tentacles are

also shown. X90.

5. Torus of a young individual. X90.

6. Fourth segment from tip of arm, showing skeletal parts, as seen from below. X90.

7. A very young arm, with only one segment; seen from below. X90.

8. Vertebra of fifty-fifth segment from mouth in an arm with 70 joints; seen from

above. X40.

9. Vertebra of fiftieth segment of same arm; from above. X40.
ID. Vertebra of fortieth segment of same arm; from above. X40.

11. Vertebra of thirtieth segment of same arm; from above. X40,

12. Vertebra of tenth segment of arm with only 18 joints; from above. X90.

13. Vertebra of twentieth segment of arm with 70 joints; from above. X40.

14. Vertebra of tenth segment of same arm; from above. X40.

15. Same vertebra as fig. 14; seen from aboral end. X40.

16. Same vertebra; seen from below. X40.

17. Same vertebra; seen from adoral end. X40.

18. Same vertebra; seen from side, X40.

19. Vertebra of second segment of arm with 70 joints; seen from above. X40.

Plate 2.
Figs, i to 6. — Amphipholis squamata (Delle Chiaje).

Fig. I. One of the second pair of adambulacral plates, with its adjoining oral tentacle,
from a specimen 0.65 mm. across the disk; to show posterior (aboral) spinous
projection. X375-

2. One of the oral plates ("jaw") and its adjoining adoral plate, from a specimen

0.80 mm. across disk; to show origin of first 2 oral papillae, in connection with
them, but from separate centers of calcification. X375.

3. A jaw, and adjoining parts, from a specimen I mm. across disk; to show Amphiura-

stage of development. X90.

4. A jaw, and adjoining parts, from a specimen 1.35 mm. across disk; to show the

Amphipholis character attained.

5. Side view of torus and first tooth, from a specimen 0.65 mm. across disk. X375.

6. Side view of torus and first tooth, also from a specimen 0.65 mm. across disk but

somewhat older. X375.

Figs. 7 to 12. — Ophiolhrix angulata (Say).

Fig. 7. Vertebra of sixth segment from tip of arm; from the side. X90.

8. Typical vertebra; from the side. X40.

9. Same vertebra; from above. X40.

10. Sam.e vertebra; showing aboral surface. X40.

11. Same vertebra; from below. X40.

12. Same vertebra; showing adoral surface. X40.

126 Papers from the Marine Biological Laboratory at Tortugas.

Plate 3.
Figs, i to 10. — Lowest arm spine of Ophiothrix angulata {Say). Xioo.

Fig. I. From first (next to terminal plate) segment.

2. From second segment.

3. From third segment.

4. From fourth segment.

5. From sixth segment.

6. From middle portion of the arm (any segment).

7. From twelfth segment, counting from mouth.

8. From ninth segment from mouth.

9. From sixth segment from mouth.

10. From third segment from mouth.

These spines are all shown in the same position. Figure 6 shows the appearance
of the typical spine characteristic of the genus; figures 1-5 show its development,
and figures 7-10 its senescence. Spines, like figures 9 and 10, are very httle
different from the ordinary arm-spines. ^, . . \ r u- 1

11. Next to the lowest arm-spine of Amphipholis squamata (Delle Chiaje) from third

segment, counting from tip of arm. X375-

Figs. 12 to 14. — Lowest arm spine of Ophiactis savignyi (M. & T.). X375»

Fig. 12. From third segment from tip of arm.

13, From sixth segment.

14. From middle of arm (any segment).

Figs. 15 to 19. — Ophiothrix angulata {Say).

Fig. 15. Torus of specimen 2 mm. across disk. X90.

16. Torus of adult, 5 mm. across disk. X40. ...

17. Torus, teeth, and dental-papillae of young, 2 mm. across disk, seen trom the side.

18. Central plate, radial shields, and part of disk of a specimen 2 mm. across disk.

X40.

19. The same of a specimen 3 mm. across disk. X40.

CLARK

■SS ftt„'

€^

./^4^Xft^

J

ViUy:

J-^^ ^r

<y<

4

H

1^

fT\

^j'-X '■

^
i

4

,UnAR

-A

'"■ ; \i-';' -, LAB.

7 ^"^■^- '-^

Zo

■ 10

^..h^

,-f^

9

n

■M,

if w

n

12

13

15

z

A

AH

14^ '^

-■£p 2c

MH

--Z.

r

"^Mh

LAl?

AH

18

<r

CLARK

E -/

Z<i

/

?w

A\*

PLATE 2

"P AH

4 t .

-/VfW

^..^-p

— Zc

- T

7
J- 2-

^MH

AH

■-AH

-T.

■-Br

X'

.^

.y-^

I&

.TV---'

'\

' ' >A

/

—

- -^Zc

If
in

'2

AH
-MH

12

-?

MH

U

Ot^^

.-^-'

. "" . ,.Vi

-Ve

2P " - fff Mh

a ^^ ^ -^^

^

5

/■^

13

x'-'

17

rT)^ f^

19

VIII.

THE EARLY INFLUENCE OF THE SPERMATOZOAN UPON
THE CHARACTERS OF ECHINOID LARV^.

BY DAVID H. TENNENT,
Bryn Mawr College.

Eleven text-figures.

127

THE EARLY INFLUENCE OF THE SPERMATOZOAN UPON THE
CHARACTERS OF ECHINOID LARV/E.

By David H. Tennent.

The question of the influence of the spermatozoan upon the character of
the larva has been the basis of a vast amount of discussion among students
of embryology. From the time of the early disagreement between the animal-
culists and the ovists it has proved a fruitful subject for speculation.

In more recent times, since the evolution of the school of developmental
mechanics, various attempts have been made at an exact determination of
the r61e played by the spermatozoan or by some of its parts. It is conceded
that the characters resident in the male parent are transmitted to the off-
spring by the spermatozoan, either by the chromatin or by the cytoplasm,
or possibly by both. But there is no general agreement as to the time in
the individual history when "paternal" characters may become evident.
There is a fairly general opinion that the egg conditions development, that
it is responsible for at least the early developmental processes, and that the
sperm, although instigating development, places no paternal stamp on the
offspring until relatively late.

Boveri's (1895) classical experiment, criticized adversely by various
investigators, but confirmed by Herbst (1907), in which enucleated egg
fragments of one species of sea-urchin, fertilized by sperm of another species,
gave rise to larvae which were paternal with respect to the skeleton, showed
that in the absence of female nuclear material the sperm nuclear material
might assume the direction of some phases of development by the time of
the pluteus stage, but there has been little evidence, one way or the other,
regarding ^^ounger stages. Here the evidence has been negative in so many
instances that a positive generalization is often heard to the effect that in
development the influence of the spermatozoan is not shown in the character
of early embryos.

The discussion between Boveri and Driesch on this subject, in 1903-04,
resolved itself finally into a simple difference of opinion in regard to the
earlier or later time at which certain characters made their appearance in
Echinoid hybrids. Boveri (1895, 1903, 1904) presented evidence on the
form of the larvae, the skeleton, the number of chromatophores, the pigment
content of the chromatophores, the arrangement of the chromatophores,
the number of primary mesenchyme cells, and, under certain circumstances,
9 129

130 Papers from the Marine Biological Laboratory at Tortugas.

the size of the larvae. Driesch (1903) considered the same points, but by
the evidence neither was able to change essentially the belief of the other.

About five years ago the controversy was given a somewhat different
form as a result of the researches of Loeb (1908) on heterogeneous hybrid-
ization. In this work an echinoid egg, Strongylocentrotus , was fertilized
by molluscan sperm, Chlorostoma, and the resulting plutei were taken as
evidence that the inheritance was purely maternal, i. e., from the egg alone.
The reason for this was shown by Kupelwieser (1909), who worked out the
cytology of the cross Echinus microtuherculatus female by Mytilus gallo-
provincialis male and was able to show that true fertilization did not take
place, the male nucleus never fusing with the female nucleus, but instead
lying at one side of the mitotic figure during division, the larvae resulting
from the cross being, in effect, parthenogenetic.

Herbst (1906) obtained a closely similar result by a combination of
artificial parthenogenesis and fertilization, in which the larvae were held
to the maternal type. During segmentation the paternal nuclear material
lags and may fail to take part in the segmentation. Development in such
cases may be in reality maternal, but it is parthenogenetic.

Results again resembling these were obtained by Giinther Hertwig
(19 1 2), when he inseminated sea-urchin eggs with spermatozoa which had
been subjected to radium rays and found that here again the male nucleus
failed to take a normal part in the processes of fertilization and cleavage.

We must distinguish between the cases afforded by this work of Loeb,
Kupelwieser, and Hertwig, which deals in fact with a type of partheno-
genetic development, no fertilization taking place, and cases in which true
fertilization occurs, i. e., the union of an egg and sperm nucleus, although
here again we must distinguish between cases in which the union is only
temporary and those in which it is permanent. From work done by
Baltzer (1909, 1910), Herbst (1909, 1912), and Tennent (1912), evidence
has been gained which has proved that even though the two nuclei unite,
some of the chromosomes may be eliminated later.

In Godlewski's (1906) fertilization of the echinoid egg with crinoid sperm
the union of the nuclei seems to have been permanent. In his fertilization
of enucleated fragments of sea-urchin eggs with crinoid sperm, the young
stages (gastrulae, four cases) were larvae which followed the sea-urchin type.

Finally "partial fertilization," discussed by Boveri and Herbst, also rep-
resents a true fertilization, although the union between the nuclei may be
delayed until after segmentation has begun, and the resulting pluteus may
be in part maternal, in part hybrid in nature.

The consideration of these points makes it clear that we have no estab-
lished evidence of the influence of the sperm on the characters of stages
younger than the pluteus and that new observations based on favorable
material is desirable. I count myself fortunate in obtaining material of
this nature while on the expedition made by the Department of Marine
Biology to Montego Bay, Jamaica.

Early Influence of Spermatozoan Characters of Echinoid Larvce. 131

MATERIAL AND METHODS.

One of the abundant reef sea-urchins at Montego Bay is the primitive
form Cidaris trihiiloides. The living egg of Cidaris is exceedingly transparent
and the union of the male and female nuclei may be followed in it without
difficulty; it may be fertilized readily in the laboratory with its own sperm
or with the sperm of other sea-urchins abundant in the vicinity, among these
being Toxopneustes (Lytechimis) variegatus and Hipponoe {Tripneustes)
esculenta. These crosses are easily made. The eggs may be fertilized at
once after their removal from the gonad, without artificial aid, although as
a check against the possibility of chance fertilization with Cidaris sperm
they were kept for 2 hours before making the fertilizations. A portion of
the eggs was always kept as an unfertilized control.

Many investigators will recognize the convenience and advantage of
supplying developing larvae with an almost complete change of water from
time to time. I found it possible to do this by centrifuging the contents
of the finger bowls, which I used as aquaria, withdrawing the water from
above the larvae and replacing with fresh water. This method was used
only after the third day. This material was centrifuged only long enough
to throw the larvae down to the end of the tubes. The plutei treated in this
manner kept in better condition than those not centrifuged and thus not
given as complete a change of water. Owing to my limited time no attempt
was made to rear the larvae to metamorphosis. Every possible precaution
was taken to avoid chance fertilization or contamination of any nature of
the cultures.

Since the completion of the investigation is being delayed by the pressure
of teaching it seems best to present at this time the results that were ob-
tained with the living material, leaving the cytological analysis of fixed
material for a later time. The outline figures used in illustrating the paper
are from camera-lucida sketches made from living larvae.

INVESTIGATION.

In its normal development Cidaris proved of interest, (i) because of its
slowness of development when compared with Hipponoe or Toxopneustes;
(2) in the difference in the site of its mesenchyme formation; (3) in the
place of the appearance of the larval skeleton.

The anaphase of the first cleavage begins about 50 minutes after in-
semination. The cleavage is like that of other echinoids. The blastula
stage is reached in 16 to 18 hours, depending on the lot of eggs; gastrulation
begins in 20 to 23 hours; mesenchyme formation begins in 23 to 26 hours,
the mesenchyme cells arising from the inner end of the archenteron; chro-
matophores appear in about 44 hours; the enterocoele arises as a single
pouch in 44 to 50 hours; in 55 hours two enterocoeles may be seen, formed
by the division of the single vesicle; in 72 to 73 hours the beginning of the
formation of the skeleton may be noted. Even at this time, the beginning
of the fourth day, the body has not begun to assume the form of an echino-

132 Papers from the Marine Biological Laboratory at Tortugas.

pluteus, and it is not until the fifth day that the arms begin to push out.
These facts are of interest when compared directly with those regarding the
development of Toxopneustes.

The processes are here shown in parallel. The hours mentioned indi-
cate hours after fertilization.

Cidaris.

Hours.
Blastute (swimming) i6 to i8

Gastrulae (beginning) 20 to 23

Mesenchyme 23 to 26

Chromatophores 44

Skeleton (beginning) 72 to 73

Pluteus 120

Toxop}ieustes.

Hours.

Blastulae (swimming) 5.5

Mesenchyme 8

Gastrulae (beginning) 9

Chromatophores 15 to 16

Skeleton (beginning) 15 to 16

Pluteus 24

The earliest striking difference between the two forms is in the time
and place of mesenchyme formation, Cidaris in this respect resembling the
crinoids more closely than the echinoids.

No mesenchyme is formed until the archenteron has pushed well into
the blastocoele (fig. i). The cells then push out from the end and from
just behind the end of the archenteron. During the next few hours the

Fig. I. — Optical section, Cidaris gastrula, showing beginning of mesenchyme formation.
Fig. 2. — Optical section, later Cidaris gastrula. 41 hours.

23 hours.

migration of cells from the archenteron continues, while the cells within the
blastocoele increase in number by division. A network formed by fine
protoplasmic processes, as in other echinoids, is formed (fig. 2).

In Toxopneustes, it will be recalled, the micromere end of the blastula
becomes flattened and the primary mesenchyme cells push into the blasto-
coele before gastrulation begins (fig. 3). This mesenchyme is first scattered
throughout the blastocoele, the greater number of the cells remaining, how-

Early Influence of Spermatozoan Characters of Echinoid LarvcB. 133

ever, at the posterior pole. When the archenteron pushes in they lie as
a ring about it, later becoming aggregated in two groups, one at each side
of the base of the archenteron; in each group the characteristic skeletal
spicules arise. The secondary mesenchyme is given off from the end of the
archenteron after this has grown well toward the anterior end of the blastula
(fig. 4).

Fig. 3. — Optical section, Toxopneusles blastula, showing beginning of primary mesenchyme formation.

8 hours. Drawn from preserved material.
Fig. 4. — Optical section of Toxopneusles gastrula, showing formation of secondary mesenchyme and

of skeleton. Chromatophores in blastocoele and wall. 41 hours.
Fig. S. — Young Cidaris pluteus, showing character and position of early skeleton. 73 hours.
Fig. 6. — Cidaris pluteus. 6 days.

In the case of Cidaris, after the formation of the primary enterocoele and
the dorsal water-tube (50 hours), the gastrula begins to elongate and to
become broader at the posterior end. When the primary enterocoele sepa-
rates into right and left enterocoeles (the left remaining in connection with
the water-tube (55 hours), two ciliated bands make their appearance, the
posterior band about one-third the length of the body from the posterior
end of the body, the preoral band somewhat less than one-third the length

134 Papers from the Marine Biological Laboratory at Tortugas.

of the body from the anterior end. Between these bands the body now begins
to be somewhat concave. In general appearance at this time the larva
resembles that of Asterias more than it does that of an echinoid. At 66 hours
the bands are well developed and the region between the bands has become
still more concave.

Between this time and 72 hours a further increase in size and change
in form of the body is apparent, the posterior half becoming much broader
and the posterior end more rounded, while by reason of this change of form
the anal region has been shifted forward. During the period from 60 to 65
hours the oesophagus, stomach, and intestine differentiate from the archen-
teron. About 72 hours after fertilization the first indication of a larval
skeletal system may be noted in the triradiate spicules arising in groups of

Fig. 7. — Toxopneustes pluteus. 6 days.

mesenchyme cells in the broadest part of the body, about the level of the
anterior half of the stomach, and just beneath the postoral band of cilia.
The laterally directed axis is seen to be double and gives rise to a rod of
the latticed type (fig. 5). During the next day the body continues to
increase in size, the anterior and posterior ciliated bands become connected
about the edges of the ventral depression, and the postero-lateral (?) arms
begin to push out laterally from the body. A short medio-dorsal rod may
be seen developing in the skeletal system (fig. 6).

A brief inspection of a figure of a Toxopneustes pluteus of the same age,
6 days (fig. 7), is sufficient to convince one of the striking difference in form.
It is evident from a comparison of the development of these two forms that
if a cross between the two were possible valuable data might be afforded.

Early Influence of Spermatozoan Characters of Echinoid Larvce. 135

THE CIDARIS FEMALE BY TOXOPNEUSTES MALE CROSS.

The cross is easily made, the Cidaris eggs requiring no treatment in order
that fertilization may be accomplished. The eggs may be fertilized with Tox-
opneustes sperm immediately after their removal from the ovary, but
as a check against the possibility of chance
fertilization with Cidaris sperm they were kept
for 2 hours before inseminating with the foreign
sperm. More than 90 per cent of the eggs fer-
tilized; the typical fertilization membrane ap-
peared shortly after the addition of the sperm.
In this cross no difference from the characteristic
normal Cidaris development was noted before

, , . . _,,..,.. Fig. 8. — Optical section of Portion

the beginnmg gastrula stage. Ihe lertilization- of Toxopneustes cf xci<fam9 gas-

, . , 1-1 crula, showing primary mesenchyme

cleavage mterval was not lessened, as m the case teiisatbaseof archenteron. 24 hours,
of certain other echinoid crosses that I have de-
scribed. In point of time the appearance of the mesenchyme seemed
slightly hastened, although the gain in time is not great enough to war-
rant a general conclusion, since it is quite within the range of variation
between different lots of straight-fertilized eggs. My statement regard-
ing a possible slight hastening refers to comparative straight and cross
fertilization in the same lot of eggs. Even here I prefer to draw no con-
clusions, since I have made the comparison with only eight lots of eggs.
As to the place of mesenchyme formation there is no chance for either
personal equation in the observer or individual variation in the larvae to
play a part.

Figs. 9 and 10. — Optical section of Toxopneustes cf >< Cidaris 9 gastrula,
showing primary mesenchyme cells. 24 hours.

In the hybrids the mesenchyme arises from the sides and around the
base of the archenteron, close to its point of union with the blastopore
(figs. 8, 9, 10). This was true in every instance. The difference between

136 Papers from the Marine Biological Laboratory at Tortugas.

the normal and the hybrid was so striking that it needed no confirmation
by other observers, but purely as a matter of personal interest I asked four
other investigators at the laboratory to look at the two sets of living larvae
and describe their observations to me. The independent observation of
each of these investigators was a confirmation of my own.

The succeeding stages show considerable variation. In some the
growth of the archenteron ceases and the blastocoele becomes filled with a
mass of opaque cells (fig. 11). In others gastrulation continues. In a few

Fig. 1 1 . — optical section of Toxopneustes c? xCidaris 9
abnormal gastrula; blastocoele nearly filled with
opaque cells. 40 hours.

cases a small skeletal spicule was formed in the mass of mesenchyme cells
at the base of the archenteron. I was unable to keep the hybrid material
alive beyond the gastrula stage.

GENERAL DISCUSSION.

We have already noted Boveri's conclusions regarding the early influence
of the sperm on development. My results confirm his general thesis as
to an earlier influence than that commonly accepted, although my con-
clusion is based on evidence not afforded by his crosses.

In the nature of the case the results given by inseminations in which the
sperm nucleus does not unite with the egg nucleus are without direct value
in this connection. They may have an indirect value in showing that the
sperm cytoplasm, if it has become a part of the egg cytoplasm, has not
influenced the early development. As has already been indicated, the
larvffi obtained from such inseminations are essentially parthenogenetic.

The only investigation of which I know, which is closely comparable
to mine, is that of Godlewski (1907), and even here we are comparing the
reciprocal cross between forms whose early development is similar. In
Godlewski's cross the fertilization of echinoid eggs (primary mesenchyme

Early Influence of Spermatozoan Characters of Echinoid Larvce. 137

formed in the blastula stage) with crinoid sperm (crinoid mesenchyme
formed in the gastrula) resulted in larvae which followed the maternal
course of development; this was true even for the fertilized enucleated egg
fragment.

In my own work a cross made on material with similar differences of
development, namely, the Cidaris egg (mesenchyme formed in the gastrula),
with Toxopneustes or Hipponoe sperm (echinoids in which the primary
mesenchyme is formed before gastrulation), gave larvae in which the mesen-
chyme was formed at the beginning of gastrulation, the mesenchyme arising
at the sides and base of the archenteron.

We may ask this question: Is the foreign sperm better able to hasten
than to retard certain characteristic processes? It is not now possible to
answer this question, but it is one which should incite investigation.

The objection may be raised that as I was unable to keep the hybrid
larvae alive beyond the gastrula stage, the cells which I observed passing
into the blastocoele were abnormal and indicated an abnormal condition
of the larvae. Certainly they were not normal Cidaris larvae; they were
hybrids, and as hybrids they were normal. The larvae were not the formless
aggregations of cells obtained in some crosses, but proceeded on a perfectly
regular course of development. Even granting that when compared with
larvae derived from straight fertilized Cidaris eggs, they were not normal,
it must be admitted that the abnormality was in the direction of normal
Toxopneustes mesenchyme formation and also that the condition was
brought about by the use of the foreign sperm and that the result indicates
that the sperm is able to exert an influence on the character of development
at a stage earlier than that heretofore demonstrated.

Shearer, De Morgan, and Fuchs (191 1), as well as some other recent
workers on the hybridization of echinoids, have come to the conclusion
that the early larvae are of too variable a nature to afford any definite
evidence of parental influence. In order that I may not be misunderstood,
let me quote directly from the paper mentioned :

As the result of extensive investigation of the early larval history of our various crosses,
we have come to the conclusion that these are too variable to afford any definite evidence
of parental influence and especially is this true with regard to the skeleton, heretofore
considered the chief index of inheritance.

This conclusion was reached after the investigation of three species of
Echinus. The authors make the statement with regard to the forms upon
which they have worked and do not apply it to all cases. The investigation
of a sufficient number of properly chosen forms will convince any investigator
that parental influence may be indicated even in early larval stages, and
especially is this true with regard to the skeleton.

138 Papers from the Marine Biological Laboratory at Tortugas.

LITERATURE.

Baltzer, F. 1909. Die Chromosomen von Strongylocentrotus lividus uiid Echinus
jnicroiuberctilatus. Arch. f. Zellforsch., Bd. 2.

1910. Uber die Beziehung zwischen dam Chromatin und der Entwicklung.

Arch..f. Zellforsch., Bd. 5.
BovERi, Th. 1895. tJber die Befruchtungs- und Entwickelungsfahigkeit kernloser
Seeigeleier und die Moglichkeit ihrer Bastardirung. Arch. f. Entwick-
..elungsmech., Bd. 2.

1903. Uber den Einfluss der Samenzelle auf die Larvencharaktere der Echiniden.

Arch. f. Entwickelungsmech., Bd. 16.

1904. Noch ein Wort iiber Seeigelbastarde. Arch. f. Entwickelungsmech., Bd. 17.

Driesch, H. 1903. ijber Seeigelbastarde. Arch. f. Entwickelungsmech., Bd. 16.
GoDLEWSKi, E. 1906. Untersuchungen iiber die Bastardierung der Echiniden- und

Crinoidenfamilie. Arch. f. Entwickelungsmech., Bd. 20.
Herbst, C. 1906. Das Beherrschen des Hervortretens der miitterlichen Charaktere.
Arch. f. Entwickelungsmech., Bd. 22.

1907' Auf der Suche nach der Ursache der grosseren oder geringeren Ahnlichkeit

der Nachkommen mit einem der beiden Eltern. Arch. f. Entwick-
elungsmech., Bd. 24.

1909. Die cytologischen Grundlagen der Verschiebung der Vererbungsrichtung

nach der mutterlichen Seite. i. Mitteilung. Arch, f, Entwickelungs-
mech., Bd. 27.

1912. Ibid. 2. Mitteilung. Arch. f. Entwickelungsmech., Bd. 34.

LOEB, J. 1908. Uber die Natur der Bastardlarve zwischen dem Echinodermenei und

Molluskensamen. Arch. f. Entwickelungsmech., Bd. 26.
Kupelwieser, H. 1909. Entwicklungserregung bei Seeigeleirn durch Molluskensperma.

Arch. f. Entwickelungsmech,, Bd. 27.
Hertwig, G. 1912. Das Schicksal des mit Radium bestrahlten Spermachromatins im

Seeigelei. Arch. Mic. Anat., Bd. 79.
Shearer, DeMorgan, and Fuchs. 191 i. Preliminary notice on the experimental

hybridization of echinoids. Jour, of the Marine Biol. Association, vol. 9.
Tennent, D. H. 1912A. The behavior of the chromosomes in cross-fertilized echinoid
eggs. Jour. Morph., vol. 23.

1912B. Studies in cytology. II. Jour. Exp. Zool., vol. 12.

IX.

STUDIES OF JAMAICA ECHINI.

By ROBERT TRACY JACKSON.

Twenty-one figures, including one plate.

139

STUDIES OF JAMAICA ECHINI

By Robert Tracy Jackson.

It was my privilege to join the Carnegie Institution of Washington
expedition to Jamaica in the spring of 1912. While there, studies were
made of the later stages of development and the variation of a number of
species of regular Echini. All the material studied was collected at Montego
Bay, close along shore in shallow water of less than a fathom in depth. This
locality is very favorable for studying these animals, as abundant material
can be obtained, on the reefs or grass-covered bottom, by wading in water
usually less than 3 feet in depth. The observations are in extension of those
recently published in the " Phylogeny of the Echini " (Memoirs Boston
Society of Natural History, vol. 7, 1912). As only a few species are in-
volved, they are taken up systematically, giving under each the results of
the observations.

Elxidaris tribuloides (Lamarck).
This species is fairly common at Montego Bay, so that sufficient material
was available for laboratory work, but it was not so abundant as the other
species of regular Echini studied. It occurs in shallow water and is sur-
prisingly dull and sluggish, presenting none of the activity of motion seen
in Centrechinus or the more passive Tripneustes. In collecting Eucidaris,
one simply picks them up, and in no case observed did the animal cling
to the ground by its tube-feet. The adults at Montego Bay are rather
small for the species, not attaining the size of specimens seen from the
Bahamas. An average adult from Montego Bay measures about 30 to
35 mm. in diameter. Material from Port Antonio, Jamaica, also is small,
but is very abundant, as I had over 500 specimens collected at that locality.
Of a limited number of specimens of this species from the Bahamas, the
largest noted is 52 mm. in diameter. Only a few young specimens of
Eucidaris were found at Montego Bay. In individuals measuring up to
10 mm. in diameter the ocular plates are all exsert and the genital pores are
wanting. There was not sufificient young material to study the periods at
which the ocular plates travel in and reach the periproct. In the adult of
this species, typically (58 per cent), oculars V, I, IV are insert, as previously
ascertained from an examination of 849 specimens from several localities.
A specimen measuring 11 mm. in diameter has all five genital pores in
place, as have also all larger specimens. Whether the genital pores would

141

142 Papers from the Marine Biological Laboratory at Tortugas.

come in at this exact period of growth could only be known from obser-
vations on a larger series of specimens than was available. In living
specimens, opened when perfectly fresh, it was found that the teeth extend
only very slightly above the base of the foramen magnum. As this struc-
ture is very shallow in Eucidaris, the dorsal border of the tooth very
nearly coincides with the upper limits of the lantern, but the tooth does
not extend horizontally over the top of the lantern as in Strongylocentrotus
(Phylogeny of the Echini, p. 179, plate 5, figs, i, 6). The dental capsules,
which are the fleshy bags inclosing the base of the teeth, are small in Euci-
daris, not large and inflated as they are in Strongylocentrotus and in the
Echinidse.

The Cidaridae, the sole family representing the order Cidaroida, is the
only group of modern Echini that extends back into the Palaeozoic. The
Palaeozoic species, which belong to the genus Miocidaris, present only
slight differences from recent typical members of the order. As an ancient
group that has suffered comparatively little change in its whole geological
history the structure and development of the Cidaridae is of especial inter-
est. Something is known of the post-embryonic development of cidarids,
due principally to Loven's critical study of young Goniocidaris (Echinologica,
Stockholm, 1892), but further knowledge of the developing stages of this
primitive archaic type is much to be desired.

Centrechinus setosus (Leske;.
The most abundant sea-urchin at Montego Bay is Centrechinus^ setosus
(Leske), which abounds in countless profusion in shallow water on the
reefs. In strong distinction from the quiescent habits of Eucidaris, Cen-
trechinus is a most active animal, moving about freely and actively and
waving its long spines in a threatening fashion. The spines are so sharp
and cause such poisonous wounds that they are a constant menace ta
bathers, and specimens have to be collected with caution or serious results
will follow. The tube-feet have only the slightest hold on the sea-floor, so
that specimens can be picked up with long-handled forceps without any
perceptible resistance. In this Centrechinus differs markedly from Trip-
neustes, Strongylocentrotus, or Echinonietra, which on a rocky surface cling
tenaciously by the tube-feet. Young specimens of Centrechinus have spines
which are strongly banded ; white and very dark, nearly black, alternating.
The light and dark bands are of about equal width. The spines are long
in young as well as in adults, and a specimen 12 mm. in diameter has spines
36 mm. in length. A specimen 60 mm. in diameter has spines up to 150 mm.
in length. From this it occurs that a small specimen looks rather big with
spines extending radially, and a large specimen is menacing indeed when the
needle-like sharpness of these weapons is considered. Banded spines as a

i Centrechinus is a name which I gave (Phylogeny of the Echini, p. 27) to replace the name Diadema,
which in post-Linnsean usage is preoccupied for a crustacean. To replace an old, established name 18 unfor-
tunate, and some objection has been raised to my action. As Diadema has been in current use as a gener ic name
for sea-urchins. Gastropoda, and Lepidoptera, It seems that the rule of priority must be mamUmed and a new
generic name substituted for Diadema in these three groups.

Studies of Jamaica Echini. 145

character are retained until individuals are about 20 to 25 mm. in diameter^
after which they are typically a deep uniform black. As an exception, a
specimen 40 mm. in diameter was seen, in which all the spines were banded
as in youth. Frequently in immature specimens the older ventral spines are
banded when the dorsal spines, situated on the later-added and therefore
younger plates are quite black. In adults, no case of banded spines was
seen, but occasionally pure white spines are interspersed with the black, and
in a few cases observ^ed all the primary spines were white in the adult.
Whatever the color of the spines, the test is of a dark, uniform black in life.

In the family of the Centrechinidse, Centrostephanus longispinus (Phil-
lipi), from the Mediterranean, in adults has spines which are banded
white and light brown, the pattern recalling those of young Centr echinus.
In Echinothrix calamaris (Leske), from the Pacific, the spines are very
variable in adults, being all dark, purple or green, all white, or banded;
great diversity may occur in the spines of one individual. In other families
of Echini, banded spines occur occasionally as species characters; they
are a feature of Ccelopleurus maculatus A. Agassiz and Clark, and are common
in species of the Temnopleuridae.

In my previous studies of ocular plates and their variation in this
species, there was a limited number (no) of developing specimens, and
of these very few were really small. Of adults, 1,168 were examined from
various localities. With a good series of developing specimens from a
single locality, certain differences are brought out from the earlier limited
observations. There are also differences in the percentages of adult char-
acters from what obtained in the species as a whole from different localities.
In table 2 (page 156) are given the characters of developing series, and the
variations of adults of this and other species are considered. The aber-
rant variations are given in detail only under the consideration of the
species in the text.

As to the developing series of Centrechinus setosus, in 50 specimens
from 4 to 15 mm. in diameter, all the oculars are exsert (fig. i). In 103
specimens 15 to 20 mm. in diameter, 78 per cent have all oculars exsert,
4 per cent have ocular I insert, and 5 per cent have ocular V insert. In 2
per cent oculars I , V are insert, and in 7 per cent oculars I, V, IV have reached
the periproct. Of aberrant variants of this size there are 5 per cent, one
specimen having ocular IV only insert; one having I, IV insert; two having
V, IV insert; and one has oculars I, V, I\^ III insert. It is interesting to
see that at this youthful period, while a high percentage have oculars all
exsert, many stages of development as regards the oculars are presented by
some of the specimens. Of the next larger size, 20 to 25 mm., 270 speci-
mens, 51 per cent still have all the oculars exsert; i per cent have ocular
I insert; and 16 per cent have ocular V insert; 11 per cent have oculars I,
V, and 17 per cent have oculars I, V, IV insert. In one specimen, 0.4 per
cent, oculars I, V, IV, II are insert. There are eight aberrants, or 3 per
cent; of these, i has ocular IV insert; i has I, IV, and 6 have oculars \^ IV
insert.

144 Papers from the Marine Biological Laboratory at Tortngas.

In the next larger series, 25 to 30 mm. diameter (240 specimens), 21 per
cent have all oculars exsert; 3 per cent have ocular I, and 8 per cent have
ocular V alone insert ; 14 per cent have oculars I, V, and 45 per cent have
oculars, I, V, IV insert. As this last is the species character, it is of note
that this is the first stage at which this species feature has become a dominant
character. In i per cent, oculars I, V, IV, II, and in i percent all oculars
are insert. Seventeen specimens, or 7 per cent, have an aberrant arrange-
ment. Of these, 4 have ocular IV only insert; 11 have oculars V, IV, one
has I, V, II, and one has oculars I, V, IV, III insert. The occurrence of
I, V, II insert as an aberrant variant is of interest because this, as previously
shown, is a common aberrant when three oculars reach the periproct in the
Echinidse and Strongylocentrotidae, and occurs also in the Stomopneustidse.

1 have not found a case of I, V, II insert in the Cidaridse, and it is the only
case seen in Centrechinus setosus, although in this species, 1,398 cases of
I, V, IV insert have been observed.

In the next series of Centrechinus setosus, 30 to 35 mm. diameter, 245
specimens, 10 per cent have all oculars exsert; 2 per cent have I only;
and 12 per cent have V only insert; 5 per cent have I, V, and 63 per cent have
I, V, IV insert; 4 per cent have I, V, IV, II insert. Of aberrants, there are
4 per cent, these consisting of 8 specimens with oculars V, IV insert, and

2 with oculars I, V, IV, III insert.

The next series, measuring 35 to 40 mm. in diameter, 169 specimens,
has 5 per cent with all oculars exsert; 2 per cent with I only; and 3 per cent
with V only insert; 5 per cent have I, V, and 71 per cent have I, V, IV insert;
6 per cent have I , V, IV, 1 1 , and 2 per cent have all oculars insert. Aberrants
are 5 per cent; of these, two specimens have ocular IV alone insert; i has
I, IV; 4 have V, IV; and 2 have I, V, IV, III insert.

The series 40 to 50 mm. in diameter is considered the last of the devel-
oping series. Of this size, 162 specimens, 4 per cent have all oculars exsert,
I per cent have ocular I, and 0.6 percent have ocular V only inseit; 6 per
cent have oculars I, V, and 70 per cent I, V, IV insert; 9 per cent have
oculars I, V, IV, II, and 6 per cent have all oculars insert. Of aberrants
there are 3 per cent, one specimen having V, IV and 4 having I, V, IV, III
insert. The series 35 to 40 mm. and 40 to 50 mm. in diameter make a close
approach to the mature series as regards the species character of I, V, IV
insert, but these two younger series have more of the developing and fewer
of the progressive characters than are seen in the older series.

While Centrechinus setosus is profusely abundant at Montego Bay,
specimens do not attain as large a size as in Florida, from which locality
many individuals exceed 100 mm. in size. Of the adult series from Montego
Bay, 50 to 80 mm. in diameter, 162 specimens, one specimen, or 0.6 per cent,
had ocular V only, and one had oculars I, V insert, both of these specimens
are between 50 and 60 mm. in diameter; 73 per cent have I, V, IV insert,
the species character; 13 per cent are progressive variants with I, V, IV, II
insert and 10 per cent are extreme progressive variants with all oculars

Studies of Jamaica Echini. 145

insert; 3 per cent are aberrant variants, all five specimens having oculars
I, V, IV, III insert.

Considering the ocular-plate arrangement in Centrechinus from Montego
Bay as a whole, it is seen, as shown clearly in table 2 (page 156), that from
all exsert.the character of the young, there is a steady progressive series
with growth in the development or traveling in of ocular plates. Centre-
chinus is a member of the most primitive suborder, the Aulodonta, of the
Centrechinoida, and as such is worthy of special attention, for, as shown by
Professor Hyatt, primitive types in a group have a slow development as com-
pared with specialized types, and in character approach nearest to the next
lower series of their own phylum. The development of Centrechinus amply
bears out this important truth.

Of the total 1,401 specimens of Centrechinus setosus listed from Mon-
tego Bay, the aberrants are 59 in number, of which 8 have ocular IV only
insert; 3 have oculars I, IV insert; 32 have oculars V, IV insert; i has
oculars I, V, II insert, and 15 have oculars I, V, IV, III insert. It is note-
worthy that in the mature series (50 to 80 mm. diameter) the only aber-
rants that occurred were cases of oculars I, V, IV, III insert, thus indicating
that the other aberrants seen in immature individuals may be considered
as irregularities in the order of sequence of the coming in of ocular plates,
rather than as aberrants that would have retained the given character if
they had lived to grow up.

In specimens of Centrechinus up to 14 mm. in diameter (48 individuals),
all the genital plates are imperforate (fig. i), whereas in those 14.5 mm. and
all larger with very few exceptions, all genital plates have the genital pores.^
In comparison with this, the specialized Strongylocentrotus drohachiensis
attains its genital pores when between 5 and 10 mm. in diameter (Phylogeny
of the Echini, pp. 131, 170). Centrechinus has all oculars exsert up to 15
mm. diameter, and with intermediate stages still
has 4 per cent all exsert when 40 to 50 mm.
diameter. Strongylocentrotus drohachiensis, on
the contrary, has typically oculars all exsert only
in very young specimens and has only 29 per
cent with all exsert in specimens (51) between 2.5
and 4 mm. in diameter; further, it has only 4 per
cent with oculars all exsert in specimens (82)
between 4 and 5 mm. diameter. Centrechinus
first attains the species character of oculars I,
V, IV in'^ert as a dominant feature (45 per cent)
when it is 25 to 30 mm. in diameter. Strongylo- A

centrotus drohachiensis attains I, V insert, the '°" Montego'^Bay!" jamai^. x ^2!

, . ^ • J. c X. /-^ - Diameter ii mm. Oculars all

species character, as a dominant leature (52 per exsert. No genital pores, large
cent) when 5 to 10 mm. in diameter. In Centre- Se^^ youthful deleters.

» I regret to say that by error in the figure published of a young Centrechinus (Phylogeny of the Echini,
fig. 88, p. 106), genital pores are shown. The specimen has no pores and they were inadvertently drawn.

146 Papers from the Marine Biological Laboratory at Tortugas.

chinus when oculars begin to enter the periproct, the first to enter is V or I,
but in the Montego Bay series V is by far the more common/ they being in
the ratio of 102 of ocular V to 23 of ocular I. In this feature Centrechinus
makes an approach to the Cidaridae, in which group, when one ocular is
insert, it is almost always ocular V. As previously shown, when in the
Centrechinoida four oculars reach the periproct, it is typically I, V, IV, II,
and when any other combination exists it is nearly always I, V, IV, III.
This combination, while rare in the Centrechinoida as a whole, is rela-
tively more frequent in Centrechinus and is a typical feature or a frequent
variant in some species of the Cidaridae. In this feature also, therefore,
Centrechinus makes a certain approach to the Cidaroida.

In Centrechinus, when opened fresh and alive, it is found that the teeth
extend above the base of the foramen magnum about to the upper line of
the lantern, but the proximal base of the teeth does not extend horizontally
over the lantern as in the Echinidse, Strongylocentrotidae, and Echinome-
trldse. Further, the dental capsule, inclosing the base of the tooth, is small.
In the limits of the teeth dorsally and the small dental capsule, as well as
in the grooved teeth, Centrechinus makes a close approach to the characters
of the Cidaridae as represented by Eucidaris trihuloides.

M

3

Figs. 2-6. — Perignathic girdle in Aspidodiadema and its development in Centrechinus.
2. — Adult Aspidodiadema meijerei (Doderlein), Pailolo channel, Hawaiian Islands. The auricles are

separate spur-like styles.
3. — Centrechinus setosus (Leske), Montego Bay, Jamaica. Specimen 5 mm. diameter. Auricles are

separate styles, apophyses faint.
4- — The same. Specimen 10 mm. diameter. Auricles are more developed and arch over the ambulacra.
S. — The same. Specimen 25 mm. diameter. Auricles meet over the ambulacra in a slender arch.
6. — The same. Specimen 70 mm. diameter. Auricles developing a high plate-like cliaracter, apophyses

also are high.
7. — Centrechinus setosus, Bermuda. Specimen no mm. diameter. Auricles have developed a high

plate-like character; apophyses are high ridges.

Very little attention has been paid to the development of the perignathic
girdle excepting by Loven, who described it in Cidaris and Strongylocentrotus
(Echinologica) , I also have published a few observations on the develop-
ment and variation of the perignathic girdle. Centrechinus, while being a
primitive type in many respects, has in the adult a specialized perignathic
girdle having high auricles which meet in a broad plate-like arch over the
ambulacra and also high apophyses on the interambulacral areas (fig. 7).
Such being the character of the adult, it is of much interest to follow the
development of this, which is one of the most specialized perignathic girdles
known in Echini. When a specimen is 5 mm. in diameter, the apophyses
show only slight indications of development and the auricles stand up as
separate, blunt, spur-like processes on either side of the ambulacrum (fig. 3) .

1 In previous observations from various localities with limited material, ocular I predominated over V
numencally.

Studies of Jamaica Echini. 147

The upper ends of the auricles at this early stage are slightly flattened, but
do not arch over the ambulacrum at all. In a specimen 10 mm. in diameter,
the apophyses have developed somewhat, and the auricles are inclined over
the ambulacral area, but do not meet (fig. 4). In specimens 25 mm. in
diameter the apophyses have developed into distinct elevated ridges, and
the auricles have met over the ambulacral areas as slender arches (fig. 5).
From this point on to the adult the development of the perignathic girdle
is marked by the increase in height of the apophyses and concurrently the
increase in height and the lamellar expansion of the auricles (figs. 6, 7).

In seeking comparisons to these stages of Centrechinus in the adults of
related forms, it is found in Aspidodiadema meijerei (Doderlein), A. nico-
haricum Doderlein, and Dermatodiadema horridum A. Agassiz, representing
the family Aspidodiadematidse, that the auricles exist as small conical spurs
situated on either side of the ambulacral areas, and apophyses are not
developed (fig. 2). This condition corresponds with the first stage noted
in the development of Centrechinus (fig. 3). This relation of the perignathic
girdle is of much interest because in the Aspidodiadematidse the simple
ambulacral plates, the large primordial ambulacral plates on the peristome,
the large apical disk and simple spur-like auricles are all primitive characters,
indicating that the family is more primitive structurally than is the still
primitive family of the Centrechinidse. While primitive in many char-
acters, the Aspidodiadematidse is specialized in that all the oculars reach
the periproct, and Centrechinus, while belonging to the primitive suborder
Aulodonta, is specialized in its great development of the perignathic girdle.
This shows how primitive and specialized features may be combined in one
and the same type.

Mr. Agassiz (Panamic Echini, 1904, p. 59) says of the Aspidodiade-
matidse: "The auricles are most irregularly developed. They are either
wanting or mere projections, slightly raised." I have examined three rep-
resentative species and a number of specimens of this family, and find
no evidence that the auricles are irregularly developed. In adults they are
not wanting except where broken off. They are perfectly definite spur-
like projections, and of much interest as being structurally the most primi-
tive auricles known in any of the adult Centrechinoida. As the Cidaroida
have apophyses but no auricles, and as these structures make their first
known appearance in the Centrechinoida, the simplicity of their structure
in adults of the Aspidodiadematidse and in the young of Centrechinus is of
particular interest.

ToxoPNEUSTES VARIEGATUS (Lamarck).

A common species at Montego Bay is Toxopneiistes variegatus (Lamarck) ,
which is found abundantly on the grass-covered bottom in shallow water.
Specimens attain a large size, some measuring 70 mm. in diameter. Few
young specimens were found, though this is perhaps due to the fact that
special search was not attempted, from limitations of time. The test is

148 Papers from the Marine Biological Laboratory at Torttigas.

mottled green and white, with green spines as the prevailing color. Some
young specimens show on the green and white groundwork purple pri-
mary tubercles and purplish spines. A peculiar color-pattern was observed
as a relatively rare variation in the fact that in the interambulacral
areas of some specimens the green followed a zigzag pattern on a white
ground and extended thus from the apical disk adorally, but in no case
observed extended further than the ambitus. The zigzag green color-
pattern on a cream-white ground of a specimen measuring 45 mm. in diam-
eter is shown in fig. 8. The lower figure on the same plate is of a somewhat
larger specimen, measuring 48 mm. in diameter, in which the zigzag pattern
occurs only near the apical portion of the test. Both of these specimens
with others figured in this paper are now in the collections of the Museum
of Comparative Zoology, at Cambridge, Mass. This type of color-pattern
was quite new to Dr. Clark and Professor Tennent, both of whom had
worked on the species, and I had never seen it before, although I have
studied many specimens from different localities.

Being busy with other species, no attempt was made to collect a large
series of Toxopneustes variegatus for a study of ocular plates, but 400 speci-
mens are tabulated. These are of various sizes, but none were young and
they were not graded into sizes for comparison. Of this series, 2 specimens,
0.5 per cent, have ocular I only insert as arrested variants, and one specimen,
0.3 per cent, has ocular V only insert; 84 per cent have oculars I, V insert
(figs. 8, 9), which is the species character in all localities; 15 per cent are pro-
gressive variants with oculars I, V, IV insert. Four specimens, i per cent,
are aberrant variants and all of these have oculars I, V, II insert. The
ocular character is practically the same as I showed previously in this species
from several localities, but the Montego Bay material has a somewhat lower
percentage of I, V and corresponding higher percentage of I, V, IV insert.
My previous observations on i ,043 specimens gave 90 per cent I , V insert
and 8 per cent I, V, IV. Comparison of the tabulation of this species may
be made with that of the closely allied Toxopneustes atlanticus (A. Agassiz)
from Bermuda, in which I showed (Phylogeny of the Echini, p. 161) that
the mature series (45-77 mm. diameter) has a much higher percentage (28
to 29 per cent) of oculars I, V, IV insert than has Toxopneustes variegatus.
The Montego Bay experience in the cases of several species brought out the
fact clearly that for a close study of variation it is very desirable to tabulate
ocular plate characters in specimens from a single definite locality because
considerable difference in a given species may occur in different localities.

In Toxopneustes variegatus, when opened fresh, it is found that the proxi-
mal part of the testh extend horizontally over the top of the lantern and the
teeth are bent back on themselves in the same plane, also the dental capsules
which inclose the base of the teeth are large and bladder-like, in these char-
acters differing from Eucidaris and Centrechinus , but agreeing with Strongy-
locentrotiis as I previously described it. In a young specimen 12 mm. in
diameter, the auricles are still separate, but in larger specimens they are

JACKSON

Studies of Jamaica Echini. 149

joined in anarch over the ambulacral areas. In large adults the apophyses
are developed as a moderate ridge and auricles are joined in a delicate arch
which never becomes heavy, as in large specimens of Tripneustes.

Tripneustes esculentus (Leske).

One of the abundant species at Montego Bay is Tripneustes esculentus
(Leske). It occurs on grass bottom and also on the reefs in shoal water,
and is a strikingly handsome sea-urchin when alive, with its white spines
against the dark purplish groundwork of the fleshy covering of the test.
The larger individuals live out in the open in full view, but young specimens
are found only in crevices or on the under side of rocks. The young are
not common, and although sought for diligently, no very young specimens
were found and only a limited number of the youngest seen were obtained.
Perhaps some other time of year, as summer or autumn, would be more
favorable for young material of this species. In Tripneustes esculentus from
this locality the ocular plates continued to travel in or enter the periproct,
from the youngest to the largest series of specimens observed, as shown by
the tabulation. For this reason, as regards the feature of ocular plate
development, what is called the developing series for this species in the table
is arbitrarily drawn at a maximum of 70 mm.

Taking up the developing series: Ten specimens of 20 to 25 mm. diam-
eter have 50 per cent with ocular I only insert, and 50 per cent with oculars
I, V insert. Younger specimens would doubtless have all oculars exsert as a
developing stage. The next size, 25 to 30 mm. diameter, 21 specimens,
has 24 per cent with ocular I only insert and 76 per cent with I, V insert.
The relations of these two younger series shows that it is a period of rapid
development. The next series, 30 to 40 mm. diameter, 60 specimens, has
12 per cent with I only insert; 80 percent with I, V insert; 2 per cent with
I , V, IV and 2 per cent with I , V, IV, 1 1 insert. Three specimens, 5 per cent,
are aberrant with I, V, II insert. This stage, while holding the I only insert
as a developing character, has attained the I, V; the I, V, IV; and the I, \',
IV, II insert; which are each respectively the feature of the species as a
character in one or more of the several localities from which material was
studied, but the relative proportions of the three characters differ from
that of any known adult series.

The series 40 to 50 mm. diameter, 42 specimens, has lost the I insert as
a youthful character; 90 per cent have oculars I, V insert; 2 per cent have
I, V, IV, and 2 per cent have I, V, IV, II insert. Five per cent are aberrant,
one of the two specimens having I , V, 1 1 insert and the other \', 1 1 insert. This
stage has the highest percentage of I, V insert of any age of material from
Montego Bay. The series 50 to 70 mm. diameter, 34 specimens, has 53
per cent with oculars I, V insert; 18 per cent with I, V, IV; and 6 per cent
with I, V, IV, II insert. Aberrants are 24 per cent, eight specimens, all of
which have oculars I, V, II insert. The developing series in its later phases
may be compared broadly with material from Bermuda or other localities

150 Papers from the Marine Biological Laboratory at Tortugas.

for the species in which oculars I, V insert is the adult local character, as
it is the youthful character in these immature stages from Montego Bay.

The next larger series, 70 to 90 mm. diameter, 125 specimens, has one
specimen with ocular I insert as an arrested variant; 30 per cent have
oculars I, V insert. The latter, which was the leading character in the later
developing stages, has now passed into the phase of an arrested variant.
34 per cent have oculars I, V, IV insert, and this is the dominant character of
this age. As a phase it can be compared with a series of adults from the

13

14

Figs. 10-14. — Normal ocular plate arrangement in Tripneustes esculenlus (Leske),
Montego Bay, Jamaica. All figures nearly X 2.
10. — Diameter 99 mm. Ocular I only insert, an extreme arrested variant.
II. — Diameter 114 mm. Oculars I, V insert, an arrested variant for the locality.
12. — Diameter no mm. Oculars I, V, IV insert, an arrested variant for the locality.
13- — Diameter ns mm. Oculars I, V, IV, II insert, the typical character for the locality.
14. — Diameter 106 mm. All oculars insert, a progressive variant.

Bahamas in which locality I, V, IV insert is the typical character. Twenty-
eight per cent have oculars I, V, IV, II insert, which at this stage may be
considered a progressive character as it is at all other known localities for
the species. In 8 per cent the ocular arrangement is aberrant, all of the
ten specimens having oculars I, V, II insert.

The next series, 90 to no mm. diameter, 200 specimens, has one speci-
men with ocular I only insert as an arrested variant (fig. 10), and one
specimen with ocular V alone insert. For ocular V only to be insert is a
rare variation in Tripneustes and is unusual in the family Echinida?. In 22
per cent oculars I, V are insert. This, which was the character of the later
developing series and is the character of adults at Bermuda and Florida,
has dropped considerably in its percentage as an arrested variant. In 24
per cent oculars I, V, IV are insert. This character for Montego Bay has

Studies of Jamaica Echini.

151

at this stage dropped into the phase of an arrested variant, though it was
the dominant character in the series 70 to 90 mm. in diameter and is the
adult character in a series from the Bahamas. Forty per cent have oculars
I, V, IV, II insert. This is the first series in which this character is a domi-
nant feature at Montego Bay, and for all earlier stages in this locality and
for all other localities known it is a progressive character. Two per cent
have all oculars insert as a progressive variant (fig. 14). It is somewhat
striking that with four oculars insert as a dominant character, there should
not be more specimens with all oculars insert. In this series, 25 specimens,
13 per cent, have an aberrant arrangement of ocular plates. Of these
aberrants, 22 have oculars I, V, II insert; two have I, V, IV, III insert (fig.
16), a very rare variant in the Echinidae when four plates reach the peri-
proct, but although rare here, it is a common character in the Cidaridae
and a not rare variant in Centrechinus setosus. One aberrant specimen of

TV if

IL W M

15

16

17

Figs. 15-17. — Aberrant ocular plate arrangement in Tripneustes esculentus (Leske). Montego
Bay, Jamaica. All figures nearly X 2.

Fig. is. — Diameter lis mm. Oculars I, IV, II insert, a common aberrant variant, a right-lianded equiva-
lent of the normal character I. V, IV insert. ..,.■.. ^i, 1 i,->„

Fig. 16.— Diameter 106 mm. Oculars I. V, IV. Ill msert, a rare aberrant variant, but the usual one when
four oculars but not I, V, IV, II are insert. _ . . ., t. j j

Fig 17 —Diameter 102 mm. Oculars I, V, II, III insert, a unique aberrant variant, a right-handed equiva-
"lentof I. V. IV, Illinsert.

Tripneustes has oculars I, V, II, III insert (fig. 17). This, which is a
right-handed equivalent of I, V, IV, III (just as I, V, II is a right-handed
equivalent of I, V, IV) is a unique variant for Echini as far as my experi-
ence goes. In the Centrechinoida, as tabulated in my preceding memoir
and in the present paper, 619 specimens are recorded, in which four plates
reach the periproct. Of these, in 554 specimens (89.5 per cent) it is oculars
I, V, IV, II the bivium and posterior pair of the trivium, in 64 specimens
(10.3 per cent) it is the aberrant arrangement I, V, IV, III, and in the one
specimen cited (0.16 per cent) it is I, V, II, III. (Compare figs. 13, 16, 17.)
The series containing the largest specimens, no to 132 mm. diameter,
82 specimens, has 13 per cent with oculars I, V insert as arrested variants
(fig. 11); 24 per cent with I, V, IV insert as arrested variants (fig. 12) ; and
43 per cent with I, V, IV, II insert as the local character of the species (fig.
13). Five per cent have all oculars insert as a progressive variation, and 15
per cent, 12 specimens, are aberrant in ocular arrangement. Of these

152 Papers from the Marine Biological Laboratory at Tortugas.

aberrants, 11 specimens have oculars I, V, II insert (fig. 15), and one has
V, IV, II insert. In this specimen genitals 5 and i are fused, mechanically
shutting out ocular I from access to the periproct.

The series of Tripneiistes esculentus from Montego Bay is very interest-
ing from the point of view of ocular development on several accounts. The
specimens show a direct progressive increase in the number of oculars insert
and the relative percentages of the same from the youngest to the largest
series observed. In specimens fully matured, as regards this character,
and including from 90 to 132 mm. in diameter, they have as the dominant
character (40 to 43 per cent) the oculars I, V, IV, II insert. This is the only
locality known for this species, and it is the only recent species known'
in which four oculars reach the periproct as a typical character. In pre-
vious studies of Tripneustes esculentus, many specimens were tabulated from
several localities, but without detailed measurements of the size of the
specimens. As the material was from various sources, it is impossible to
go over it again to ascertain the sizes but it is fair to say that the material
was practically all adults. With this qualification it is interesting to com-
pare the tabulations of Tripneustes esculentus from the several localities from
which there was sufficient material to give any reasonable basis for con-
sideration.

T

ABLE

I.

"0 v

00

Locality.

Ocular I

insert; v,

IV, 11, III

exsert.

Ocular V

insert; i,

rv, II, III

exsert.

Oculars I,
V insert;

IV, II, III

exsert.

Oculars i,
V, IV insert;
II, III ex-
sert.

Oculars i,

V, IV, II

insert; iii
exsert.

All oculars
insert.

Arrange-
ment of
oculars

aberrant.

p. ct.

No.

p. ct.

No.

p. ct.

No.

p. ct. No.

p. ct.

No.

p. ct.

No.

p. ct.

No.

193
112
160
282

Bermuda. .
Florida

2

3

61
46
22
19

117
52

35

54

35
29
46

24

67
33
73
67

2
12
27
40

3

13
43

114

0.5

I

I
12

3
13

2
13

S
37

I

0.6

0.4

I

I
I

1 2

i ^

3

Jamaica!. .

0.4

I

> 90 to 123 mm. diameter.

As shown in table i, Tripneustes from Bermuda is characterized strongly,
61 per cent, by having oculars I, V insert, with 35 per cent I, V, IV insert as
progressive variants, also in 2 per cent I, V, IV, II, and 0.5 per cent all
oculars one insert as progressive variants. The aberrants are few, i per
cent, one specimen having I, V, II and one having V, IV, II insert. This
Bermuda lot is the most primitive of the species that I know of as regards
ocular arrangement in adults.

Tripneustes from Florida has 46 per cent with oculars I, V insert, which
is still the dominant character, but with a lower percentage than at Bermuda.
As progressive variants, 29 per cent have I, V, IV insert, and 12 per cent
have I, V, IV, II, The aberrants are 12 per cent, all being cases of I, V, II
insert. Florida specimens, while having oculars I , V insert as the character,
have a lower percentage of this character and a higher percentage of progres-

> In Acrosalenia pseudodecorata Cotteau, from the Bathonien of France, oculars I, V, IV, II are said to
be insert as a species character (Phylogeny of the Echini, p. 112).

Studies of Jamaica Echini. 153

sive variants than Bermuda material, hence it is structurally an advance
on that and a step towards the next higher advance as represented in the
series from the Bahamas.

Material from the Bahamas collected by C. J. Maynard has only 22 per
cent of I , V insert. This is not the character of the locality and has therefore
passed into the phase of an arrested variant. There are 46 per cent with
oculars I,V, IV insert. This is the dominant character of the Bahama mate-
rial, while it would be a progressive variant for Bermuda and Florida, and
further, would be an arrested variant for the Montego Bay series. In the
Bahamas, 27 per cent have I, V, IV, II insert, a common progressive variant,
and 2 per cent have all oculars insert. The aberrants are relatively few,
3 per cent, four of the aberrants having I, V, II insert and one having
I, V, IV, III insert. Structurally, in its percentages of ocular arrangement,
the Bahama material is intermediate between the Florida and Montego Bay
series.

The Montego Bay material, considering only the two largest-sized series
tabulated on page 156 and treating them as a single series of 90 to 132 mm.
diameter, has 19 per cent with oculars I, V insert as a moderately common
arrested variant, but with the lowest percentage of any of the localities
given. In 24 per cent oculars I, V, IV are insert, which, for this locality, is an
arrested variant, though it is a progressive variant for Bermuda and Florida,
and a dominant character for the Bahamas. In 40 per cent, oculars I, V,
IV, II are insert, the dominant character for this locality, though it is a
progressive variant for all the other localities; 3 per cent have all the
oculars insert, which as a progressive variant, is only a slight advance on
that existing in the other regions cited. The aberrants, 13 per cent, have
already been considered above, and do not need further mention excepting
that it is interesting that most of them are cases of I, V, II insert, as in other
localities.

In all the localities given Tripneustes esculentus is an abundant form,
and in all attains a large size. The specimens from Bermuda were collected
for me and averaged very large, the largest measuring 145 mm. in diameter,
yet this locality has the lowest ocular index. The cause of the progressiv^e
evolutionary series as marked by the higher and higher percentages of ocular
plates reaching the periproct is unknown, but it is a fact that it exists and
points to the desirability of studying large series of specimens from different
localities.

In Tripneustes esculentus, when opened alive, it is seen that the teeth
extend above the top of the lantern and horizontally over the top of the
same. The proximal end of the tooth is curved back on itself and is grooved,
the keeled character developing as one passes orally along the tooth. The
dental capsules are relatively large, all this structure coinciding with that
previously described in Strongylocentrotus drobachiensis. In a young
specimen of Tripneustes, 24 mm. in diameter, the auricles are slender, but
already meet in an arch over the ambulacra. In large adults the auricles

154 Papers from the Marine Biological Laboratory at Tortiigas.

are quite massive and are fused over the ambulacra in a long median
symphysis. They may attain a height of 14 mm. or more. The apophyses
are also relatively high and massive in large specimens.

ECHINOMETRA LUCUNTER (Linn6).

Close in to shore on the coral rock at Montego Bay and in very shallow-
water Echinometra lucunter (Linn6) occurs in abundance. Not only adults
but individuals of all sizes, down to very young, live in the same exposed
situation, where they get the full beat of the waves on shore. This species
clings tenaciously to the rock and has to be pulled off from the rock with
some effort. The ocular development of this species is very interesting,
but unfortunately this was not appreciated until it was too late to gather
more material and only a limited series was collected. These were supple-
mented by a few specimens collected at the same time by Mr. G. M. Gray.
A large series could easily be obtained in this locality. As Echinometra
lucunter is more or less elliptical in horizontal outline, being elongated in
the plane of the axis of interambulacrum 3 and ambulacrum I, the measure-
ments of sizes are given in length instead of in diameter, as in the other
species considered.

Considering the developing series of Echinometra, a set of 34 specimens
measuring 4.5 to 15 mm. in length have all the oculars exsert. Of this
series, those from 4.5 to 10 mm. in length (26 specimens) were all wanting
in genital pores, but these pores exist in all the larger sizes examined. The
series 15 to 20 mm. in length, 23 specimens, has 91 per cent with all oculars
exsert and 9 per cent with ocular V insert. While ocular I is the first to
enter the periproct in the Echinidae and Strongylocentrotidae, it is ocular
V that typically enters the periproct first in the Arbaciidse and the Echino-
metridae, thus indicating family distinctions in ocular development.

In the series 20 to 25 mm. in length, 58 specimens, in 71 per cent oculars
are all exsert and 29 per cent have ocular V insert. In the 25 to 30 mm.
series, no specimens, 34 per cent have oculars all exsert; 2 per cent have
ocular I insert. These are the only cases with ocular I only insert found in
Montego Bay material and it is a rare variation in the genus and family, as
previously shown. In 54 per cent ocular V is insert. This series is the first
stage in growth in which the typical species character of V insert is a dom-
inant feature. In 10 per cent, oculars V, I are insert and in the limited
number of specimens this is the first stage at which this character appeared.
With more material some specimens would probably be found having I, V
insert at an earlier stage of growth. This series is considered the last of
the developing stages, but the line as regards ocular development is arbitrary
because, as seen in the later series, oculars continue to come in with increas-
ing size of the individuals.

Of the series 30 to 40 mm. in length, 175 specimens, 17 per cent have
oculars all exsert as arrested variants, 58 per cent have ocular V alone
insert as the typical species character, and 25 per cent are progressive

Studies of Jamaica Echini.

155

variants with oculars V, I insert. The percentage of V insert at this age is
quite near that of the next larger series, but there are more arrested and fewer
progressive variants than in the older series. The series 40 to 52 mm. in
length, 103 specimens, includes the largest specimens collected at Montego
Bay. From other localities specimens may attain a length of 90 or more
millimeters. Of this series, 7 per cent have all oculars exsert as arrested
variants ; 66 per cent have ocular V insert as the maximum development of
the species character in the locality; 26 per cent have oculars V, I insert as
progressive variants. One specimen, i per cent, has oculars V, II insert
as an aberrant variant, and this was the only aberrant found in the whole
series of 503 specimens collected at Montego Bay.

Considering the ocular arrangement of Echinometra lucunter at Montego
Bay as a whole: It is a desirable type to study because its species character
is to have one ocular insert, and this is a very unusual feature in Echini,
otherwise occurring only in the recent Saleniidae and in Echinus magel-
lanicus Philippi from South America. Other Echini, both recent and
fossil, are characterized as the species feature of adults by having all oculars
exsert, or two oculars insert, or three or more insert, usually three or five.
Five insert is a not unusual feature in both living and fossil types as
far back as the Palaeozoic. In Echinometra lucunter it is seen that oculars
are all exsert up to a comparatively large size, 15 mm., after which oculars
travel in gradually with increasing size, the highest percentages of insertness
being in full-grown specimens. Comparing the present tabulation of this
species with that previously published (Phylogeny of the Echini, p. 163),
it is seen in both that ocular V insert is the species character, though in
Bermuda material oculars V, I insert is nearly as frequent. In both lots
ocular I insert is a rare feature, and aberrant ocular arrangement is rare.
As material previously tabulated was not graded by size, no comparison
can be made in this respect.

18

Figs. 18-21. — Development of the perignathic girdle in Echinometra lucunter (Linne).
Fig. 18. — Young specimen from Montego Bay. Jamaica. Specimen s mm. long. Auricles are erect separate

styles.
Fig. 19. — The same. Specimen 8 mm. long. Auricles arched over the ambulacrum but separate, apophyses

developing.
Fig. 20. — The same. Specimen 15 mm. long. Auricles joined in suture over the ambulacrum.
Fig. 21. — Large adult from Bermuda. Specimen 6s mm. long. Auricles produced vertically as high spoon-like

plates, apophyses high ridges.

The teeth of Echinometra lucunter extend horizontally over the top of
the lantern and the proximal end of the tooth is bent back on itself and in
the same plane, as previously described in Strongylocentrotus. The proximal
part of the tooth is grooved, the keel developing as one passes adorally;

156

Papers from the Marine Biological Laboratory at Tortugas.

^

1

s

K)

O .S-3

<^

s -^2 =

Arrange-
ment of
oculars

aberrant.

1

; lOOO t^ O Oi "5

Ifl

■*

: fONoo

CO

OWN

W P« M

5

-

"

d

, U5 CO t» ^"5 ro

:

CO

M

00 com

H l-l

•

H

1

M

■*o

HI

-o

:■*•*

00

u

6.

M

«o

o

n uj

:

Oculars i,

V, IV, II

insert; in
exsert.

i

M COO M w

1-1 M H

O

H

M H N

•*

ioo\>o

cot^ CO

la

O' iH ■*o o.

CO
1-1

P) ««

00 O CO
P« ■*■*

Oculars i,

V, IV insert;

II, III

exsert.

i

r-o Oi •*o (»)

1

M

00

M H\0 00

P) t~ O

a
o

4J

o

d

t~i~in t<5M o

:

CO

v>

N PlOO ■

H

j ;

ION N ;

Oculars i,
V insert;

IV, II, III

exsert.

6

« O ro roOiO
CO ro w H

"

CO
CO

■OOOOOOOO Ifl

M ^CO W PI

M

t^CO M

CO"* l-l

H

a

i

«

«

CO r^
■*p<

o

d

« M -i-WlONO

00

OO O O lO
Wt^OO Olio

O M CO

COM M

o

\n\0
fi p)

PI 00

o-o

Ocular V
insert; I,

IV, II, III

exsert.

1

lO to Ol Ov "> M N
■<tM M O

M

w

M •

M

PI t^o

a

o

M

tJ
d

lo-ooo « fod

d

CO

d

d :

OvOv*

00 \o
U50

Ocular I
insert; v,

IV, II, III

exsert.

6

■fS-vO ■VfON

CO

n

\n\nt.

M

MH :

PI

P)

p)

-

d

M-w r<5M « M

d

o ■*?«

WW M

00 lo ;

d d ;

P«

2

6
2

O ooo w looOO
»O00 rolOO)

=0
CO

CO N -^CO

CO

O 1^

CO

CO

•4J

o
d

OOO M M o «0 >*

O «^>0« M

•1

O W M Tj-

o a t~ CO

M

Size of specimens.

.loowoinoo o !

. QiHN«nro^ir> 00

s go :

Q ■<tio o w o w o o I

H N M c«3 fO M- >r>

moooo COM .inoi/)o OM

NcO'*iOt~ 0>"ro -MPlNCO Ttio

M H <-> O

oioooo ooo §"?
NP»ro->*io t^ Ov " nJ^iooifl OO
M M N M ro •^t

Speci-
mens
ob-
served.

OrOOOWOiP) N

M N « N M 1-1 M

O 11 O « •* irt O « «i-c000O Ifl CO
MNvO'^co N ooo coNinw t~o

03
U

1

c

o

a.

"33
>

Q

s
a

CO

i

S
a

1

>_ ■ '

1 s

II

w >

t ^

J

1

Tripneustes esculentus:
DevelooinE series -i

'V

d
]

•z

c
o
■a3

a

s

•2 1

s i

s c

1

>

3

t
1

c

i

3

i
i

>

D

5

Total
No. of
speci-
mens.

M

%

O

o o t-

to eo
p> t^

PJ M

CO

o

lO

Studies of Jamaica Echini. 157

the dental capsules are large, embracing the base of the tooth as in the
Echinidse and Strongylocentrotidae. The perignathic girdle of Echino-
metra hicunter in the adult is highly specialized, having relatively high ridge-
like apophyses and extravagantly developed auricles which meet over the
ambulacra and are produced dorsally as spoon-shaped plates with a long
median symphysis (fig. 21). In young specimens, 5 mm. in length, the
auricles exist as simple erect styles situated on either side of the ambulacra
but not arched over the same (fig. 18). In specimens 10 mm. in length,
the auricles arch over the ambulacra, but are still separate, not meeting in
median suture (fig. 19). In a specimen 15 mm. in length, the auricles have
met over the ambulacra and are joined by median suture, and the apophyses
exist as moderately developed ridges (fig. 20). From here on to the adult
condition the apophyses increase in height and the auricles gradually take
on an extensive vertical enlargement of their upper border (fig. 21). In
Echinometra viridis A. Agassiz, E. ohlonga Blainville, and E. vanhrunti A.
Agassiz, in adults, the auricles are relatively slender, but are united in
suture over the ambulacra; there is no vertical extension in a spoon-like
fashion, however, and the auricles closely resemble the condition of E.
hicunter when about 15 mm. in length.

In Echinometra mathei Blainville, in a young specimen 8 mm. long from
Mauritius, the auricles are still separate, but in the adult the auricles are
joined in suture and with moderately developed spoon-like processes, but
not as high processes as are those in adult E. lucunter. In Heterocentrotus
trigonarius (Lamarck), from Mauritius, a young individual 9 mm. in
length, has the auricles still separate, but in the adult of this species
auricles are arched and united over the ambulacra and with moderately
developed spoon-like processes.

SUMMARY AND CONCLUSIONS.

Montego Bay, Jamaica, is an excellent locality for studying Echini on
account of the number of species there available and the abundance of
material. It is an ideal place for a student who wishes to supplement
studies on northern species by observations on tropical forms.

Some 2,878 specimens of regular Echini were collected and studied, but
in no case did any specimen show a departure from the typical pentamerous
system. This is mentioned as in previous studies it was found that, on
the average, completely or partially trimerous, tetramerous, or hexamerous
specimens occurred a little oftener than once in a thousand specimens.

In the " Phylogeny of the Echini," page 91, it is stated that "all the evi-
dence goes to show that the full number of oculars that are to become
insert are developed early in the life of the individual and apparently later
no change in this respect takes place." It was assumed that in Echini
specimens of about half the mature size could be accepted as showing the
mature characters as regards oculars, and the tables of ocular plate vari-
ation, while based principally on adults, were made up on the basis of con-

158 Papers from the Marine Biological Laboratory at Tortugas.

sidering specimens more than half-grown as being fully developed in this
respect. This view was based largely on studies of Strongylocentrotus dro-
bachiensis, and in this species the view was fully justified. It is doubtless
true of many species, perhaps most species of Echini, that specimens half-
grown have practically the full character as regards ocular development.
It is not entirely true, however, of all species, for as shown in this paper,
in Centrechinus setosus, Tripneustes esculentus, and Echinometra lucunter
(table 2), ocular plates continue to come in or enter the periproct as a
developing character until nearly or quite grown; therefore, as a result in
these species, the largest numbers in percentage of oculars insert may be
found in the series of largest specimens. This is not true of selected indi-
viduals. That is, the largest specimen does not necessarily have more
oculars insert than a smaller individual, but taking a large series of indi-
viduals of these species, as shown in the table, the average of each size up
to the largest has more oculars insert than any of the preceding sizes. This
late coming in of oculars in some species is of interest as showing the mobility
in adjustment of the test to allow these changes to take place. It is also
of interest in support of the view urged so much by Professor Hyatt, that
changes are taking place constantly throughout the life of the individual.

Centrechinus setosus is slow and very gradual in developing its characters.
It is slow in acquiring genital pores and very slow relatively in having
ocular plates enter the periproct. Up to 15 mm. in diameter no oculars
have yet entered the periproct. In the series 20 to 25 mm. in diameter, a
little more than half of the specimens still have all oculars exsert. When
35 to 40 mm. in diameter, which is about half-grown for the locality, the
species character of I, V, IV insert is developed to a high point, 71 per cent,
practically the maximum percentage; but at this age there are more speci-
mens representing developing characters and fewer representing progressive
characters than in older series. The maximum of the species character and
also the maximum of the progressive characters of I, V, IV, II, and all oculars
insert, is seen in the largest specimens of 50 to 80 mm. diameter. The
slow rate of development of Centrechinus as here shown (p. 156) is graphically
brought out by comparing it with that of the development of Strongylo-
centrotus drobachiensis given in the " Phylogeny of the Echini," page 142.
Centrechinus is a relatively primitive (and geologically old) genus in the
lowest suborder (the Aulodonta) of the Centrechinoida. On the other
hand, S. drobachiensis is a highly accelerated species in the highest suborder
(the Camarodonta) of the Centrechinoida. As Professor Hyatt showed
fully in his studies of fossil Cephalopoda, primitive genera have a slow
development and specialized genera have a highly accelerated development.
I have shown the same thing in Pelecypoda where the primitive Avicula and
Pecten have a relatively slow development, whereas the specialized Spondylus
and Ostrea have an accelerated development.

Another feature of note in the ocular arrangement of Centrechinus setosus
is the fact that while ocular I or ocular V may be the first to enter the peri-

Studies of Jamaica Echini. 159

proct, V is by far the more common, and in so far makes an approach to the
character of the Cidaridae where ocular V alone is frequently insert, but I
alone is rarely insert. In higher groups of the Centrechinoida, when one
ocular is insert, the order is more definitely fixed, as in the Arbaciidse and
Echinometridae, where V is typically and I is rarely insert; or in the Echinidae
and Strongylocentrotidse, where I is typically and V is rarely insert. Another
ocular peculiarity of Centrechinus is that where four plates reach the peri-
proct, I, V, IV, III insert is a rather frequent aberrant variant. This is a
very rare character in the higher Centrechinoida but is a common character
in the Cidaridse.

Of the ocular characters of Toxopneustes variegatus, no special comment
is called for. It is simply the record from a definite locality, whereas my
previous record of the species was based on specimens from several locali-
ties. The peculiar zigzag pattern of color ornamentation shown recalls the
somewhat similar pattern seen in some of the Temnopleuridse.

In Tripneustes esculentus the ocular arrangement (figs. 10-17) is one of
the most interesting that has been found in any Echini. Two oculars
insert is commonly given as the species character, and it is the dominant
character in some localities (Bermuda, Florida). In Montego Bay this
character is acquired up to 90 per cent in specimens 40 to 50 mm. in diam-
eter, much less than half-grown. But in Montego Bay, as a local char-
acter, the species has gone much further and passed through a phase in which
oculars I, V, IV are insert like Bahama material, and finally takes on oculars
I, V, IV, II insert as the fully matured adult character. It is an exceptionally
interesting case as being the acquirement in late stages of growth as a typical
feature of a character which is a normal progressive variant of its own
species in other localities and also a typical advanced character in all
Echini which have four oculars insert. Looking at the test of a nearly grown
Tripneustes with its solid heavy plates, it is astonishing to realize that they
are still plastic and capable of allowing the readjustment of parts indicated
by the separation of genitals and traveling in of oculars so as to reach the
periproct. Not less surprising is the power of the law of ocular develop-
ment, that not any ocular, but a certain specific ocular, is the one that is
selected by the laws of growth to change its relative position. I know of no
more striking case of the definiteness of action in the growth of organisms.

As stated, Ecliinometra lucunter is one of the very few species of Echini
characterized by having one ocular insert; therefore a series shov/ing its
development is of interest. Four of the six species in the genus Echino-
metra are characterized by having all oculars exsert as the species feature,
and in these to have one ocular insert is a progressive variant. Also as a
family character, all exsert is the leading feature. Therefore for the genus
and for the family, Echinometra lucunter can be considered an advanced
type as regards ocular arrangement. In accordance with this fact it is not
strange, it might even be expected, that it would take on its advanced
character of oculars insert at a somewhat late stage as it does. The domi-

i6o Papers from the Marine Biological Laboratory at Tortugas.

nance of V insert, 54 per cent, is first attained when the specimens are 25
to 30 mm. in diameter, about half grown, and the relative increase of V
insert progresses somewhat up to the largest series recorded. Having
oculars V, I insert is a progressive character for this species, and this feature
comes in late, progressively increasing in frequency up to the largest size
recorded. To have oculars V, I insert as a typical character is the feature of
Echinometra van hrunti A. Agassiz, from Lower California and South America,
and in this respect it is the most specialized species in the genus and family.
In the " Phylogeny of the Echini " much attention was given to ocular
plate arrangement and development of the same. The results for the Cent-
rechinoida are there briefly summarized in pages 92-94. As observations
on the Montego Bay series have materially extended the results in some
important particulars, a brief statement of the laws of ocular arrangement is
given here. In the " Phylogeny " the results of tabulating ocular plates in
young and adults of the order Centrechinoida are given for 48,541 specimens.
Those given above include in addition 2,878 specimens, or including both,
51,419 specimens. By the law of ocular development as worked out in
this order, the oculars when entering the periproct, do so in the sequence
I, VorV, I, IV, II, III (figs. 10-14), and it is seen by the following how closely
this rule is adhered to. Of the total 51,419 specimens of Centrechinoida
observed, oculars are all exsert in 6,763 specimens, all others having one or
more oculars insert. In 3,708 specimens, one ocular is insert, and of these
in 3,679 cases it is either ocular I (fig. 10) or ocular V, that is, 99.22 per cent
are correct by rule. There are only 29 exceptions.^ In 35,951 specimens two
oculars are insert, and of these, in 35,611 cases, it is oculars I and V, or the
bivium (fig. 11), that is, 99.05 per cent are correct by rule. In all of the
340 exceptions^ one of the two plates insert is either I or V. In 3,820 speci-
mens, three oculars reach the periproct, and of these, in 3,426 cases, it is
oculars I, V, IV, the bivium and the left posterior plate of the trivium (fig. 12),
that is, 89.68 per cent are correct by rule. Of the 394 exceptions in 360
cases, the order is I, V, II insert (fig. 15), that is, the right posterior plate of
the trivium is insert instead of the left posterior as usual. These may
therefore be considered as right-handed specimens. Of the total 3,820
specimens with three oculars insert, 3,786, or 99.37 per cent, have either
oculars I, V, IV, or oculars I, V, II insert.^ In 619 specimens, four oculars
reach the periproct, and of these in 554 cases, it is oculars I, V, IV, II the
bivium and posterior pair of the trivium (fig. 13); that is, 89.50 per cent
are correct by rule.'* In 558 cases all five oculars are insert (fig. 14). In the

1 Of these exceptions, 18 have ocular IV alone insert, three have ocular III alone insert, and eight have
ocular II alone insert.

2 Of these exceptions, in loi cases, oculars I, IV are insert; in one case oculars I, III are insert; in s6
cases oculars I, II are insert, which is the species character in Gymnechinus robillardi (Loriol) and G. pulchellus
Mortensen; in 113 cases oculars V, IV are insert; in 4 cases oculars V, III are insert; and in 65 cases oculars
V, II are insert.

' Of the other 34 exceptions to the rule where 3 oculars are insert, in 4 cases oculars I, V. Ill are insert;
in one case oculars I, IV, II are insert; in 11 casesoculars V, IV, III are insert; and in 18 cases oculars V,
IV, II are insert.

* Of the 65 exceptions in 64 cases, the arrangement is oculars I, V, IV, III insert (fig. 16); this character
which is rare in the Centrechinoida as a whole, occurred principally in Centrechinus setosus (47 cases), but it
is a common character in the Cidaroida, the other 17 cases of I. V, IV, III insert are scattered through a number
of species of the Centrechinoida. One aberrant with four oculars insert has oculars I, V, II, III reaching the
periproct (fig. 17). This, which may be considered a right-handed equivalent of I, V, IV, III as far as experience
goes, is a unique variant in Echini.

Studies of Jamaica Echini. l6i

" Phylogeny " a detailed discussion is given of the bearing of aberrant variants
which, although aberrant, follow quite definite lines of their own. Of the
total 51,419 specimens considered above, 50,591 , or 98.39 per cent, are correct
by rule as regards ocular arrangement, and 828 specimens (of which 360 are
I, V, II as above noted), or 1.61 per cent, have an aberrant arrangement.

From the above records it is seen that ocular plates in the Centrechinoida
follow a very definite line as regards which ones reach the periproct, and by
their arrangement emphasize a bilateral symmetry through the axis of
ambulacrum and ocular III and interambulacrum and genital 5. This is
the axis on which the irregular types develop an elongate form, eccentric
periproct and other features of bilaterality, and it was urged by Loven as
the correct axis for orienting Echini. Mr. A. Agassiz {Challenger Echini,
pp. 4-6) criticizes Loven's determination of the antero-posterior axis in
Echini. His argument, however, simply shows that he misunderstood
Loven's point of view.

In accordance with what I previously showed in Strongylocentrotus it is
seen that the perforation of the genital plates by the genital pores originates
at quite definite periods. In Centrechinus setosus the genitals are imper-
forate in specimens up to 14 mm. in diameter, after which with few ex-
ceptions all specimens have developed the genital pores. In Echinometra,
genitals are imperforate up to 10 mm., after which all plates are perforate.
Both of these are later in developing than Strongylocentrotus drobachiensis,
which is imperforate only up to 5 mm. in diameter, after which in most
cases the genitals are perforate.

In Eucidaris tribuloides and Centrechinus setosus the teeth extend to
the upper line of the lantern, but do not extend horizontally over the top
of the same; also, in both the dental capsules embracing the base of the
teeth are small. These are both doubtless primitive conditions, and are
opposed to the condition found in the Echinidse and Strongylocentrotidae.
In Toxopneustes variegatus, Tripneustes esculentus, and Echinometra lucunter
the teeth extend horizontally over the top of the lantern, the proximal end
of the tooth is grooved, not keeled, the keel appearing further down the
tooth, all as in Strongylocentrotus. This grooved character is of interest as
a localized stage in development, as it shows that the young growing point
of the tooth has the character retained throughout life in the Aulodonta,
Cidaroida, and Palaeozoic genera, in which the teeth are always grooved.
In Toxopneustes, Tripneustes, and Echinometra the dental capsules are
large, here again being like the character of Strongylocentrotus and diff'ering
from the condition in Eucidaris and Centrechinus.

In Centrechinus setosus the perignathic girdle is specialized in the adult,
with high apophyses and high laminar auricles joined over the ambulacra.
In the young, the apophyses are very slight and the auricles are simple
spur-like processes, as in the adult Aspidodiadematidae, which represent
the simplest auricles known. Later, in Centrechinus, the auricles arch over
the ambulacra, but without fusion. Then they fuse in delicate arches, a

1 62 Papers from the Marine Biological Laboratory at Tortugas.

condition comparable to that of the adults of ChcBtodiadema pallidum A.
Agassiz and Clark and Centrostephanus rodgersi A. Agassiz. Later still, in
Centrechinus, the auricles develop laminar expansions on their upper border,
which in large specimens become extensive.

In Echinometra lucunter, the perignathic girdle is also very specialized,
attaining an extravagant development of the auricles. In the young,
however, the auricles exist as simple separate styles. Later they arch
over the ambulacra, but without being in contact; still later they join over
the ambulacra as delicate arches, a condition comparable to that seen as the
adult character in simpler species of the genus. Finally, the auricles take
on a great development of the dorsal border, producing the spoon-like arches
characteristic of the species.

As shown and urged by Hyatt from his studies of Cephalopoda, and as
I have shown in previous studies, in the present study of Echini, and in
plants, development, that is the addition of differential characters, con-
tinues more or less markedly throughout the life of the individual, and a
study of such development yields facts of value in the structure, the mor-
phology, and the phylogenetic relations of the group in hand.

X.

THE SPERMATOGENESIS OF THE MONGOOSE; AND A
FURTHER COMPARATIVE STUDY OF MAMMALIAN
SPERMATOGENESIS, WITH SPECIAL REFER-
ENCE TO SEX CHROMOSOMES.

BY H. E. JORDAN,
Professor of Histology and Embryology, University of Virginia.

One plate and nine text-figures.

163

THE SPERMATOGENESIS OF THE MONGOOSE ; AND A FURTHER COM-
PARATIVE STUDY OF MAMMALIAN SPERMATOGENESIS, WITH
SPECIAL REFERENCE TO SEX CHROMOSOMES.

By H. E. Jordan.

INTRODUCTORY.

In view of the recent discovery of heterochromosomes in a number of
the higher vertebrates it seemed desirable that the spermatogenesis of still
other forms should be carefully studied with particular attention to the
possible presence and variable behavior of homologous "accessory" chro-
mosomal elements. While this study was in process opportunity was
afforded to collect in Jamaica, British West Indies, material of the mon-
goose.^ This material is interesting in a number of respects and serves
well as a starting-point for such comparative study as will appear from its
description. Other material at hand includes that of cat, squirrel, pig,
rabbit, white mouse, sheep, horse, mule, bull, and dog.

HETEROCHROMOSOMES LACKING IN MALE.
MONGOOSE.

Spermatogonlal nuclei more generally contain three plasmosomes
(fig. i), though there may be more or fewer. Only a very few nuclei were
seen in mitosis. Many stages of apparent amitosis are present (fig. 2).
The rarity of mitotic figures and of binucleate cells indicates that the pre-
vailing type of division among the spermatogonia is amitotic. The presence,
however, of a few binucleate spermatocytes and spermatids shows that
cell-body divisions do not always follow the nuclear divisions.

The primary spermatocytes contain a single nucleolus (plasmosome),
seen only in the resting condition of the nucleus. The fact that this body
never appears subsequently makes the synapsis and division stages of the
mongoose especially favorable for study with reference to the presence of
an "accessory chromosome." There can be no confusion here between a
plasmosome and a chromosome-nucleolus, as has apparently been the case
in some forms. Moreover, the early disappearance of the plasmosome in
mongoose gives support to the general conclusion that synaptic and post-
synaptic chromatin elements are chromosomes, more especially when
invariably on the nuclear membrane instead of being disposed at random.

The synizesis phase is characterized by a very compact, deeply chro-
matic mass of simple threads (fig. 4). The mass is situated at the pole of

> This opportunity occurred during a month's stay at the temporary Marine Biological Laboratory of the
Carnegie Institution of Washington at Montego Bay in the spring of 1912. I take ttds occasion to express
my thanl£3 for the privileges of the Institution's expedition to Jamaica, and for daily kindnesses on the part of
the director, Dr. Alfred G. Mayer.

165

1 66 Papers from the Marine Biological Laboratory at Tortugas.

the nucleus, frequently, but not invariably, next the idiosome. The re-
mainder of the nuclear space remains empty, except for a few delicate linin
threads, more abundant in the vicinity of the mass. All about the idiosome
are pale spherical mitochondria, now apparently present for the first time;
but no clue appears as to their origin.

The fact that the synizesis nucleus is not appreciably larger than that
of the resting stage shows that synizesis is not due to mechanical factors
consequent upon a great increase in nuclear sap. On the contrary, this
phenomenon is unquestionably the result of the specific activity of the
chromatic nuclear thread. Its frequently polar position may be due to
idiosome influence.

In the bouquet stage the simple threads emerge from the synizesis mas»
in pairs (fig. 5) and unite side by side (parasynapsis) . The nucleus is still
of approximately the same size; and again the chromatin threads give evi-
dence of a specific activity. The process continues until the previously
empty nucleus is filled with double threads crossing in every direction
throughout the nucleus (fig. 6). Subsequently the nucleus enlarges some-
what, and the double threads shorten and thicken (fig. 7). The beginning
of division of the idiosome marks this as the early prophase. The bivalent
chromosomes appear more chromatic and knobbed at the ends, an early
indication of tetrad formation. Figure 8 illustrates a later prophase stage,
in which the chromosomes have become still shorter, uniformly chromatic,
and more compact, a number appearing as complete tetrads.

The point of special interest is that nowhere throughout these early
phases has the slightest evidence of an accessory chromosome appeared.
No evidence would be expected to appear subsequently, and this is actually
the case, as shown in figures 9 to 17. At metaphase the chromosomes are
all arranged in a very compact equatorial plate (figs. 9 and 10). No
chromosome seems marked by peculiar behavior or uncommon size. An
accurate count seems impossible. Innumerable division figures appear,
but the chromosomes are so closely compacted as to. preclude definition of
limits. A most careful study warrants only the statement that the number
of chromosomes is somewhere around 24.

A resting stage ensues, but without chromosome-nucleolus, in contra-
distinction to the usual condition in forms with an accessory chromosome,
or without even a plasmosome. The division figures of the second mitosis
are smaller, somewhat less compact, and seem to consist of fewer chromo-
somes. We may be dealing here, as in the opossum, with a second pairing,
or a hemioid group (Jordan, 191 1). The smaller size of the chromosomes,
however, makes this a little doubtful.

Figure 14 illustrates a resting spermatid, with conspicuous idiosome.
At the next step in the metamorphosis (fig. 15) the cell has elongated, the
nucleus has become more dense and in consequence smaller, and the idio-
some has given place to the archoplasmic sphere. The latter contains
centrally a chromatic granule, presumably the centrosome (fig. 15). The

Spermatogenesis of the Mongoose, etc. 167

sphere seems to have arisen directly from the idiosome by process of vesic-
ulation and rejection of one or several small plastin particles. The con-
densing nucleus moves into an elongating projection of the cell, and then to
and beyond the periphery (figs. 16 and 17). Meanwhile the "sphere" has
flowed backward over the elongating nucleus, its lateral limits being marked
by a distinct line for some distance on either side of the growing axial fila-
ment. The filament is marked proximally by two granules (the products
of the original centrosome, fig. 15), the anterior one at the point of attach-
ment to the nucleus, the posterior one passing along the axial filament as
the "ring centrosome" (figs. 17 and 18) to the posterior end of the middle
piece. At the later and adult stages the middle piece is attached to the
head by a delicate neck (fig. 20, drawn in the living condition), in which
appears a knob, the proximal centrosome, or "end knob" (figs. 19 to 21).
The posterior limit of the middle piece is defined by the termination of the
spiral filament. The latter arises by a coalescence of mitochondria included
within the limits of the backward-extending sphere (middle piece). The
remaining mitochondria fail of inclusion and are cast off with the discarded
cytoplasm. This observation, made also in a number of other forms
(e. g., opossum, Jordan, 191 1; Euchistus, Montgomery, 1912), invalidates
any hypothesis attributing specific hereditary significance to mitochondria.

CAT.

The later steps in the spermatogenesis of the cat, including the spermatid
and subsequent stages, have recently been described by Leplat (1910).
As to the prespermatid stages, it may be said in brief that they are essentially
similar to those described for mongoose, and no convincing evidence appears
at any stage of heterochromosomes. Respecting the postspermatid stages
also, the similarity amounts practically to an identity; and my own obser-
vations on these stages agree essentially with those made and illustrated by
Leplat.

More recently Gutherz (1912) has described and illustrated (in two
figures) first spermatocytes of the cat stained both in the iron-hematoxylin
and the Biondi mixtures. His conclusion confirms my own as to the
probable absence of heterochromosomes. However, he figures a synaptic
(synizesis) and a postsynaptic stage in which appear heterochromosome-
like elements in iron-hematoxylin preparations. The true acidophile
nature of these bodies, however, is revealed in the Biondi material. On
the basis of this observation he casts doubt upon the interpretation as a
heterochromosome of the element described by Winiwarter and Sainmont
in their iron-hematoxylin preparations of the oocyte of the cat. I can only
add that my material (also stained with iron-hematoxylin) reveals elements
somewhat similar to those described and illustrated by Gutherz for the
first spermatocytes, but which, on grounds of morpholog>% location, relation
to spireme, and general behavior, I can not interpret as typical X-elements.
Moreover, in well-decolorized preparations their non-chromatic nature is

l68 Papers from the Marine Biological Laboratory at Tortugas.

likewise revealed. The illustrations of the oocytes of the cat by Winiwarter
and Sainmont (see figs. 27a, 33, and 44) answer to the main criterion, namely
morphological, for the presence of a heterochromosome, "split monosome."
A comparison of the illustrations of the spermatocytes and oocytes of the
cat as given by Gutherz and by Winiwarter and Sainmont respectively
does not seem to warrant the conclusion of an homology between the hypo-
thetical X-element in the male and that in female germ-cells of the cat.
In short, the oocytes would seem to contain, the spermatocytes to lack,
typical X-elements.

SQUIRREL.

The spermatogenesis of the squirrel has been described in two papers by
Van Molle (1906, 1907). The cells are relatively very large and the idio-
some is very distinct. My observations agree with the illustrations of
Van Molle with respect to the absence of heterochromosomes. Duesberg
(1910) has recently questioned the accuracy of Van Molle's observations
with respect to the details of the later stages in the metamorphosis of the
spermatid. Interest in these details is outside the scope of this contribution,
and my observations on these points have not yet been carried to an extent
where I might confidently presume to express an opinion. The absence of
any accessory chromosomal element during the growth period appears un-
questionable.

PIG.

It will suffice for present purposes to state simply that evidence of an
accessory or heterotropic chromosome at early stages of the spermatogenesis
of the pig is completely lacking in my material. The cells here are relatively
small; but the fixation is good and reveals clearly all the details of structure,
and every phase is abundantly represented prior to the division phases ; oc-
casionally first maturation groups are also present.^

RABBIT.

Exactly the same statement holds for the rabbit as for the pig; and I
make it with greater confidence, since the material is more extensive
including besides adult testes those of new-born and young individuals.

In all of the foregoing instances the spermatogonia contain plasmosomes.
The resting primary spermatocytes also contain either plasmosomes or
karyosomes (net-knots), or both, all of which disappear in early presynapsis.
It seems legitimate to assume, on the basis mainly of this observation, that
in mammals generally plasmosomes almost invariably disappear before

1 However, no chromosome counts were attempted. In view of J. E. Wodsedalek's recent work ("Acces-
sory Chromosomes in Pig;" Science, N. S., vol. xxxviii. No. 966, pp. 30-31) reporting two additional chro-
mosomes or accessories in half of the secondary spermatocytes — this number (10) corresponding with the re-
duced number in the eggs — the pig can no longer be listed under "exceptions." In the absence of definite
evidence during the growth period, the final criterion for the presence of heterochromosomes must be an actual
count.

Since the appearance of Wodsedalek's complete paper (Biol. Bull., 25: i), I have carefully re-examined
my sections of the pigs' testes. Mitoses are too infrequent to warrant any statement concerning the presence
or absence of accessory chromosomes on the basis of appearances at metaphase, or of chromosome counts.
My conclusions rested upon evidence drawn from a study of appearances during the growth stages including
the early prophase. As to the resting spermatogonial and primary spermatocyte nuclei I find no such con-
stancy with reference to nucleoli as Wodsedalek records. The number, both of the large and smaller nuc-
leoli, is very variable. Three or four large nucleoli are common. Moreover, with reference to the chro-
mosome-nucleoli I can not find in my preparations (of young testes) any cells to parallel his illustrations 21, 23,
24, and 25. My material does not confirm his finding that "two large round nucleoli remain very conspicuous
throughout the process of growth of the primary spermatocyte."

spermatogenesis of the Mongoose, etc. 169

synapsis, and can thus produce no confusion at those stages where hetero-
chromosomes when present are most conspicuous, i. e., synaptic and early
postsynaptic stages. The conclusion that heterochromosomes are wanting
in the male of the mongoose, cat, squirrel, pig, and rabbit is, of course, based
on the assumption that if present they would be conspicuous at the same
stages in which they are so strikingly in evidence in the group of mammals
next to be described. The final test must, however, in these cases be an
actual count of chromosomes.

HETEROCHROMOSOMES PRESENT IN MALE.

WHITE MOUSE.

Spermatogonial and spermatocyte resting nuclei agree in being vesicular
and having a delicate chromatic reticulum with numerous net-knots (figs.
22 and 23) ; they differ in that the spermatocyte nucleus has a bilobed or
double chromatin (chromosome) nucleolus.

The Sertoli cell has a still more vesicular nucleus, which is very con-
spicuous by reason of a characteristic trilobed, deep-staining nucleolus, the
central section being invariably the larger. This peculiar type of nucleolus
is as characteristic of the Sertoli cell as its peculiar shape, and suggests
interesting speculations concerning the relationship between nucleolar
morphology and cell functions.

My earlier observations led me to the conclusion that the synapsis
phenomenon {i. e., synaptic knot, contraction or bouquet phase, or synizesis
stage) 1 was lacking in the auxocytes of the white mouse. Further careful
search has disclosed a few cells in the condition illustrated in figure 24.
This is tentatively interpreted as synapsis (polarized amphitene). If the
nucleus is normal, as it appears to be, the paired threads unmistakably
indicate synapsis. I find that Regaud (1909, 1910) reports his failure to
find a synapsis phase also in the rat. This phase is assuredly not accentu-
ated in white mouse. Nothing more closely resembling synapsis than the
nuclear arrangement illustrated in figure 24 could be found in my material.
This, however, seems sufficiently suggestive of true synapsis to show that
something similar to synapsis actually occurs in the mouse, perhaps too
rapidly to appear except very occasionally in "fixed" preparations. If the
phenomenon is really lacking the "synapsis" figure does not appear essen-
tial to the pairing of the chromosomes in the formation of the haploid group.
During synapsis (parasynapsis) and early postsynapsis (figs. 24 and 25)
the chromatic accessory body (chromosome) is apparently usually single,
usually oval, sometimes slightly bilobed, and occasionally wedge-shaped or
even irregular. During later postsynaptic (fig. 26) and prophase (fig. 27)
stages, this element again appears invariably double, frequently in the form
of a longitudinally split rod. There is no unequivocal evidence that the

> Synapsis is assumed to take place during the period of contraction or " synizesis "; hence the terms are
here used interchangeably.

170 Papers from the Marine Biological Laboratory at Tortugas.

pair may entirely and widely separate, as is the case in the mule for example.
The double nature of the heterochromosome suggests a pair of idiochro-
mosomes; but the fact that during and immediately after synapsis this
element appears single invalidates somewhat such an assumption, and
suggests rather a double accessory chromosome, or X-element of Wilson.
In leptotene and diplotene phases it very frequently lies in a clear space
free of threads, a condition similar to that described by Wilson (1912) for
Largus and Oncopeltus.

Figure 27a would seem to permit of not the slightest doubt of the presence
of a "sex chromosome" (Wilson) in the white mouse. Its later history,
however, is still obscure. Division figures are very abundant in my material
and the preservation is nothing short of superb, but at no later stage can I
with certainty identify this body. Occasionally a spindle appears with a
large chromosome in advance of the main metaphase group (fig. 276), but
the instances seem too rare to have much significance apart from morpho-
logical marks of identification from among the chromosomes of the meta-
phase plates.

SHEEP.

In sheep I am able to distinguish the "X-element" during the presyn-
aptic (fig. 28), synaptic (fig. 29), and postsynaptic (fig. 30) stages. Both
in the first and second stages it is usually attached to the chromatic spireme,
or one of its segments. During the postsynaptic stages it is usually split
or bilobed, and, as at earlier stages, frequently but not invariably lies close
to the idiosome. Synapsis is effected by side-by-side union. In the sheep,
also, this accessory chromosome eludes later certain identification.

HORSE.

As previously briefly noted (Jordan, 191 1, 191 2), during postsynaptic
stages the horse shows frequently a tripartite X-element (fig. 33), usually
next the idiosome. In the resting phase of the primary spermatocyte
(fig. 31), and during synapsis, it appears double. Further clear tracing of
this body seems impossible. The evidence is thus far from sufficient to
warrant even an inference as to whether we are here dealing with a multiple
X-element (as in Sinea, for example, according to Payne, 1909) or a pair of
idiochromosomes. Both conditions have been described for mammals;
for example, man as having a double X-element (as in Syromastes — ^Wilson,
1909) by Guyer (1910), and guinea-pig with a pair of heterochromosomes, by
Miss Stevens (191 2). The only point sought to establish is the presence of
"sex chromosomes" of some sort.

Kirillow (1912) has recently published the first number of his projected
studies on the spermatogenesis of the horse. The accompanying illustra-
tions, including all the chief stages, give no indication of a heterotropic
body so conspicuous in my preparations; nor is any mention made of such
body. His observations on the germ-cells of the horse lead him to accept
the opinion of Regaud (1909, 1910) that "the appearance or failure of

spermatogenesis of the Mongoose, etc. 171

synapsis is a variable factor in different species" (Kirillow, p. 145); that is,
synapsis (contraction phase) is regarded as a normal phase in the auxocytes
of some forms (species), but is thought not to appear in others.

MULE.

The presence of an accessory chromosome in the mule was also noted
briefly in an earlier paper (Jordan, 191 1, 1912). Here it usually appears
double in pre- and post-synaptic stages (figs. 34 and 36). At synapsis it
appears as a single large oval chromatic body (fig. 35). It is very clearly
recognizable among the prophase chromosomes as a large paired element
(fig- 37)- Since at certain phases (synapsis) it may be a single compact
body, it is perhaps more legitimate to regard it as a double X-element.

In the case of the mule it will be impossible further to trace this body
with any degree of certainty. As first noted by myself (see quotation
from my unpublished manuscript in a paper by Dr. R. H. Whitehead, "A
Peculiar Case of Cryporchism, and its Bearing upon the Problem of the
Function of the Interstitial Cell of the Testis," Anat. Record, Aug. 1908;
vol, II, No. 5), spermatogenesis in the mule does not ordinarily proceed
beyond the early prophase. This is the cause of the sterility of mules bred
inter se; the mule does not generally produce ripe spermatozoa. My obser-
vations were subsequently confirmed by Poll (191 1), who succeeded in
observing a few first maturation spindles at metaphase.

BULL.

The spermatogenesis of the bull has been fully described by Schoenfeld
(1901). My observations agree essentially with his illustrations. Schoen-
feld figures a stage similar to my illustration of postsynapsis (fig. 38), where
a single large, oval, chromatic " accessory chromosome " appears among the
bivalent threads, close to the idiosome. This element, at this as well as
at earlier and later stages, Schoenfeld describes as the "corpuscle intra-
nucleare." In the light of what we now know of "sex-chromosomes"
generally, and in mammals more particularly, this body may, I believe, be
regarded confidently as an accessory chromosome.

DOG.

Figure 39 illustrates a resting primary spermatocyte with pale plasmo-
some and chromatic, bilobed chromosome nucleolus. No very satisfactory
synapsis stages could be observed in my material.^ Figures 40 and 41
illustrate two successive postsynaptic phases in which the X-element is
conspicuously present as a chromatic, sharply contoured, irregularly oval
body close to the idiosome. We are probably here again dealing with an
accessory chromosome. The attempt further to trace this body, however,
was very disconcerting by reason of the absence of definite marks of identity,
lack of clear delimitation of the relatively large number of small chromo-
somes, and confusing contradictions of trying observations. In short, any
further analysis was unsuccessful.

» That is, from the standpoint of a conspicuous " X-element."

172 Papers from the Marine Biological Laboratory at Tortugas.

DISCUSSION.

Heterochromosomes are apparently lacking in mongoose, cat, squirrel,
pig, and rabbit, and appear to be present in white mouse, sheep, horse,
mule, bull, and dog. On the basis of what was already known regard-
ing man (Guyer, 1910; Gutherz, 1912),^ rat (Guyer, 1910), armadillo
(Newman and Patterson, 1910), guinea-pig (Stevens, 191 2), and opossum
and bat (Jordan, 1911, 191 2), heterochromosomes were suspected in mam-
mals generally. With respect to heterochromosomes, then, mammals fall
into two groups: those lacking and those possessing the X-element in the
male germ-cells. Under the first group five^ mammals can be classified;
under the second, twelve. Both groups include higher as well as lower
mammals. Measured by no other criterion known to me can these mammals
be similarly grouped. Presence of heterochromosomes obviously bears no
relationship to natural affinities or evolutionary levels of mammals. The
proportion of 12 to 5 among examined and recorded cases of mammals in
favor of those possessing heterochromosomes suggests that in those forms
apparently lacking them they have simply eluded detection. They might
presumably be so small as to escape recognition, or lack the peculiar
morphological marks by which they are usually characterized, or be too
labile to be identified.

Regarding a similar condition in Culex, Stevens (191 1, 1912) suggests
that here, "where no heterochromosome differentiation of any kind has
been detected, the members of a pair of chromosomes may differ by a sex
unit or some other character unit, and the difference in size comes within
the probable error of most careful observation" (p. 165).

The observation of the attachment of the accessory chromosome to the
parasynaptic spireme (figs. 28, 29, and 40) controverts the plausibility of
Stevens's (1912) suggestion that " the condensed condition of the odd chro-
mosome and the unequally paired heterochromosomes of the growth stage of
the first spermatocytes might be due to their unpaired or unequally paired
condition preventing them from joining with the other bivalent chromo-
somes to form a spireme, especially in cases of parasynapsis " (p. 166).

It would seem to be required, then, for a proper grounding of later
hypotheses, that all uncertainty be disposed of, first with respect to the
possible presence in some form or at some stage of an X-element in those
animals in which it seems to be wanting; and second, with respect to possible
confusion with a persisting plasmosome in those instances where an X-
element is believed to obtain. Though the above order would seem to be
the more logical, the best approach to the argument and the proof is per-
haps by way of the second point, though in fact they sustain a reciprocal
relationship.

> Gutherz (1912) confirms Guyer's (1910) observation respecting the presence of a heterotropic element in
man, but interprets it as a pair of equaJ idiochromosomes undergoing neither heterokinesis nor giving rise to
a dimorphism of spermatozoa as described by Guyer.

* This must be a tentative grouping; only four mammals can now be temporarily placed in this group.

spermatogenesis of the Mongoose, etc. I73

Regarding a possible confusion between plasmosomes and accessory
chromosomes, or X-elements, it must be noted first that in the stain em-
ployed (Heidenhain's iron-hematoxylin ; checked in doubtful instances by
Auerbach's methyl-green-acid-fuchsin stain) the plasmosome when present
(as in spermatogonia and occasionally in primary spermatocytes) is always
in properly decolorized preparations considerably less deeply stained than
the X-element. Moreover, the latter is frequently attached to the spireme
or one of its segments, and generally holds a position close to the nuclear
wall and close to the point where the idiosome lies. Again, it is frequently
bilobed or compound. In certain instances, e. g., mouse more particularly,
one of the first-division chromosomes occasionally passes to the pole in
advance of the mass; but in no instance could the X-element be followed
with any satisfaction or certainty beyond the early prophase. At this stage,
however, it is in at least a number of instances clearly recognizable among
the pale, mossy chromosomes as the body earlier identified as the X-element.
The evidence would seem amply adequate to support the interpretation
here as a typical X-element.

Heterotropic nuclear elements must be identified by other criteria than
staining reaction. At best, staining capacity can only give confirmatory
evidence. Morphological marks would seem to be the most certain grounds
for basing distinctions. The complete nuclear history is of course necessary
for full certainty. Chromosomes undergo alterations in their chromaticity
and consequent staining capacity at different phases of the nuclear cycle.
Again, true chromatin-nucleoli occur in certain forms {e. g., oocytes of
Asterias forbesii, Jordan, 1908; oocytes of Echinaster and Cribrella, Jordan,
1 910), whose function at least in part is to contribute chromatic material
to the chromosomes just before they enter the first maturation spindle.
The plasmosomes of certain forms may thus be of true chromatic nature.
Moreover, true nucleoli never assume — except perhaps very exceptionally
and atypically — the characteristic bilobed and split forms of the chromo-
some-nucleoli during the growth stages. Degenerating nucleoli become
ragged of outline, frequently fragment, and undergo karyolysis, but never
show the series of phenomena characteristic of heterochromosomes : sharp
contour, deep basophily, frequently split form, frequent attachment to
spireme, and usual location close to the nuclear membrane. A presumptive
chromosome-nucleolus may meet the test of basophilic staining reaction in a
differential dye, but if it does not meet at least the majority of the above
structural and spatial tests it very probably is simply a chromatic plasmo-
some; on the other hand, if it answer to the latter tests, confirmatory— but
perhaps not crucial, in view of the undoubted chemical changes which
chromosomes undergo — evidence accrues from a chromatic reaction to a
"specific" stain.

Regarding the first stated point of possible doubt, then, nothing further
can be said than that in the five forms enumerated as lacking an X-element,
the tissue was similarly well preserved and stained, but notwithstanding

174 Papers from the Marine Biological Laboratory at Torlugas.

careful search, not the sHghtest evidence of such an element in its usual
morphological form and peculiar behavior with reference to spireme and
idiosome could at any stage be detected ; but this negative evidence obviously
can not remove all uncertainty with respect to this first point.

Heterochromosomes (allosomes, Montgomery) in mongoose, and other
forms showing like growth-stage conditions, may possibly have disappeared
by close union with one of the ordinary chromosomes and thus forced into
the usual behavior of the autosomes during auxocyte phases. Transition
stages in such process of disappearance are apparently shown in certain
orthoptera (McClung, 1905) and amphibia {e. g., Necturus, King, 1912).

Militating further against the objection of mistaken identity is the fact
that in this group the plasmosome has almost invariably disappeared^ before
synapsis, when the X-element first exhibits its characteristic behavior. If
plasmosome could be confused with X-element in mammals, a group of five,
with no relatively closer affinity than obtains among the contrasted group,
could hardly be expected to give consistent and unequivocal evidence in
favor of complete absence. Moreover, if the plasmosome generally disap-
pears before or about the time of synapsis in a relatively no more closely
related group of five forms, it seems quite legitimate to infer that it disap-
pears at about the same time in six other — and as closely related — forms,
and thus leaves little opportunity for confusion. And the above deduction
is supported to some extent by direct observations (exception: man).

Having disposed of these objections as far as possible within the limita-
tions of microchemical technic, it remains to consider further bearings and
implications of the facts established. The observation of an X-element in
the ovary of young cats (fig. 42) by Winiwarter and Sainmont (1909) has
important significance in this connection. If my negative evidence for the
male germ-cells in the cat can be regarded as equally certain with the
positive evidence of these investigators for the female germ-cells, then it
would seem that in those forms in which the males lack a typical X-element
this is possessed by the female.^ An analogous instance from the inverte-
brate group is that of certain sea-urchins reported by Baltzer (1909).
Here the female possesses a pair of idiochromosomes, while the male lacks
any indication of such. The above hypothesis for mammals must be tested
by examination of the young ovaries of those forms in which the male
appears to lack the accessory.

This hypothesis would be somewhat weakened if cases were known in
which both male and female germ-cells possessed typical^ X-elements.
The meager evidence available is contrary to such a hypothetical condition.
Wilson (1905, 1906) was unable to find a chromosome nucleolus in certain
genera among the insects, including Euchistus, in which he was confirmed
by Foote and Strobell (1909). Buchner's (1909) evidence with respect to
Gryllus — the only presumptive contradictory evidence known to me —

* A delayed disappearance occasionally occurs, more especially in cat and man.

* Presumably by approximately half of the ova.

« That is from the standpoint of behavior during the growth stages.

spermatogenesis of the Mongoose, etc. 175

judged from his illustrations, seems far from conclusive. In fact, the
hypothetical so-called X-element ("accessory body") in the ovary lacks
much of being similar in structure and behavior to the undoubted X-element
in (Edipus (described and figured in the same paper) and the male Gryllus.
Moreover, Gutherz (1910) has shown that the two X-chromosomes, char-
acteristic of the female, are both present in the metaphase plates of the
oogonia, in addition to the outlying "accessory body." As is now estab-
lished beyond any doubt the accessory, for example, of the male germ-cells
passes into all females. Its homologue is present in both males and females ;
but the additional male accessor>% at any rate, would seem to be able again
to comport itself during the growth stage of the maturing oocyte as it
originally behaved in the growing spermatocyte; and similarly with respect
to its female homologue, since that in the male, coming originally also from
the egg, behaves again in typical "accessory" fashion. It seems fairly
well established that the female homologues of the male accessory do not
behave in the growing oocytes as does the X-element in the spermatocyte.
As indicated by observation on the cat, the reverse of the usual relation-
ship here obtains: the oocytes are probably dimorphic with respect to an
accessory; the spermatocytes apparently monomorphic. Here, then, as in
certain sea-urchins, the female is apparently digametic, the male apparently
homogametic. Further investigations must establish the fact whether the
same reverse condition obtains with respect to an X-element in the case
of the remaining four mammals of the first group.

Male and female sex (together with secondary sexual characters) appear
to be hereditary in a manner similar to other organic characters, and may
be regarded as a pair of "unit characters." Since the femaleness is associ-
ated with, or determined by, the presence of X-elements, the female (con-
taining a pair of homologous X-elements) is presumably homozygous,^ the
male heterozygous for sex; or, in accordance with a more recent termin-
ology, the female represents a duplex (but apparently recessive like a nuUi-
plex) condition, the male a simplex (dominant) condition. Maleness is
accordingly characterized by the absence, femaleness by the presence of a
"determiner" (presumably an inhibitor), the X-element. The former is
the negative or minus (lacks accessory, contributed to female at fertili-
zation), the latter the positive or plus, condition. Sex inheritance con-
forms thus to the Mendelian scheme, in which maleness represents the DR
and femaleness the RR condition; or, stated in terms of the presence (-f )

and absence ( — ) of an inhibiting accessory, or X-element, maleness is -^ ,

and femaleness the -f -j- condition. Femaleness may therefore be plus or
positive in an inhibiting sense ; a germinal plus condition prevailing in the
female, inhibiting the appearance of a positive somatic condition or male sex.
This modified Mendelian interpretation is suggested in part by an
analogy with conditions in crosses between horned and hornless cattle,
where the positive character (horns) is recessive to the negative or hornless

I That i3, in the group where the male germ-cells contain an X-element.

176 Papers from the Marine Biological Laboratory at Tortiigas.

character; meaning, according to the general interpretation, that hornless-
ness is determined by the presence of some determiner which prevents the
development of horns. Similarly, femaleness may be due to the presence
of the X-element, preventing the development of maleness.^ It would be
immaterial whether it were contributed by the male or by the female gamete.
This idea is further suggested by, and seems in perfect accord with, the
numerous observations indicating that femaleness is undeveloped maleness,
e. g., embryological facts; relatively female characteristics of male infants;
the assumption of male secondary characters after spaying, or after the
menopause; etc.

SUMMARY AND CONCLUSIONS.

Among the forms examined, including mongoose, cat, squirrel, pig,
rabbit, white mouse, sheep, horse, mule, bull, and dog, heterochromosomes
are apparently lacking in the male germ-cells of the first five, and present in
the remainder. The available evidence favors more the interpretation in
terms of a bipartite univalent or compound X-element than of an asso-
ciated X and Y group (idiochromosomes).

In view of the fact that heterochromosomes have recently been reported
in man (Guyer, double X-element; Gutherz, equal pair of idiochromosomes,
or X- and Y-elements), rat (Guyer), armadillo (Newman and Patterson),
guinea-pig (Stevens), and opossum and bat (Jordan), the evidence indicating
similar elements in the above-enumerated group of six common mammals
would seem to warrant the conclusion that sex-chromosomes are very
generally present in mammals. Universality of presence seems vitiated for
the present by the fact that in another group of five mammals, carefully
studied, such elements seem apparently lacking. It might be assumed
that such elements are actually present in the male germ-cells, but are so
small or labile as to elude detection by ordinary methods, or do not present
the usual morphology of heterochromosomes during the prophase stages.
The unmistakable presence, however, of a "split-accessory" in the female
germ-cells (primary oocyte) of the cat, as recorded by Winiwarter and Sain-
mont, and the absence of any X-element in the male (confirmed by Gutherz,
1 91 2), suggests very forcibly that sex-chromosomes are present in all
mammals, generally in the male, exceptionally in the female. The same
result would follow (that of numerical sex-equality) whether present in one
or the other sex. If this hypothesis can be further sustained, it would seem
cogently to reinforce the evidence for an essential sex-determining function of
heterochromosomes. Interpreted in terms of Mendelian heredity-formulae,
in those mammals in which an X-element is present in the male, the female
sex is homozygous, the male heterozygous. The facts would seem to fit the
hypothesis that the accessory chromosome acts as a deterrent to the develop-

» However, in the case of horn-heredity, the inhibitor, according to this explanation, is dominant; in sex-
heredity, recessive. In cases where the female is the heterozygote, the condition might be parallel to that of
horn-heredity. In these instances the female would be simplex, the male nuUiplex. Recession of a duplex
condition seems a contradiction in terms. The suggestion can have significance in only an approximate or
general sense.

spermatogenesis of the Mongoose, etc. 177

merit of maleness; or more accurately, and in keeping with a quantitative
interpretation of sex in a final analysis, the accessory with its egg-homologue
(two X-elements) inhibits male-sex development, the single egg-homologue
in males being insufficient to counteract the male tendency, thus giving
origin to male individuals. In those instances where the female can be
shown to be the heterozygote, one X-element, according to the hypothesis,
may be assumed to be able to counteract the male tendency.

ADDENDUM.

The opportunity recently presented itself for the study of very favorable
human material. I am indebted to Dr. H. T. Marshall, professor of
pathology at the University of Virginia, for the specimen, a thin slice of
testicle fixed in Zenker's fluid, and obtained at autopsy from a negro aged
38 years, who died as the result of drinking wood alcohol. The material
is in excellent state of preservation. It was stained for study with iron-
hematoxylin. Primary spermatocytes in early postsynaptic phases (text-
figs. I to 9) are especially abundant and clear. Mitoses, however, both in
spermatogonia and spermatocytes, are infrequent, and, though a sufficient
number can be found at metaphase in the latter cells to furnish a fairly
satisfactory opportunity for attempting chromosome counts, I can come
to no definite conclusion concerning the actual number. My specimens of
primary spermatocyte mitoses, however, compare very favorably with the
illustrations given by Guyer and by Gutherz, every one of which I can
practically duplicate.

^ An impartial judge of the figures must admit, I believe, that certainty is
impossible at least within an error of two — and two makes all the difference
between the presence of an accessory as urged by Guyer and denied by
Gutherz. The inherent difficulty of an accurate count is indicated by the
fact that the diploid chromosome number of the male cells has been given as
22 (Guyer), 24 (Flemming, Duesberg, Branca, Gutherz), 16 (Bardeleben),
18 (Wilcox), 33 or 34 (Wieman), and 47 (Winiwarter). It is significant,
however, that Guyer, Gutherz, Duesberg, and Branca agree on 12 as the
haploid (reduced) number; but Guyer, in contrast to the others, regards 2 of
these as univalent accessories; also Guyer, Gutherz, and Winiwarter agree
in claiming the presence of a heterochromosome ; and Guyer and Winiwarter
both report a dimorphism of spermatozoa, the former giving the number of
chromosomes as 10 and 12, and the latter as 23 and 24.

It seems clear that a heterochromosome of some valency is present; it is
equally clear that the true nature of this body can not yet be established on
the basis of numerical relationships; the count is still uncertain as regards
at least 2 chromosomes. Study of my material convinces me that the
reduced number is not less than 12, but perhaps several more. As regards
conclusions respecting the presence of an accessory chromosome drawn from
a study of side views of metaphase spindles, it may be pointed out that
Guyer's and Gutherz's figures are substantially alike ; but Guyer interprets

178 Papers from the Marine Biological Laboratory at Tortugas.

the appearance as demonstrating, Gutherz as contradicting, the presence of
an accessory. My own figures occasionally show a double chromosome at
one pole in advance of the main complex, and the evidence favors Guyer's
interpretation rather than Gutherz's, but there is here nothing so striking
as one sees in certain insects (e. g., the phasmid, Aplopus mayeri) and the
opossum. Strong evidence for the presence of a heterochromosome in
man is given by Gutherz in his illustrations of early postsynaptic nuclei
(text-figs. 2 to 6). But neither he nor Guyer seems to offer satisfactory
evidence as to its composition from chromosome counts and behavior at
metaphase. Both, moreover, regard it as a double structure. My own
evidence confirms the findings of Guyer and Gutherz in so far as pertains to
the actual presence and bipartite nature of a heterochromosome or X-ele-
ment, and it emphasizes the additional points of similarity between this

Fig. I.— Human primary spermatocyte; nucleus in early diplotene post-synaptic phase, showing a deeply
staining oval chromatin (chromosome) nucleolus on the nuclear wall and attached to one of the threads; the
true nucleolus is a spherical, less deeply staining body, and only very rarely placed close to the nuclear wall.
The cytoplasm is homogeneous, finely granular. . , ... r »t,„ i,„*»,«^t,,„

Figs. 2 to o.— OutUnes of nuclei at same stage to show the general form and location of the heterochro-
mosome. It is almost invariably bipartite and attached to one of the threads, and on the nuclear membrane.
The nucleolus when present is indicated in outline.

and undoubted sex-chromosomes, namely, frequent connection with the
diplotene thread and close spatial relationship to nuclear wall (exception,

text-fig. 3).

Judging from the illustrations, Winiwarter (191 2) has had the advantage
of possessing by far the best human material. His illustration (fig. 25) is
practically a duplicate of my own of the postsynaptic diplotene (passing
into the pachytene) phase. However, the later history of the accessory
element as traced by Winiwarter is not at all stages perfectly clear. But
figures 39 and 40 show pairs of sister spermatids, one pair with, the other
without, a deep-staining nuclear body (accessory chromosome?). If this
spermatid element actually represents, as seems very probable, the postsyn-
aptic nuclear bipartite body, or heterochromosome, then cogent additional
evidence here also accrues favoring the legitimate interpretation as a
heterochromosome of the similar (structurally and tinctorially) peculiar
postsynaptic nuclear element wherever found in mammals.

LITERATURE.

Baltzer, F. 1909. Die Chromosomcn von Strongylocentrotus lividus und Echinus micro-

tuberculatus. Arch. f. Zellf., Bd. 11, Heft 4.
BucHNER, P. 1909. Das accessorische Chromosome in Spermatogcne?e und ovogenese

der Orthoptern, zugleich ein Beitrag zur Kenntnis der Reduiction. Arch.

f. Zellf., Bd. IV, Heft 2.
DuESBERG, J. 1910. Nouvelles recherches sur I'appereil mitochondrial des cellules

seminales. Arch. f. Zellf., Bd. vi, Heft I.
Foot and Strobell. 1909. The nucleoli in the spermatocytes and germinal vesicles of

Euchistus variolatus. Biol. Bull., vol. xvi. No. 5.
Gross, J. 1912, Heterochromosomes and sex-determination. Zool. Jahrb., vol. xxxii.
GuTHERZ, S. 1909. Wird die Annahme einer Beziehung zwischen Heterochromosomen

und Geschlechts bestimmung durch das Studium der Cxryllus-oogenese

wiederlegt? Sitzungsber. d. Gesellsch. Naturforsch, Freunde, Nr. 9.

1912. Ueber ein bemcrkenswertes Structurelement (Heterochromosom?) in der

Spermiogenese des Menschen. Arch. f. Mikr. Anat., Abteilung 11, 79, 2.
GuYER, M. F. 1910. Accessory chromosomes in man. Biol. Bull., vol. xix. No. 4.
Jordan, H. E. 1908. The spermatogenesis of Aplopus mayeri. Carnegie Institution of

Washington Pub. No. 102. Abstract in Anat. Anz., vol. xxxii.

1908. The relation of the nucleolus to the chromosomes in the primary oocyte

of Asterias forbesii. Carnegie Institution of Washington Pub. No. 102.
■ 19 10. The relation of nucleoli to chromosomes in the egg of Cribrella sangnine-

olenta Lutken. Arch. f. Zellf., Bd. v. Heft 3.
191 1. The spermatogenesis of the opossum {Didelphys virginiana), with specia

reference to the accessory' chromosomes and the chondriosom.es. Arch.

f. Zellf., Bd. VII, Heft i.

19 1 2. Notes on the spermatogenesis of the bat. Anat. Anz., Bd. XL.

King, Helen Dean. 1912. Dimorphism in the spermatozoa of Necturiis macidosus.

Anat. Record, vol. vi, No. 10.
KiRiLLOW, S. 1912. Die Spermiogenese beim Pferde. I. Arch. f. Mikr. Anat., Bd.

Lxxix, Abteilung 11, Heft 3.
Leplat, G. 1910. La Spermiogenese chez le Chat. Arch, de Biol., Tome xxv, Fasc.

2 and 3.
McClung, C. E. 1905. The chromosome complex of orthopteran spermatocytes.

Biol. Bull., vol. IX.
MoLLE, J. VAN. 1906. La Spermiogenese dans I'ecureuil. La Cell. Tome xxiii, Fasc. I.

1907- Les Spermatocytes dans I'ecureuil. Ibid. Tome xxiv, Fasc. 2.

Montgomery, T. H. 1910. On the dimegalous sperm and chromosomal variation of

Euchistus, with reference to chromosomal continuity. Arch. f. Zellf.,
Bd. V, Heft I.

1909. A biological and cytological study of sex determination in

phylloxerans and aphids. Jour. Exp. Zool., vol. vii. No. 2.

and J. T. Patterson. 1910. The development of the nine-banded

armadillo from the primitive streak to birth, with special reference to

the question of specific polyembryony. Jour. Morph., vol. xxi. No. 3.

1909. Some new types of chromosome distribution and their relations to sex.

Biol. Bull., vol. XVI, Nos. 3 and 4.
191 1. Mischlings Studien V: Vorsamenbildung die Mischlingen. Arch. f.
Mikr. Anat., Abteilung 11, Bd. Lxxvii, Heft 2.
1909-1910. Etudes sur la Structure des Tubes Seminiferes et sur la Sper-
matogenese chez les Mammiferes. Arch, de Anat. Micr., Tome xi. Paris.
1902. La Spermatogenese chez le taureau, et chez les Mammiferes en
general. Arch, de Biol., Tom.e xviii, Fasc. i.

1909. An unpaired heterochromosome in the aphids. Jour. Exp. Zool.,
vol. VI, No. I.
Further studies on heterochromosomes in mosquitoes. Biol. Bull., vol. xx.
Heterochromosomes in the guinea-pig. Biol. Bull., vol. XXI, No. 3.

1913. Chromosomes in man. Am. Journ. Anat. 14: 4.
1905A. The chromosomes in relation to the determination of sex. Science,
n. s., vol. XX, 564.

The behavior of the idiochromosom.es in Hemiptera. Studies on Chromo-
somes, I. Jour. Exp. Zool., vol. 11, No. 3.

The paired microchromosomes, idiochromosomes, and heterotrophic
chromosomes in Hemiptera. Studies on Chromosomes, II. Ibid.,
vol. II, No. 4.

Morgan, T.
Newman, H

Payne, F.
Poll, H.
Regaud, C
Schoenfeld, H.
Stevens, N. M.

1911.

1912.

Wieman, H. L.
Wilson, E. B.

1905B
1905 c

179

I So Papers from the Marine Biological Laboratory at Tortugas.

WiLSOX, E. B. 1906. The sexual differences of the chromosome groups in Hemiptera,
\vith some considerations on the determination and heredity of sex.
Studies on Chromosomes, III. Ibid., vol. iii, No. i.

1909A. The accessory chromosome in Syromastes and Pyrrhocoris with a com-
parative review of the types of sexual differences of the chromosome
groups. Studies on Chromosomes, IV. Ibid., vol. vi. No. i.

190913. The chromosomes of Metapodius. A contribution to the hypothesis

of the genetic continuity of chromosomes. Studies on Chromosomes, \'.
Ibid., vol. VI, No. 2.

1910. The chromosomes in relation to the determination of sex. Science Prog-
ress, vol. IV, No. 16.

191 1. The sex chromosomes. Arch. f. Aiikr. Anat., Bd. Lxxvii, Abteilung 11,

Heft 2.

1912. Observations on the maturation phenomena in certain Hemiptera and

other forms, with considerations on synapsis and reduction. Studies on
Chromosomes, VIII. Ibid., vol. xiii, No. 3.

Von Winiwarter, H. 1912. Etudes sur la spermatogenese humaine. Arch. Biol.
Tome xxvii.

\'oN Winiwarter, H., and G. Sainmont. 1909. Nouvelles recherches sur I'ovagenese
et I'organogenese de I'ovaire des Mammiferes (chat). Arch. Biol.,
Tome XXIV, Fasc. 2 and 3.

DESCRIPTION OF PLATE.

Fixation: Flemniing's strong solution. Stain: Heidenhain's iron haematoxylin.
Magnification: 1500 diameters.

Mongoose.

1. Spermatogonium; the nucleus contains three

small plasmosomes.

2. Spermatogonium; the nucleus is in process of

amitotic division.

3. Primary spermatocyte; the nucleus contains a

centrally located plasmosome; the delicate
chromatic reticulum consists of an apparently
continuous single thread; leptotene-nucleus.

4. Bouquet stage; the slightly thickened spireme has

aggregated in a close-meshed mass at the
idiosome pole of the nucleus; the cytoplasm
surrounding the idiosome contains pale mito-
chondria; synizesis, polarized amphitene.

5. Synapsis: the disentangling threads are pairing

side by side; the nucleus is still approximately
of the same size as in the resting spermatocyte,
hence synizesis and synapsis signify activity
(motion) on the part of the nuclear reticulum;
synaptene-nucleus.

6. Postsynapsis; early prophase; diplotene-nucleus.
7 and 8. Later prophases; at no stage do hetero-

chromosomes appear.
9 and 10. Side and polar views respectively of
metaphase plates; no chromosome is conspic-
uous for unusual size or behavior; the number
of chromosomes is approximately twenty-
four, the haploid group.

II. Resting secondary spermatocyte.

12 and 13. Side and polar views respectively of
metaphase plates; the number of chromosomes
is approximately twelve, forming a hemioid
group.

14. Spermatid; the cytoplasm has become filled with

pale mitochondria.

15. Appearance of "sphere" and centrosome.

16 and 17. Appearance of a-xial filament and middle
piece.

18 and 19. Formation of spiral filament by coales-
cence of mitochondria.

20 and 21. Ftice and profile views respectively of
mature, living spermatozoa.

White Mouse.

22. Spermatogonium.

23. Primary spermatocyte, nucleus at leptotene

phase; the bilobed nucleolus may be a pair of
heterochromosomes.

24. Synapsis, polarized amphitene, bouquet or

' synaptene phase; the heterochromosome is
apparently single.

25. Postsynapsis; early prophase; heterochrome (or

accessory) conspicuous; pachytene-nucleus.
26 and 27a. Early and late prophase groups respec-
tively, both including conspicuously the pair of
heterochromosomes almost invariably situated
on the nuclear wall and close to the idiosome;
26, diplotene-nucleus.

276. Metaphase; one of the chromosomes has moved
in advance of the others to one pole; this has a
tetrad form and is still clearly bivalent; ac-
cordingly, it fails to divide in this mitosis and
may represent the "accessory" of the growth
stages.

Sheep.

28. Resting primary spermatocyte, showing the

accessory at the idiosome pole; perhaps post-
synaptic "confused stage" of Wilson.

29. Bouquet (synapsis) stage; the accessory chromo-

some is attached to a double thread, and in-
variably lies close to the nuclear wall.

30. Postsynapsis; early prophase, showing a bipartite

accessory (or pair of heterochromosomes) .

Horse.

31. Primary spermatocyte; the nucleus contains

two irregular karyosomes and a bilobed
chromatic nucleolus, probably a pair of hetero-
chromosomes.

32. Synapsis stage, showing the heterochromosomes

among the pairing threads close to the nuclear
wall.

33. Postsynapsis, showing a tripartite X-element,

the heterochromosome group; pachytene-
nucleus.

Mule.

34. Resting primary spermatocyte, showing a

divided, or paired, heterochromosome.

35. Synapsis stage, showing a heterochromosome.

36 and 37. Early and late prophase stages respec-
tively, showing a pair of heterochromosomes.

Bull.

38. Postsynapsis, showing an accessory chromosome
among the bivalent threads, close to wall and
idiosome; diplotene-nucleus.

Dog.

30. Primary spermatocyte; presynaptic resting
phase, showing plasmosome and bilobed
monosome.

40. Postsynaptic phase, pachytene-nucleus; mono-

some attached to one of the threads.

41. Early prophase, diplotene-nucleus; chromosomes

in form of long bivalent threads; monosome
close to nuclear wall.

42. Oocyte from ovary of 23-day cat. Synapsis.

True nucleolus; monosome divided longitu-
dinally (after H. von Winiwarter et G. Sain-
mont, fig. 43).

JORDAN

XI.

THE BRYOZOA OF THE TORTUGAS ISLANDS, FLORIDA.

BY RAYMOND C. OSBURN.
Twenty-three text-figures.

i8i

THE BRYOZOA OF THE TORTUGAS ISLANDS, FLORIDA.

By Raymond C. Osburn.

In the summer of 1908 the writer had the privilege of spending the
month of June at the Carnegie Institution Laboratory for Marine Biology,
located on Loggerhead Key of the Tortugas Islands. Owing to the short
time at my disposal the entire period was devoted to a close search for the
bryozoa inhabiting the shallow waters about the reefs, on the piles of the
old government dock on Garden Key, in the moat of old Fort Jefferson
on the same key, and in dredging the shallow waters about the islands
down to 22 fathoms. For much of the work a skiff or a small launch
was used, and for the deeper dredging (10 to 22 fathoms) the schooner
Physalia was employed. The Tortugas Islands are islets of a coral atoll
and in the diversified bottom afforded by such a region a rather abundant
bryozoa fauna was found.

Comparatively little work has been done on the bryozoa of the Florida
and West Indian regions. Smitt's papers on the Florida bryozoa (1872-73)
deal with 87 species collected by Count L. F. de Pourtales and afford the
only extended record of the bryozoa of this region. Pourtales (1867), in
a paper entitled "Distribution of the Fauna of the Gulf Stream at Great
Depths," listed and described as new 7 species of bryozoa, 2 of which have
been proved to be synonyms. Verrill, in his Bermuda papers, has listed 21
species from the Bermuda Islands. Levinsen, in his important work,
"Morphological and Systematic Studies on the Cheilostomatous Bryozoa"
(1909), records 6 species from the West Indian and Florida region, describing
two of them as new. Otherwise this vast region remains untouched. As
the collections of Pourtales were made in the deeper waters of the Florida
region, it was presumed that careful collecting in the shallow waters would
disclose the presence of a somewhat different fauna, a presumption well
borne out by the results of my collecting. There is, indeed, a remarkable
disparity between the list given by Smitt and that forming the basis of the
present paper. Not only did 41 of the species described by Smitt fail to
appear in the shallow waters, but 40 others, whose presence was not hitherto
suspected in the Florida region, and many of them not even in America,
were taken. This brings the list of the species at present known from
Florida and West Indian waters up to 127.

By comparing with lists from other regions where the bryozoa have been
carefully worked, it will be seen that the bryozoa fauna of the Tortugas
and of the Florida- West Indian regions is fairly rich in species and fairly
representative of tropical and semi-tropical regions. Careful collecting for

183

184 Papers from the Marine Biological Laboratory at Tortugas.

a number of years in a limited region about Woods Hole, Massachusetts^
and extending down to about 20 fathoms, yielded 83 species (see Osburn,
1 91 2). Waters (1909-10) listed 73 species from the Red Sea. Norman
(1909) records 139 species from Madeira and neighboring islands. This
latter list is of special interest for comparison, since the Madeira bryozoa
have been collected with some care for about 50 years and it may be supposed
that it is fairly complete. In making comparison, however, it must be
noted that the dredging about Madeira extended down to 200 to 300 fathoms,
while those of the present list go only to 22 fathoms, so that to properly
compare the Florida with the Madeira fauna the present list of shallow-
water forms should be combined with that of Smitt.

The Tortugas bryozoan fauna (shallow water) presents a very different
facies from that of Woods Hole at the same depth, since the two lists of 76
and 83 species, respectively, have only 13 species in common. The Woods
Hole fauna is distinctly northern, while that of the Tortugas, at least for
the shallow waters, is distinctly tropical. This is readily understood when
we recall that the Tortugas are within 0.5 degree of the tropics, and that
these islands receive, directly along their shores, the warm waters of the
Gulf Stream which sweep up from the Caribbean Sea.

It is interesting to note in this connection that, of the 40 species here
added to Smitt's list, 33 are limited entirely or chiefly to a tropical distri-
bution.

The arrangement of the bryozoan species, genera, and families has under-
gone so much alteration in recent years, as a result of the studies of Norman,
Waters, Calvet, and Levinsen, as to be wholly unintelligible to anyone except
the systematist familiar with the changes of the last ten years. It is im-
possible that this should be otherwise, since the older classification was
based almost entirely upon the structure of the skeleton of dead and dried
material, without reference to the general morphology.

In the following list I have followed as far as possible the classification
adopted by Levinsen (1909). In the cases of a few imperfectly understood
species the older genera have been retained as "catch-alls." Thus the
genera (?) Lepralia and Phylactella have been used in the old Hincksian
sense for certain species which, with more complete knowledge, must un-
doubtedly go elsewhere.

The figures were drawn by Mr. S. Shimitori, except figures 7, li, 22,
and 23, which are the work of Mr. H. Murayama.

ENTOPROCTA.

Genus Pedicellina Sars, 1835.

Pedicellina cernua (Pallas).

Osburn, 1912, p. 213, Synonymy and previous records of the occurrence of this species
on the American coast.

Apparently not common, but several small colonies were taken on the
piles of docks and dredged at 10 fathoms.

The Bryozoa of the Tortugas Islands, Florida. 185

Genus Barentsia Hincks, 1880.
Barentsia discreta (Busk).

Busk, 1886, p. 44 (Ascopodaria discreta). — Jullien, 1888, p. 13 [PedeceUina auslra-
lis). — (?) Verrill, 1900, p. 594 {Barentsia timida sp. nov.)- — Waters,
1904, p. 99. — OsBURN, 1912, p. 214.

This species, first described from Tristan da Cunha in the Challenger
reports, is now known to be widely distributed, as follows: Tristan da
Cunha, 100 to 150 fathoms (Busk); China Sea, 27 fathoms (Kirkpatrick) ;
Cape Horn, 26 fathoms (Jullien); He Londonderry, Magellanes, Chile
(Waters) ; Woods Hole, Massachusetts (Osburn) ; Beaufort, North Carolina
(Osburn); Tortugas, Florida (Osburn).

If Verrill's Barentsia timida from the Bermudas is the same species, as
I strongly suspect, this will add another locality. Verrill states in regard
to his B. timida that it is closely allied to B. discreta, but that the latter has
a shorter and more annulated basal cylinder and also several annulations
of the stem below the base of the cup. In my experience, the annulations
of the stem are of little importance, if any, as there is much individual vari-
ation which may represent only differences in contraction at this point.

The individuals are small and the species may readily escape observation.
It appears only once in my collection made at the Tortugas, being dredged
in 18 fathoms directly north of the island on a bottom of coral mud; several
fragments of colonies attached to shells.

CYCLOSTOMATA.
Genus Crisia (part) Lamouroux, 1816.
? Crisia denticulata (Lamarck).

Lamarck, 18 16, p. 137 (Cellaria denticulata). — Stimpson, 1853, p. 18. — Smitt, 1872,
p. 4 (Crista eburnea). — Jelly, 1889, p. 73. — Harmer, 1891, p. 136. — Verrill,
1879, p. 28; 1900, p. 592. — Whiteaves, 1901, p. no. — Osburn, 1912, p. 216.

Numerous small colonies were taken at from 10 to 15 fathoms in various
places about the island, attached usually to sponges and shells. Smitt
records the species as C. eburnea from 7 to 60 fathoms. Stimpson (1853),
Verrill (1879), and Whiteaves (1901) have recorded it from New England
to Canada, though these records are questionable on account of failure
to consider the ooecia, and the writer has recorded it questionably from
Woods Hole, Massachusetts (ooecia wanting). Verrill has reported it as
common at Bermuda.

Harmer (/. c), referring to Smitt's Florida record, states: "I do not feel
certain that the form described is really identical with C. ramosa, although
it can hardly be regarded as C. denticulata,'' and suggests that it may be
Stimpson 's C. cribraria. It does not seem to agree with the latter species,
however, in any essential point (see Osburn, 1912, p. 215. for redescription
of C. cribraria), nor does it agree with C. ramosa, in which the radical
fibers have long internodes with yellow or colorless joints. It does agree
with C. denticulata in having short internodes in the radical fibers and the

i86 Papers from the Marine Biological Laboratory at Torttigas.

joints of these fibers, as well as those of the zoarium, conspicuously jet-
black. As the form of the colony and of the zooecia agree fairly with
dentiadata, I retain that name for the species. It is unfortunate that
ovicells are wanting, so that it is impossible to make a positive determina-
tion, for, in spite of the number of times the species has been recorded from
North America, there seems to be not one of these records based on an
unquestioned identification.

Genus Lichenopora Defrance, 1823.
Lichenopora hispida (Fleming).

Fleming, 1829, p. 530 {Discopora hispida).— ]elly, 1889, p. 134, synonymy.

A very minute specimen was taken on an alga at 2 fathoms and placed
with some question in this species. A well-developed colony taken by Dr.
Paul Bartsch at Biscayne Key, Florida, has recently been examined, and
after comparison there seems to be no doubt as to the identity of the Tortu-
gas specimen. The species hitherto has not been noted in the Florida or
West Indian regions.

CHEILOSTOMATA.

Genus Aetea Lamouroux, 1812.
Aetea truncata (Landsborough).

Landsborough, 1852, p. 288 {Anguinaria truncata).— Jelly, 1889, p. 5, synonymy
and references.— Cornish, 1907, p. 75.— Norman, 1909, p. 283.

Common in shallow water and down to 5 fathoms, creeping over shells

and seaweed. It has been recorded at Canso, Nova Scotia (Cornish, I. c).

Aetea sica (Couch).

Couch, 1844, p. 102 {Hippothoa sica).— Jelly, 1889, p. 5 i^^tea recta) synonymy and
references.— Waters, 1909, p. 129 {^tea recto).— Norman, 1909, p. 283.

Tortugas at 10 fathoms on shells. Not previously noted in American
waters.

Some authors prefer to consider this only a variety of Mtea anguina.
The latter has not been taken in the Florida region, though it is common
northward from Beaufort, North Carolina.

Genus Bugula Oken, 181 5.

Bugula neritina (Linne).

Linne, 1758, p. 38 {Sertularia neritina).— Ve-rrill, 1878, p. 304 {Acamarchis neritina);
1900, p. 588.— Robertson, 1905, p. 266.— Waters, 1909, p. 135.

This widely distributed warm-water species is abundant in the shallow
water, growing on piles, shells, seaweed, etc. Colonies 1.5 inches in height
with ocecia abundantly developed were taken on the bottom of a skiflf
which had been in the water only from May i to June 23.

Verrill has recorded the species for the Bermuda Islands and from Fort
Macon, near Beaufort, North Carolina. Otherwise it has not been noticed
on the American side of the Atlantic. Robertson found it common on
the Pacific coast as far north as Monterey Bay, California.

The Bryozoa of the Tortugas Islands, Florida.

187

Bugula neritina var. minima Waters.

Waters, 1909, p. 136,

One small colony with ovicells abundantly developed and with numerous
avicularia (as described by Waters in Red Sea specimens) was taken at a
depth of 8 fathoms. Previously known only from the type locality. This
variety is distinguished from the typical neritina by the smaller size of its
zooecia and by the presence of avicularia situated near the lower end of the
zooecia.

Bugula flabellata (Gray).

Gray, 1847, p. 106 {Avicularia flabellata).— Smri, 1872, p

1873, p. 711. — Verrill, 18796, p. 189; 1880, p.

OsBURN, 1912, p. 225.

A single small colony of this species, attached to Retepora atlantica, was

taken at a depth of 12 fathoms. The oviccll is more widely opened than

in Woods Hole specimens. A very widely distributed species, occurring

northward on our coast to Maine. Miss Robertson found it at San Diego,

California.

Bugula microcecia n. sp. (Figs

18. — Verrill and Smitt,
29 (Bugula fluslr aides). —

p,G. I. Bugula microxcia n. sp. Front view of portion of branch, showing details. Upper two oceaa not

quite complete. Lower one complete with embryo.

Fig 2 — The same. Portion of main branch showing joints. Degenerate polypides are present.

Pig, 3. ^The same. Portion of stem higher up, showing less-modified zocEcia. Remains of degenerate poly-
pides within.

Zoarium delicate, composed of a central stalk with long, narrowly
flabellate branches arising in an irregular dichotomous fashion from the
main stem. The stalk, as well as its branches, consists of very much
elongated and modified zooecia arranged biserially. That these are really
zocecial in character is shown by the fact that they contain polypides and
occasionally one of them even bears an avicularium. Notwithstanding
this, definite joints appear at the bifurcations of the main and all the
accessory branches. These conditions show an interesting connecting link

1 88 Papers from the Marine Biological Laboratory at Tortugas.

between the Bugula and Stirparia types of stems. From the lower stem
zooecia arise numerous, strong, radical fibers.

The ordinary zooecia, which are arranged biserially, are long and slender
and expand but little toward the tip; the frontal aperture extends almost
to the base and faces somewhat toward the axis of the branch; a strong
spine is situated at the distal outer edge; near the distal end, at the point
where the outer wall curves farthest over the frontal area, is situated a
short, stout avicularium with a strongly decurved beak; the avicularium is
set upon a stalk so short as to be almost wanting.

The ooecium is hemispherical, very small, and set very low down, par-
tially hidden by the spine, so as to be very inconspicuous; it is thin and
transparent, with a slightly arched aperture. Many of these contain em-
bryos in various stages of development.

Taken at i8 fathoms on a bottom of coral mud, several colonies attached
to shell fragments; the largest 2.5 inches in height. Small fragments were
also found decorating the legs of crabs {Hyas sp.?) at the same depth.

Bugula caraibica Levinsen.

Levinsen, 1909, p. 104, pi. m, figs. 2A-2N.

This species, recently described from St. Croix, Danish West Indies, is
common at the Tortugas on the piles of the government docks, and colonies
1.5 inches high were taken on the bottom of a skiff which had been in the
water only from May i to June 23, 1908.

It is one of the most conspicuous species of the region, growing in loose
tufts of a fine purple color to the height of 4 to 6 inches. Levinsen does not
recognize Stirparia as a separate genus, but if the species with a jointed
stalk are to be separated, caraibica will fall in that genus,

Bugula armata Verrill. (Figs. 4 and 5.)

Verrill, 1900, p. 588 {Bugula [Caulibugula] armata n. sp.).

Verrill has described from
Bermuda a species which I
take to be the same as one
occurring at Tortugas. His
description is as follows:

A much more delicate, white
Bugula, consisting of diverging, fan-
like branches attached to the alter-
nate sides and to the tip of slender
jointed stems, sometimes having
alternately a long joint and a very
short joint, but more frequently the
short joint is lacking and the end of
the joint is swollen, as in Stirparia.
There are usually 2 or 3 annulations
at the base of each main branch and
these arise just below the internodes. Many of the cells have a slender, distal vibraculum or
sometimes two. It should doubtless form the type of a new genus or subgenus {Cauli-

FiG. 4. — Bugula armata Verrill. Modified zooecia at base
of branch, showing arrangement of spines and position
of avicularia.

Fig. S- — The same. At tip of same branch, showing character
of spines and position of avicularium.

The Bryozoa of the Tortugas Islands, Florida. 1 89

bugula) intermediate between Biigula and Bicellaria on account of its articulated spines or
vibraculum, and related to Stirparia by its jointed stem. It may be named Bugula (Cauli-
bugula) armaia. Its zooecia are oblong and biserial, alternate. The pedicellariae are on
short pedicels, large, lateral not numerous.

Verrill in the above description overlooked some interesting and im-
portant points. He makes no mention of radical fibers which originate
at the lower ends of the internodes of the stalks. He does not mention the
ooecium, which is subglobose with a rather wide opening directed toward the
zooecial axis. The ooecium is attached by a narrow stalk to the outer edge
of the zooecial aperture near the distal extremity and is turned sidewise at a
right angle to the zooecial axis.

The avicularia ("pedicellarise" of Verrill) show a peculiar distribution.
The basal zooecium bears no avicularium, but the zooecia immediately above
have the avicularia situated on the side about half-way along the aperture.
The next few cells in order have the avicularium farther and farther removed
toward the distal end, and in all zooecia farther out on the branches the
avicularium is on the distal outer edge of the aperture. This is a feature
which I have never seen exhibited by any bryozoa. There are one or two
ong spines, jointed at the base, on the distal end of the zooecium (sometimes
wanting). Verrill errs in calling these "vibracula," a term which should
be applied only to modified avicularia with elongate mandibles. The basal
zooecium of the main branches arising from the stem differs from the others
in having usually 6 spines surrounding the aperture, 4 on the outer, 2 on the
inner edge, a feature exhibited by Stirparia occidentalis (Robertson, 1905,
pi. XIII, fig. 73).

I can see no reason for erecting a new genus for this species. The
jointed stalks, swollen at the nodes, seem sufficient to place it in Stirparia
if that name were to be maintained for the Bugulas with jointed stalk.
The jointed spines do not necessarily relate the species to Bicellaria, as this
feature is present in some species of Bugula and Stirparia.

Taken on a number of occasions at 8 to 10 fathoms, growing in a sprawl-
ing fashion over sponges. Found also on the legs of spider crabs at 10 to
18 fathoms, and dredged on a bottom of coral mud at a depth of 18 fathoms.

Genus Beania Johnston, 1838.
Beania mirabilis Johnston.

JoHNSTOX, 1847, p. 372.— Jelly, 1889, p. 17, for further reference.— Robertsok,
1905. P- 276.

Several fragments of colonies on the legs of a crab {Hyas sp.) taken at
18 fathoms. This striking and well-known species has not before been
noticed on the American side of the Atlantic.

Beania intermedia Hincks.

HiNCKS, i88ia, p. 133 {Diachoris intermedia).— Waters, 1906, p. 15; 1909, p. I37.
At 5, ID, and 15 fathoms, sprawling over hydroids and other bryozoa,
and on shells. Very small colonies consisting of only a few cells. The

190 Papers from the Marine Biological Laboratory at Tortugas.

simple, spineless character of the zooecium easily serves to distinguish it
from others of the genus.

The species is entirely southern in its distribution, being recorded from
Tasmania (type locality), Australia, Chatham Island, Ganjam coast, and
the Red Sea. The present record therefore adds the species to the Atlantic
fauna and greatly increases the known range.

Hincks described the species as having spines on each side. Waters
(1909) remarks that they are absent in Red Sea specimens, but sometimes
present in those from Chatham Island. There is no indication of them in
my material.
Beania cupulariensis n. sp. (Figs. 6 and 7.)

Small, straggling colonies growing attached to the under side of Cupularia
guiniensis. The individual cells are fairly large (about as in B. mirabilis).
On each side of the flattened frontal area there are about 6 (5 to 7) slender
marginal spines which curve somewhat over the area. At the distal end
2 smaller spines project forward. At each side of the aperture is placed an

Fig. 6. — Beania cupulariensis n. sp. Mode of branching and details of /w/f^\

zooecium. ^~ — ''

Fig. 7. — The same. Dorsal side, showing zocedal base and point of origin of
dorsal fiber.

avicularium situated on a short stalk on margin of cell; as far as observed,
these are always paired. The form is short, the beak strongly decurved.
The zooecium arises from the preceding one of the series, near the distal
part of the dorsal surface, by a tubular stalk so short that the top of one
zooecium projects over the base of the next in series. A branch arises
about midway of the zooecium near the lateral margin. Here again the
tubular portion is very short and, as far as observed, only one branch arises
from a zooecium — never in pairs. There seem to be no other processes
arising from the zooecium and many of the cells have no tubes, branches, or
other processes arising from them, except that of the zooecium next in the
series and a dorsal process for attachment.

The Bryozoa of the Tortiigas Islands, Florida. 191

The colony appears to be very loosely attached to the Cupularia by
processes from the dorsal sides of the cells. The species occurred at various
depths — in fact, wherever this species of Cupularia was found, at from 10
to 22 fathoms, but was found nowhere else. Usually only a single colony
was found on one Cupularia, but occasionally on a large specimen two or
three colonies would occur.

Genus Synnotum Hincks, 1886.

Synnotum aviculare (Pieper).

PiEPER, 1881, p. 43 {Gemmellaria avicularis).— Hincks, 1886, p. 257.— Robertson,
1905, p. 286.

This well-known species is abundant about the islands, growing espe-
cially on shells and sponges at a depth of 8 to 10 fathoms. It has not before
been recorded from North America, except on the Pacific coast, where Miss
Robertson found it at San Pedro and San Diego, Southern California.
Hincks (1886) described the sessile avicularia as alternating on Adriatic
specimens. Robertson states that they appear on every pair in specimens
from California. In the Florida material in my collection the alternate
condition is the rule, but occasionally sessile avicularia appear on successive
pairs of zooecia. Besides the above localities the species is now known from
the Mediterranean and Red Seas, South Africa, Australia, and the Andaman
Islands.

Genus Nellia Busk, 1852.
Nellia oculata Busk.

Busk, 1852, p. 18.— Lamarck, 1816, p. 135 {Cellaria lenella). —Smitt, 1873, p. 3.—
Jelly, 1889, p. 94 (Farcimia tenella, references and synonymy). — Waters,
1909, p. 167 {Farcimia oculata). — Levinsen, 1909, p. 120 {Nellia tenella).

Abundant at 10 to 18 fathoms, frequently attached to sponges or
occasionally to shells. The usual height of the colonies is 0.5 to i inch.
One large, branched colony was taken, measuring 4 inches across and of
equal height. Smitt recorded the species from 13 to 138 fathoms. It has
also been recorded from Texas and St. Thomas, West Indies (Levinsen),
and is widely distributed in the Atlantic, Pacific, and Indian Oceans down
to a depth of 550 fathoms.

Waters (1909, p. 167) objects to the use of Lamarck's name tenella on
the ground that the short description given by Lamarck is not diagnostic
and "might be applied to several species or even genera."

Genus Scrupocellaria van Beneden, 1844.
Scrupocellaria cornigera (Pourtales).

PouRTALES, 1867, p. Ill {Cauda cornigera). —?iTAiTT, 1872, p. 14 {Cellularia cornigera).—
Jelly, 1889, p. 61 {Cellularia cornigera).

One small specimen of this species was taken at 10 fathoms and several
others at 15 fathoms. The radical fibers are characteristically armed with
stout, retrorse spines, as figured by Smitt. This alone is not sufficient for
determining the species, as Busk has noticed serrate radicals in S. ferox

192 Papers from the Marine Biological Laboratory at Tortugas.

Busk and S. macandrei Busk, and Waters has named a species from the
Red Sea S. serrata on account of this character. Florida specimens of
S. cervicornis Busk show similar spines occasionally on the basal radicals.
S. cornigera is a very delicate species and the radical fibers are transparent
and thread-like, and this fact alone will serve to distinguish it from S.
cervicornis.

Pourtales took the species at 270 fathoms, so it has a rather wide range
in depth and temperature.

Scrupocellaria cervicornis Busk.

Busk, 1852, p. 24. — Smitt, 1872, p. 14 {Cellularia cervicornis). — ^Verrill, 1900, p. 594.

Occurs commonly about the islands, from low water to 18 fathoms,
growing on piles of docks, attached to stems of gorgonias, etc. Colonies
about 0.5 inch in height, with ocecia containing eggs, were taken from the
bottom of a skiff on June 23, which had been in the water only since May i.
Recorded by Smitt from 7 to 17 fathoms.

The radical fibers, which are abundant at the base of the colony, are
usually smooth, but in a few cases they bear retrorse spines like those of
S. cornigera. In such cases the species are easily distinguished by the
fibers alone, as they are much coarser in cervicornis. The cervicorn spines
on the outer distal extremity of the zooecia are well marked.

Genus Canda Lamouroux, 1816.
Canda caraibica Levinsen.

Levinsen, 1909, p. 142.

This species has been recently described by Levinsen, who gives no
further indication of its distribution than is embodied in his remark "plenti-
ful material of a West Indian species." Levinsen suggests the probability
that this species is identical with C. simplex Busk, but being unable to
determine this point he gives a new name to the West Indian material.

Several colonies were taken by the writer at the Tortugas at a depth of
15 fathoms, attached to shells. The largest of these was not much over
0.25 inch in height and no ovicells nor avicularia were present. The
absence of an opercular spine and the length of the membranous frontal
area (two-thirds of the whole zooecium) show this form to be that described
by Levinsen. The vibraculum in some cases is longer than the width of
the branch.

The other Florida species of this genus, C. retiformis Pourtales (1867,
p. no), which Levinsen wrongly attributes to Smitt, is quite distinct from
C. caraibica, since it possesses a broad opercular spine, the membranous
area is but little more than half as long as the zooecium, and there are
differences in the vibracula. C. retiformis was not found at the Tortugas.

The Bryozoa of the Tortugas Islands, Florida. 193

Genus Membranipora Blainville, 1834.

Membranipora membranacea (Linne).

Linn6, 1766-68, p. 1301 {Flustra membranacea), — Packard, 1867, p. 274. — Robertson,
1908, p. 267. — Norman, 1909, p. 286.

One colony on the bottom of a skiff. In ev^ery respect this colony
agrees with European specimens in my collection, except that there is a
very slight extension of the calcification over the area at the base of some
of the zooecia. This is so slight, however, that I can not consider it of
specific value. The species was not listed for Floridian waters by Smitt,
nor was it taken by the writer at Woods Hole, Massachusetts, in several
years collecting.

Packard recorded it from the Labrador coast. It is a common European
species, occurring as far south as Madeira (Norman). It is also found in
Australia and New Zealand, and Miss Robertson (1908) states that on the
Pacific coast of North America it occurs from California to Alaska.

? Membranipora lacroixii (Audouin).

AuDOUiN, 1826, p. 240 {Flustra lacroixii). — Busk, 1852-4, vol. 2, pi. i. — Dawson, 1859,
p. 256. — Packard, 1867, p. 8. — Smitt, 1873, p. 18. — Jelly, 1889, p. 162
{M. reticulum). — Waters, 1898, p. 697. — Whiteaves, 1901, p. 97. — Robert-
son, 1908, p. 261. — OSBURN, I912, p. 227.

Taken in drift incrusting the carapace and legs of a large crab, and at
8 fathoms on shell. The specimens taken at Tortugas are much more
heavily calcified than those from Massachusetts and between the zooecia
are numerous, roughly triangular prominences with triangular, membranous
areas, as figured by Busk (1. c, pi. i), while these are almost wanting in the
Massachusetts specimens. Examination of the reverse side of the zooecium,
however, reveals the transparent circles and the irregular projections
described by Waters (1. c).

There has been much difference of opinion concerning this species and
there can be but little certainty in regard to many earlier records, such, for
example, as those of Dawson and Packard. Smitt's figures of Florida
specimens seem, without question, to represent the same species as my own,
though the dorsal surface is not shown. The range on the eastern coast of
North America is from Florida to the Gulf of St. Lawrence, or to Labrador,
if Packard's record is valid. Elsewhere it is known from Europe, New
Zealand, and from California northward to Alaska.

Membranipora tehuelcha (d'Orbigny).

D'Orbigny, 1839, pi. 17 {Flustra lehuelcha). — Bosc, 1802, p. 118 {Flustra tuberculata) .
— Jelly, 1889, p. 167 {M. tehuelcha) and p. 168 {M. tuberculata). — Waters,
1898, p. 674. — Verrill, 1900, p. 594 {Biflustra dentata). — Norman, 1909,
p. 286 {Membranipora tuberculata) .--<)sbvrn , 1912, p. 231.

This species is found on Sargassiini bacciferum (gulfweed) wherever it

floats. It is abundant at the Tortugas. Verrill notes its occurrence at

Bermuda, under the name of Biflustra dentata, as "common on Sargassum

found upon the beaches." Smitt did not record it for the Florida region.

14

194 Papers from the Marine Biological Laboratory at Tortugas.

Norman prefers to use Bosc's older name for this species on the ground
that, while his description means nothing, Bosc stated that it occurs "en
immense quantite sur les fucus nageans dans I'Atlantique." As I am not
aware that, under the rules of nomenclature, the habitat alone is considered
a sufficient diagnosis, I prefer to follow Waters in using d'Orbigny's name.
Membranipora irregularis d'Orbigny.

D'Orbigny, 1839, p. 17, pi. vni, figs. 5 and 6.— Smitt, 1873, p. 8, pi. n, fig. 63. — Jelly,
1889, p. 151, for synonymy. — Waters, 1904, p. 31.

Incrusting shells at 8, 15, 18, and 22 fathoms. Smitt's figure is very

satisfactory and there can be no doubt that the present species is the one

which he figured as taken at 60 fathoms. The granulation of the border

varies with the amount of calcification. Ooecia are present in some numbers.

In younger stages these are quite prominent, but with later calcification they

become included in the general crust.

Membranipora savartii (Audouin).

AuDOUiN, 1826, p. 240 {Flustra savartii). — Smitt, 1873, p. 20 {Biflustra savartii). —
Waters, 1909, p. 137.

One small, branching specimen 0.5 inch in height was collected at a

depth of 10 fathoms. The broad, internal, basal denticle figured by

Waters for Red Sea specimens is the best means of determining the species.

Recorded by Smitt, 29 fathoms.

Genus Cupularia Lamouroux, 1821.

Cupularia guiniensis Busk.

Busk, 1854, p. 98. — Busk, 1859, p. 87 (C. canariensis). — Smitt, 1873, p. 10 {Mem-
branipora canariensis). — Norman, 1909, p. 289, synonymy.

Abundant at 10 fathoms, especially on soft bottom with algae. The
umbrella-shaped disks range in size from 0.125 to 0.75 inch in diameter.
The color of living colonies is horn brown, due mainly to the chitinous
bristles which form the mandibles of the avicularia. When touched the
bristles stand erect for some time.

Recorded by Smitt from Pourtales's collections from 10 to 44 fathoms.
Smitt listed this species under Membranipora because of the character of
the zooecium, holding the zoarial characters as of no importance. Just
why he should have retained the related C. lowei and C. doma in Cupularia
is not certain, but probably on account of the greater amount of calci-
fication of the front wall.

The under surface of the colony is usually clear and free of any attached
animals or plants, except that in numerous cases there occurs, closely
attached, a new species of Beania, which I have named B. cupulariensis .

Cupularia lowei Busk.

Busk, 1854, p. 99. — Smitt, 1873, p. 14 (C umbellata Manzoni). — Verrill, 1878, p. 305
(C. umbellata). — Norman, 1909, p. 290.

Only two specimens, both dead, were taken, one at 12 and the other at

22 fathoms, in company with C. guiniensis. Each was about 0.25 inch in

diameter.

The Bryozoa of the Tortiigas Islands, Florida, I95

Recorded by Smitt for 29 fathoms and also in 7 fathoms off the mouth of
the Cape Fear River. Smitt regarded C. johnsoni Busk (syn. C. donia
d'Orbigny) as a highly calcified form of C. lowei {umbellata), but later inves-
tigations do not support this view.

Verrill (1. c.) recorded the species with doubt from one worn specimen
taken at Fort Macon, near Beaufort, North Carolina. As the writer has taken
specimens at the same locality, Verrill's record may be considered good.

Genus Cribrilina Gray, 1848.

Cribrilina floridana (Smitt).

Smitt, 1873, p. 23 {Crihrilina figularis and local variety floridana). — Jelly, 1879,
p. 66 (calls attention to fact that C. figularis of Smitt is not Johnston's species,
and lists C. floridana as a distinct species).

A number of small colonies of this species taken at 5 to 15 fathoms on
shells. Smitt recorded it from 29 to 42 fathoms. It is not known to occur
outside of the Florida region.

No ooecia nor avicularia are developed on my specimens. The posterior
margin of the orifice is straight, and the orifice is completely filled by the
semicircular operculum, which is well chitinized and of a brownish color.
There are two short, stout oral spines and in young zooecia there is fre-
quently a third median spine.

There is a possibility that this species may belong to the genus Puellina
of Jullien as amended by Levinsen (1909, P- I59). but the immature colonies
in my collection seem rather to belong to Cribrilina in the strict sense.

Genus Arachnopusia Jullien, U
Arachnopusia monoceros (Busk).

Busk, 1854, P- 72 (Lepralta monoceros). — Jelly, 1889, p. 67, synonymy. — Levinsen-,,
1909, p. 160 (discusses relationship and adopts Jullien's genus Arachnopusia.
with emendations).

A single specimen of this widely distributed species was taken at 5,

fathoms, incrusting a shell. The colony was small, measuring only 0.125;

inch across, but the zooecia ranged in development all the way from the

ancestrula to adults with ovicells and from these to marginal individuals

in various conditions of calcification. The species has not before been

taken in North American waters.

Genus Smittipora Jullien, 1881.

Smittipora abyssicola (Smitt).

Smitt, 1873, p. 6 {Vincularia abyssicola). — Jelly, 1889, p. 253, synonymy and refer-
ences.— HiNCKS, 1881, p. 155; 1887, pp. 161 and 164.

Found at low water and down to 15 fathoms on shells, bases of gorgonias,
etc. Smitt recorded two specimens taken by Pourtales, one at 68 fathoms,
the other at 450 fathoms. The bathymetrical range of the species, from low
water down to nearly 0.5 mile, is worthy of note.

Only the incrusting phase was seen, with no evidence of the erect stems
described and figured by Smitt in one colony. The family relations of this

196 Papers from the Marine Biological Laboratory at Torttigas.

genus are very uncertain. Hincks (1881, p. 155) placed it in the Micro-
poridae, but later (1887, p. 154) he removed it from this family on account
of "the double layer of the ectocyst," and placed it in the Steganoporellidse.
As we understand the latter family at present, this genus can not be included
in it. It will require a careful study of sectioned material to determine the
real position of the genus.

Genus Steganoporella Smitt, 1873.

Steganoporella magnilabris (Busk).

Busk, 1854, P- 62 {Memhranipora tnagnilabris) . — Smitt, 1873, p. 15 {Steginoporella
elegans). — Verrill, 1900, p. 594 {Steginoporella elegans). — Harmer, 1900,
pp. 279-286 (revision, synonymy, and detailed description). — Levinsen, 1909,
pp. 167-168 (discussion of family and genus).

Common in low water to 15 fathoms. Beautifully convoluted, frill-like
colonies were taken on piles and on shells, sponges, coral, etc., at 12 to 15
fathoms. The erect frill-like extensions are sometimes 2 inches high and 3
inches broad. These are bilaminate and are sometimes tubular in form.
The color in life varies from pink to reddish brown.

Smitt recorded the species in Pourtales's collections from 15 to 37
fathoms; Bermuda (Verrill) ; other distribution given by Harmer (I.e., p. 280).

Steganoporella connexa Harmer.

Harmer, 1900, p. 254.

Dead colonies of what appears to be this species were taken incrusting
shells at a depth of 12 fathoms. The opercula are all lost from the colony,
so it is impossible to make use of this important structure for identification.
The character of the calcification, with the junction of the median process
with the sides of the zooecium to form a pair of opesiulse or foramina, is
similar to that described and figured by Harmer. Since this connection
of the median process with the zooecial walls is known only in S. connexa,
I refer my specimens to that species. The other points, as far as I can make
them out on the dead specimens, agree with connexa, except perhaps in the
matter of size, since the zooecia in my specimens are smaller than described
by Harmer. There is much variation in the matter of size, however, a
condition also stated by Harmer to occur in connexa. Only one form of
zooecium can be distinguished.

The species is known from a single locality, "John Adams Bank,"
the location of which is somewhat in doubt, but Harmer seems to feel
satisfied that it is south of Abrolhos Island off the coast of Brazil.

Genus Thalamoporella Hincks, 1887.

Thalamoporella rozierii (Audouin).

AuDOUiN, 1826, p. 239 (Flustra rozierii). — (?) Smitt, 1873, p. 16 {Steginoporella rozi-
erii).— Levinsen, 1909, p. 181 (discusses at length, with five varieties).

Found on shell in drift and at 10 fathoms on sponges. Only small,

unilaminar colonies without ooecia were taken. Smitt found only a single,

small specimen in Pourtales's collections. His description of this specimen

The Bryozoa of the Tortugas Islands, Florida. 197

is brief and his figure (pi. 4, fig. 102) does not show the avicularium nor the
spicules, so I include it in the above synonymy with some doubt. I have
received from Professor A. E. Verrill a fine, erect, convoluted, bilaminatc
colony taken on the west coast of Florida. Levinsen (I.e., p. 183) records
the variety lahiata from Jamaica and my Florida specimens agree fairly
closely with this variety.
Thalamoporella granulata Levinsen.

Levinsen, 1909, p. 188.

In the drift were found several dead
specimens incrusting shells and calcareous
algffi. The sinuated character of the ori-
fice with the strong lateral denticles; the
nearly equal opesiulae reaching the mar-
ginal wall and with granulated edges;

, , , . 1 . ^\^ i-U Fig. 8. — Thalamoporella granulata Levinsen.

the dorsal Opesmlar outgrowths; tne The variety in which the avicularium

, ,1 , r , 1 • 1 11^. chamber is closed basally, except for two

granulated character of the zooecial walls ; opesiute.
the form of the large, spatulate avicularian

mandible, and the character of the spicules, all approach closely to T.
granulata. The only difference to be observed is in the basal portion of the
avicularian chamber. When completely calcified this structure becomes
closed by a calcareous lamina up to the hinge denticles of the mandible,
leaving only two openings similar in appearance to the opesiulae of the
ordinary zooecia, except that they are much smaller and are removed from
the lateral wall.

Thalamoporella falcifera (Hincks).

HiNCKS, 1880&, p. 380 (Steganoporella rozieri, (orm falcifera). — Levinsen, 1909, p. 186.

A single specimen of this species, which is distinguished by the falcate
mandible of the avicularium, was taken in shallow water growing on algae.
Apparently this species grows only on algae, if one may judge from the
comparatively small number of records. It has a wide distribution and is
recorded from Campeche Bank, Yucatan; Java Sea; Geographe Bay, Aus-
tralia, and from latitude 23° 30' N., longitude 40° W.

Genus Savignyella Levinsen, 1909.

Savignyella lafontii (Audouin).

AuDOUiN, 1826, p. 242 {Eucratea lafotilii).— Jelly, 1889, p. 93 (synonymy under

Eucratea lafontii). — Norman, 1909, p. 295 (Calenaria lafontii, later synonymy

and discussion of relationships).— Waters, 1909, p. 131 {Catenaria lafontii,

later synonymy and discussion of generic name). — Levinsen, 1909, pp. 213

(note) and 274 {Savignyella, new genus and reasons for change in generic

name).

This widely distributed tropical species is abundant about the islands

from low water down to 10 fathoms on shells, piles, and sponges, in company

with Amathia, Bowerbankia, etc. Ocecia with eggs in various stages of

198 Papers from the Marine Biological Laboratory at Tortugas.

development are common. A number of colonies with ooecia containing
eggs were on June 23 taken from the bottom of a skiff which had been in the
water only since May i. The color in life is dark brick-red, which makes
the colonies quite conspicuous against the background in spite of their
small size. The presence of this species in American waters has hitherto
escaped notice.

While there has never been any question as to the validity of this spe-
cies, there has been great confusion as to its generic position ever since its
discovery. This has been sufficiently discussed by Norman (1. c). Waters
(1. c), and Levinsen (1. c). Waters remarks "though it certainly never
ought to have been placed under Catenaria, it seems best to leave it here,"
but Levinsen solves the problem by the erection of a new genus Savignyella.
This seems the only proper course, since the species does not have generic
relationship with any of the older genera {Catenaria, Eucratea, Alysidium,
Catenicella) in which it has been placed by previous authors.

Genus Hippothoa Lamouroux, 1821.
Hippothoa distans MacGillivray.

IMacGillivray, 1868, p. 130.— Jelly, 1889, p. 112 {Hippothoa flagellum), references. —
Waters, 1904, p. 54, synonymy and references.

This minute species occurs abundantly at the Tortugas from low water
to 12 fathoms, spreading over the inside of dead shells. Ooecia are abun-
dantly developed . The ' ' short processes arising from the side of the zooecium ,
below the lateral tubular branches," described and figured by Waters (1. c.)
as occurring on specimens from Sydney, New South Wales, are developed
on many of the zooecia. I am unable to add anything to Waters's descrip-
tion of this organ, except that it is constricted at the base where it is attached
by a movable joint to the zooecium.

The species is very widely distributed in European seas, New Zealand,
Australia, Pacific Ocean, Cape Horn, etc., but has not hitherto been noticed
on the American side of the Atlantic.

Genus Trypostega Levinsen, 1909.

Trypostega venusta (Norman).

Norman, 1864, p. 84 (Lepralia venusta) and 1909, p. 299. — Gabb and Horn, 1862,
p. 127 {Lepralia inornata). — Smitt, 1873, p. 37 {Gemellipora glabra, forma
striatula) and p. 61 {Lepralia inornata Gabb and Horn). — Jelly, 1889,
p. 233 {Schizoporella striatula), p. 237 (5. venusta) and p. 128 {Lepralia inor-
nata).— Levinsen, 1902, p. 23, and 1909, p. 281 {Trypostega venusta).

On shells, Steganoporella and OcuUna, at 5 to 15 fathoms. Smitt

recorded it as Gemellipora glabra forma striatula, incrusting nullipores at a

depth of 68 fathoms, one colony, and as Lepralia inornata, two colonies from

26 and 60 fathoms. After careful study of the material at hand, I am unable

to distinguish more than one species, or to separate it from the venusta of

Norman, even as a variety. The form of the aperture, the umbonate knob

below, the puncturing, the character of the dwarf zooecia and their relation

The Bryozoa of the Tortugas Islands, Florida. 199

to the ooecia, all appear to me to be identical. Levinsen (see above) has
recently made this species the type of a new genus, Trypostega, characterized
especially by having the ooecia covered by dwarf zooecia with scattered
pores and a minute orifice.

This species is widely distributed, but on the American coast is known
only from the Florida region. Levinsen (p. 281) included Gabb and Horn's
and Smitt's records for Lepralia inornata, but failed to include Smitt's
Gemellipora glabra forma striatula as a synonym. Smitt (pp. 61-62) sepa-
rated the species on the following ground: "The most characteristic differ-
ence, which has caused us to place them in different families, depends on
the shape of the zooecial aperture," and his figures show some differences, but
these may be accounted for by variation. In a single colony in my posses-
sion the variation is as great as that shown by Smitt's two specimens.

Gexus Adeona Busk, 1884.

Adeona violacea (Johnston).

Johnston, 1847, p. 325 {Lepralia violacea) . — (?) Reuss, 1847, p. 85 {Cellepora heckeli). —
Smitt, 1873, p. 30 {Porina violacea and Porina plagiopora Busk). — Jelly,
1889, p. 184 (Aficroporella heckeli), synonymy. — Verrill, 1901, p. 54 (Porina
plagiopora). — Norman, 1909, p. 296 (Reptadeonella violacea). — Levinsen,
1909, p. 283, discusses genus and species.

Taken from 5 to 18 fathoms. Both the typical form and the nominal
variety plagiopora (Busk) occur and in some colonies the avicularia are
intermediate in position between that of violacea and plagiopora. Occa-
sionally the characters of both "varieties," together with intermediate con-
ditions, may be observed in the same colony. The usual color is purple,
varying to nearly black, or through pale bluish to white.

Smitt recorded the species "from the depth of 35 fathoms W. off Tortu-
gas" (Porina violacea), and "at a depth of 60 fathoms W. off Tortugas"
{Porina plagiopora), and Verrill has noted the form plagiopora at Bermuda.
Otherwise the species has not been recorded from American waters.

Levinsen (1909, p. 282) has returned to the use of Busk's family Ade-
onidfe, thus again breaking up the arrangement established by Hincks in his
family Microporellidae. The latter family was based on the presence of a
frontal pore (ascopore) as the principal character; but, as Levinsen has
shown, this brought together a number of forms which are separated by
more fundamental characters.

Genus Bracebridgia MacGillivary, 1886.

Bracebridgia subsulcata (Smitt).

Smitt, 1873, p. 28 (Porina subsulcata). — Verrill, 1901, p. 54 {Porina subsulcata). —
Hincks, 1880a, p. 76 {Microporella stibsulcata).

Abundant from 10 to 12 fathoms. Branching colonies reach a height

of 2 inches or more. The color in life varies from yellowish pink to orange;

when dead, white. Smitt recorded the species from 10 to 48 fathoms and

a dead specimen from 471 fathoms. Verrill has recorded it from Bermuda,

but otherwise it is known only from the Florida region.

200 Papers from the Marine Biological Laboratory at Tortugas.

I have placed this species in the genus Bracebridgia with some hesitation,
since, as this genus is understood, it possesses no ascopore (Levinsen, 1909,
pp. 283, 289). In subsulcata a minute pore is usually present below the
suboral avicularium; a similar pore appears above the basal avicularium
when this organ is present. This latter pore was not figured nor mentioned
by Smitt. I have not been able to determine positively the nature of this
structure and it may be, from the fact that it is often wanting and also
because of its presence in connection with the basal avicularium, that it
is not an ascopore at all, or that it represents this organ in a vestigial
condition.

The nature of the colony and the zooecia with the marginal and other
ribs, the form of the orifice, the presence of a line of special independent
avicularia on each edge of a branch, all seem to agree with Bracebridgia.
B. pyriformis (Busk) has a flattened area below the orifice where the avicu-
larium is located in subsulcata, but the fact that MacGillivray found a
single avicularium in this position in one colony of B. pyriformis is sufficient
to show the relationship. Moreover, in the rare cases in which the oral
avicularium is wanting in subsulcata there is a flattened area similar to that
of pyriformis.

Genus Retepora Smitt, 1867.

Retepora marsupiata Smitt.

Smitt, 1873, p. 67. — Busk, 1884, p. 116 {Retepora atlantica). — Gabb and Horn,
1862, p. 138 {Phidolophora labiata). — Jelly, 1889, p. 212 (Jelly adopts Busk's
name atlantica for this species, but Smitt's publication of the name marsupiata
clearly has priority).

Taken on a number of occasions at 10 to 18 fathoms. The largest colony

measured about an inch in height. The color in life is a delicate pink.

Smitt records the bathymetrical distribution as 16 to 262 fathoms.

Genus Rhynchozoon Hincks, 1891.
Rhynchozoon tuberculatum n. sp. (Fig. 9.)

Zooecia small, rather evenly swollen, separated
by slightly raised marginal walls; the surface, ex-
cept in young individuals, strongly tuberculate.
Pores are wanting, except for occasional very
minute ones at the margin. Orifice ovate, the
broader distal border often somewhat straight;
at one side near the proximal border a strong
pointed tooth extends often more than half-way
across the orifice and curves backward ; on the op-
posite side a minute projection sometimes appears, pi^. c,.— Rhynchozoon mbercuia-
but is often entirely absent. The peristome is thin TL^^Zm^^oS Some!
and raised high above the primary orifice, usually andoviceii.
erect at the sides, but flaring slightly backward behind the orifice; above
the large tooth the peristome appears to bear a minute, short-elliptical
avicularium, but I am not able to distinguish a mandible in my specimen.

The Bryozoa of the Tortugas Islands, Florida. 201

The ovicell is globose, at first smooth, but later tuberculate like the
zooecium ; on either side at the widest part and near the base is a rounded
thin area appearing membranous; in complete calcification the peristome
rises over the front of the ovicell and fuses with the outer layer, leaving a
thin, somewhat triangular area on the front.

One colony at 18 fathoms, incrusting a shell with a single layer of
zooecia.
Rhynchozoon soHdum n. sp. (Figs. 10, 11, and 12.)

Fig. 10.— Rhynchozoon solidum n. sp. Enlarged, showing primary aperture, spines, two

pairs of denticles, and the beaded vestibular arch.
Fig. II. — The same. Operculum. ... . , j • „ii

Fig. 12.— The same. Fully calcified, showing aviculanum, penstome, and oviceii.

Zooecia small, little swollen, rather broad, closely set in a continuous
crust, thick-walled, the surface smooth or irregularly traversed by very
fine lines. As a rule only three pores are seen, one at the posterior margin
and the others on either side of the orifice at the margin. Orifice evenly
rounded, a little broader than long, with a rounded sinus in the proximal
border. This sinus is really double, being formed by two sets of denticles,
the lower pair equal in size and inclosing a semicircular sinus much like
that of many Schizoporellas . Immediately above them is another pair of
denticles, more pointed and a little longer than the lower ones, and often
one of them is a trifle larger than the other; they are turned inward more
than the lower pair, so as to inclose a narrower and more nearly circular
sinus. The operculum is well chitinized , yellowish in color, the dots showing
attachment of muscles situated nearly half-way from the circumference
toward the center. Through the operculum can be seen a beaded vestibular
arch very similar to that figured by Levinsen (1909, pi. xxiii, fig. ^e) for
R. angulatiim. In young cells there are four short, stout spines about the
anterior margin of the orifice. The peristome is high and thick, extending
upward and forward on the sides into thick, lappet-like processes, usually a
little unsymmetrical ; posteriorly the peristome extends upward and over
the orifice in the form of a broad mucro, though sometimes this is but little
developed.

Small avicularia, sometimes a pair, frequently only one, are placed
posteriorly not far from the margin, with the triangular beak pointing
laterally.

202 Papers from the Marine Biological Laboratory at Tortugas.

The ooecia are globose, raised, smooth, without pores; a thin extension
of the peristome runs upon the top and fuses with the outer layer of the
ovicell, leaving a thinner area upon the front.

One small colony incrusting a shell at 8 fathoms.

Arborella, new genus.

Zoarium erect, dichotomously branched, with flexible corneous joints
at the bifurcations. Zooecia arranged in pairs, back to back, each pair
facing at right angles to the preceding pair of the series, giving a somewhat
quadrangular appearance. Front wall porous, but without special en-
larged pores. Fertile zooecia shorter than the infertile. Orifice with a
distinct sinus. Ovicell external, but inserted partly under the front of the
zooecium, its orifice closed by the zooecial operculum. Spines and avicu-
laria absent. A clear, chitinous, uncalcified ectocyst covers the zooecium
and ovicell.

The relationships of this genus can not as yet be stated positively.
The general appearance of the colony and especially the character of the
joints and the membranous ectocyst seem to relate it to Tubucellaria, but
the character of the ooecium and the manner of origin of the zooecia at the
tips of the branches, together with the absence of a special median frontal
pore, distinctly separate it from this genus. It may be necessary to erect
a new family to accommodate it. Sectioned material will be studied to
determine the points necessary for a more complete understanding of the
relationships.
Arborella dichotoma n. sp. (Figs. I 3 to 15.)

Zoarium erect, dichotomously branching, with hinged, flexible joints at
the points of bifurcation; internodes consisting of 8 to i8 pairs of zooecia,
each pair set slightly in advance and at right angles to the preceding pair
of the series, facing in four directions. Zooecia broad fusiform, wedge-
shaped at the base where they are inserted between the preceding pair.
The front wall evenly arched, cryptocyst well calcified, porous, with no
specialized, larger pores. Ectocyst uncalcified and covering the calcareous
wall as a clear layer. The orifice is evenly rounded in front and on the
sides, somewhat straighter on the posterior margin, with a broad, shallow,
but very evident sinus. The operculum is compound, well chitinized, with
an arched rib on either side, extending inward and upward from the hinge
denticle. Avicularia and spines are wanting. The fertile zooecia are some-
what different in form from the ordinary ones, being shorter and somewhat
broader, and the ooecium incloses nearly all of the orifice. The ooecium is
rounded in outline, large and prominent, the aperture looking upward and
closed by the zooecial operculum. The ooecial wall is porous like that of
the zooecium and is covered by the clear ectocyst.

The colony was evidently broken off in dredging and no description can
be given of the base or the mode of attachment.

The Bryozoa of the Tortugas Islands, Florida.

203

One colony 0.5 inch high, composed of 24 internodes, taken at 10 fathoms
on a sponge bed between Loggerhead and Garden Keys, Tortugas.

Fig. 13. — Arbor ella dichotoma n. sp. Portion of colony, showing mode of growth.
Fig. 14. — The same, showing details.
Fig. is.- — The same, operculum.

Genus Tubucellaria d'Orbigny, 1850-52.
Tubucellaria cereoides (Solander).

SoLANDER, 1786, p. 26 {Cellana cereoides). — Waters, 1907, p. 129. — Levinsen, 1909,
P- 305-

One small colony about an inch in height was taken at a depth of 15
fathoms. This species has not before been taken on the American side of
the Atlantic.

Waters makes T. opuntioides Pallas a synonym of cereoides, but Levinsen
believes that they should be kept separate. This difference of opinion,
which has existed since the discovery of the form, does not seem likely to
reach a settlement soon. The single specimen taken by me at the Tortugas
is not sufficient to decide the point. It agrees closely with specimens of
cereoides of the same size taken at Naples, but cereoides has hitherto been
limited to the Mediterranean by those who separate the species, while
Atlantic specimens are considered to be opuntioides.

Genus Escharella (part) Gray, 1848.

flscharella costifera n. sp. (Fig. 16.)

Zooecia of moderate size, rather regularly ovoid in form, evenly swollen,
nearly hyaline. Around the margin is a row of rather large pores, between

204 Papers from the Marine Biological Laboratory at Tortugas.

Fig. i6. — Escharella costifera
n. sp. Details of front
wall, mucro, aviculariutn,
jointed spines, vestibular
arch (in upper figure),
and ovicell (in lower
figure).

which rise strong, conspicuous ribs which converge regularly toward the
central portion of the front wall. They fade out, however, before they
reach it, leaving a smooth central area devoid of ribs. The orifice, which is
placed far forward, is rounded in front, more straight
on the posterior border, and the hinge denticles are
rather conspicuous. A few zooecia seem to show a
slight median denticle, but I can not be certain of
this. The peristome, which is very thin, rises in front
and on the sides into a short tube, and is beset with
six or eight long spines which are jointed at inter-
vals. Posteriorly the peristome rises into a high,
pointed mucro which projects forward over the ori-
fice. This mucro is flat and thin, with a strong rib
on each edge, looking as though it originated from
two spines fusing at the tip, leaving an open space
between. The development of the mucro, however,
shows that this is not the case. The thin area forming
the middle of the mucro is rounded below and pointed
above and looks as though it were meant to harbor
an avicularium, but no such structure is present.

Avicularia are present on the zooecium, usually situated near the margin
on either side of the mucronate process. Occasionally only one is pres-
ent. The avicularium is comparatively large, the mandible long triangular,
projecting sidewise or a trifle backward and pointing strongly upward.
The ocecium is rounded in outline, prominent, imperforate, smooth on the
upper surface, or with fine, radiating lines. Around the margin is a strong,
raised rib, within which there Is a row of large pores with strong ribs running
upward as on the zooeclal wall. The anterior pair of spines is inclosed
within the ovicell.

Two small colonies, the largest scarcely over 0.125 ii^ch across, in-
crusting algse, at 2 fathoms. In spite of the small size nearly every cell is
adult, with a complete ocecium.

I am a little uncertain as to the generic position of this species. Es-
charella as amended by Levinsen (1909, p. 315) includes only forms without
avicularia. In spite of this difference the species seems to fit better into
Escharella than Into any related genus. The absence of zooeclal and ooecial
pores, except the marginal ones, the well-developed vestibular arch, the very
thin operculum, the character of the orifice and peristome, and the nature of
the jointed spines, seem to locate the species in Escharella. In many respects
it resembles E. diaphana MacGillivray, as figured by Levinsen (1. c),
except for the presence of the avicularia and the strong development of the
ribs, which are wanting In diaphana.

The Bryozoa of the Tortiigas Islands, Florida. 205

Genus Schizoporella Hincks, 1880.

Schizoporella unicornis (Johnston).

Johnston, 1847, p. 320 {Lepralia unicornis). — Desor, 1848, p. 66 (Lepralia vari-
olosa).— Leidy, 1855, P- 10 (Escharina variabilis). — Smitt, 1873, p. 44 (Hip-
pothoa isabelleana). — Verrill and Smith, 1873, p. 713 (Escharella vari-
abilis).— Verrill, 18756, p. 41 {Hippothoa variabilis); ibid., p. 41, pi. ill,
fig. I {Hippothoa reversa, n. sp.); 1878, p. 305 {H. variabilis); 1879b, p. 193,
and 1880, p. 30 (Escharina isabelliana d'Orbigny, E. reversa Verrill and
E. ansata Gray); 1900, p. 592 (? Schizoporella isabelliana). — Levinsen, 1909,
p. 323.— Osburn, 1912, p. 236.

Abundant from low water to 10 fathoms, in the drift and growing
profusely on the piles of docks, incrusting shells, and coral rockb, and rising
into branched, finger-like processes about worm-tubes and the stems of
hydroids, or occasionally growing free. Masses nearly as large as a man's
head were dredged. The color varies from white or pale pink or purple in
young, rapidly growing colonies, to a dark purplish red or sometimes nearly
black. Colonies 1.5 inches in diameter were found incrusting the bottom
of a skiff that had been in the water only from May i to June 23.

Smitt recorded the incrusting form "growing in a single layer of whitish
and yellowish hue on Ocidina and Nullipora at the depth of 42 fathoms,"
and, "a more compound colony, composed of several concentric layers of a
purplish-blue tint, around an axis of foreign matter," at a depth of 17
fathoms.

Schizoporella floridana n. sp. (Figs. I 7 and 18.)

Primary characters of the zooecium much like those of S. unicornis
(Johnston) but with a slightly more V-shaped sinus; avicularia situated far
forward, often in advance of the middle of the orifice, and with the mandible
usually strongly curved toward the zooecial axis; large vicarious avicularia
with long-pointed, straight mandibles on hemispherical cells raised high
above the surface of the colony.

The naked-eye appearance of the colonies is very similar to that of
S. unicornis (Johnston) and the general characters of the zooecia are not
very different from that species. It does differ remarkably, however, in
the presence of the large, vicarious avicularia, which are sometimes very
abundant. These are raised high above the surface of the colony as rounded,
knob-like prominences; the mandibles, while broad at the base, are very
much narrowed and elongated distally. These are turned in all directions
on the colony without any apparent arrangement. Schizoporella ampla
Kirkpatrick, from Mauritius, and 5. viridis Thornley, from the Red and
Mediterranean Seas, also have vicarious avicularia.

The ordinary avicularia are situated at the side of the orifice and the
mandibles are curved toward the axis of the zooecium so as to nearly con-
form to the curve of the margin of the orifice. There is much variation in
this avicularium and the mandible is occasionally straight instead of curved,
and it may be directed backward or sideways. It is usually situated well
forward near the anterior border of the orifice, but may be placed farther

206

Papers from the Marine Biological Laboratory at Torlugas.

back, occasionally behind the orifice. The zooecia are swollen about as in
unicornis, the punctures are similar to that species, and the form of the
orifice (with the sinus) might readily pass for that species. No oral spines
have been observed. In the very young stage a low, raised margin separates
the cells, but this is soon obscured by the secondary calcification. An
umbo is developed below the orifice, apparently only by secondary calci-
fication. In young colonies, where the vicarious avicularia are not yet
developed, this species may readily be mistaken for S. unicornis, but the
more anterior position and the curved form of the avicularia will serve to
distinguish it even in this condition.

The ooecium differs from that of any Schizoporella with which I am
acquainted. It is comparatively small, very short (about two-thirds as
long as wide), very high and prominent, smooth, and imperforate.

This well-marked species, which appears to be undescribed, was dredged
in 15 to 18 fathoms, incrusting shells and hard sponges, and, in one case,
forming an erect nodular growth nearly an inch in height and about half
an inch in diameter around some foreign matter as a center. The color
ranges from pure white to dark purplish-red.

Fig. 17. — Schizoporella Jloridana n. sp. Details of young zooecium. • , ■

Fig. 18. — ^The same. Highly calcified condition, showing ovicells and the long, independent avicularia
mounted on large, swollen cells.

Schizoporella sanguinea (Norman).

Norman, 1868, p. 222 {Hemeschara sanguinea).— Jelly, 1889, p. 233, synonymy.
Taken at 15 fathoms on a Vermetiis shell; one colony with ovicells.
Reported by Smitt from Pourtales's collections at 60 fathoms, southwest
of Tortugas.

The Bryozoa of the Tortugas Islands, Florida. 207

Schizoporella biaperta (Michelin).

MiCHELiN, 1841-2, p. 330 {Eschara biaperta). — Smitt, 1873, P- 4^ {Hippothoa biaperta)
and p. 47 (Hippothoa divergens n. sp.). — Verrill, 18756, p. 41 {Hippothoa
biaperta); 1878, p. 305, 1879&, p. 30, and 1880, p. 193 {Escharina biaperta). —
HiNCKS, 1880, p. 258 {S. biaperta var. divergens). — Norman, 1909, p. 303
(S. biaperta var. divergens). — Levinsen, 1909, p. 323 (discusses and amends
genus). — OsBURN, 1912, p. 237.

A few colonies of this species were taken at various depths from low
water to 22 fathoms. The appearance is characteristic and the specimens
compare very well with those from the Woods Hole region, but never reach
such a luxuriant development.

Recorded by Smitt (under Hippothoa biaperta) from 9 to 60 fathoms,
and (under H. divergens new species) from 135 fathoms "forma typica,"
and 120 fathoms "forma laxa."

Smitt's "H. divergens" has been ranked by Hincks and Norman as a
variety of biaperta, and recorded by them from the British Isles and Madeira.
All specimens examined by the writer belong to the typical biaperta.

Schizoporella spongites (Pallas).

Pallas, 1766, p. 45 {Eschara spongites). — Smitt, 1873, p. 42 {Hippothoa spongites). —
Verrill, 1900, p. 592 {Hippothoa or Schizoporella spongites). — Levinsen,
1909, p. 324 (full discussion of the species).

One of the most characteristic species of the region ; abundant from low
water to 18 fathoms; incrusting shells, coral, rocks, and growing on the
surface of harder sponges; also on piles. The very large ooecia serve to
distinguish this species at a glance from any related forms. In life the color
varies from translucent white or yellow to bright brick-red.

Recorded by Smitt from Pourtales's collections from 13 to 35 fathoms
and by Verrill (1. c.) from Bermuda. Levinsen (1. c.) records it from
St. Thomas and St. John, West Indies, and from Malacca (with some
differences).

Genus Escharina (part) Milne-Edwards, 1838.
Escharina pesanseris (Smitt).

Smitt, 1873, P- 42 {Hippothoa pes anseris). — Jelly, 1889, p. 141 {Mastigophora duter-
trei var. pesanseris). — Waters, 1909, p. 169 {Schizoporella pesanseris). —
Norman, 1909, p. 302. — Levinsen, 1909, p. 326 (discusses genus, amended,
and species fully).

One small dead specimen, composed of about a dozen zooecia, was taken
at a depth of 8 fathoms. The vibracula characteristic of the species were
broken off. OcEcia present.

Smitt described this species from "two small colonies growing on a
Nullipora, west of Tortugas" at 42 fathoms. It occurs also at Madeira
in 70 fathoms (Norman), the Island of Mauritius (Kirkpatrick) , the Red
Sea (Waters), and Siam (Levinsen).

2o8 Papers from the Marine Biological Laboratory at Tortugas.

Genus Microporella (part) Hincks, 1877.

Microporella ciliata (Pallas).

Pallas, 1766, p. 38 {Eschara ciliata). — Smitt, 1873, p. 26 (Porellina ciliata). — Packard,
1867, p. 270 {Lepralia ciliata). — Verrill, 1879^, p. 29 {Porellina ciliata);
1875c, p. 53; 1880, p. 190, and 1879, p. 29 {Porellina siellata {or a form of this
species).— OsBURN, 1912, p. 233 {M. ciliata), and p. 234 (the variety stellata
Verrill). — Levinsen, 1909, p. 328 (discusses genus).

From 5 to 18 fathoms on shells, corals, etc. One fine colony growing
on an egg case of a large mollusk. Smitt records it at various depths from
7 to 60 fathoms. It is one of the most widely distributed of all bryozoa
and occurs in all oceans.

Levinsen (1. c.) places this genus in family Escharellidae.

Genus Smittina Norman, 1903.
Smittina trispinosa (Johnston).

Johnston, 1838, p. 280 {Lepralia trispinosa). — Dawson, 1859, p. 256 {Lepralia
trispinosa). — Packard, 1867, p. 67 {Lepralia trispinosa). — Smitt, 1873,
p. 59 {Escharella jacotini) , and p. 60 {E. spathulata). — Verrill, 1879&, p. 31;
1880, p. 195 {Mucronella jacotini) ; 1875a, p. 514, 1878, p. 305, and 18796,
p. 30 {Discopora nitida). — Whiteaves, 1901, p. 106 {Smitlia trispinosa). —
Norman, 1903, p. 120 {Smittina nom. nov. for Smiltia preoccupied in Dip-
tera). — Osburn, 1912, p. 246 {Smittia trispinosa).

One of the most abundant and characteristic species in the Tortugas
region from low water to 12 fathoms. A cosmopolitan species and dis-
tributed all along the Atlantic coast northward to the Arctic region. Smitt
recorded it from 13 to 44 fathoms in Florida waters.

Smitt states (p. 60) "For the peculiarity of the development of the
avicularia, the Floridan form, as a distinct variety, I propose to be named
Escharella spathulata.^ ^ But variation in the secondary characters of this
species here as in many other regions occurs to a bewildering extent. The
typical form with the large, pointed avicularia, common in northern waters,
is rare, but these are occasionally observed. One beautiful colony growing
on the under side of Cupularia guiniensis at 10 fathoms is an almost pure
example of Smitt's spathulata. Even the smaller, triangular avicularia
figured ^ - Smitt (pi. x, fig. 200) are not represented, but in a few cases
very elongated pointed avicularia occupy the place of the long-spatulate
form figured by Smitt. Another colony found in the drift is more like
Verrill's nitida in the character of the avicularia, which are small, short-
pointed, and short oval in form, very much like specimens from the southern
New England coast. In a few cases very large spatulate avicularia, as long
as the whole zooecium, were observed. In one colony short-triangular avicu-
laria were rather regularly disposed in pairs below the orifice with the man-
dibles pointing outward. Occasionally the avicularia are almost wanting
from a colony, a condition seen in many New England specimens that bear
evidence of very rapid growth in a single layer.

The ooecia vary much with the state of calcification. In younger stages
they are coarsely and sometimes very irregularly punctured. This layer
may be entirely covered by secondary calcification, when the ooecia present

The Bryozoa of the Tortugas Islands, Florida. 209

a smooth, shining appearance or are variously roughened. The secondary
layer usually grows irregularly over the primary layer, but in some cases it
forms a rather regular raised border with a rounded area of the primary
layer exposed on the top.

This form resembles closely the variety protecta Thornely, as figured by
Waters (1909, pi. 17, fig. 5), but the large avicularia are evenly rounded at
the tip instead of being divided into points. The peristome also varies
greatly. In some cases it is scarcely raised; in others it takes the form of a
tall, broad, median projection, often forked at the tip and projecting so far
over the orifice as to nearly obscure it from above. All sorts of intermediate
conditions are observed. The peristome also occasionally shows raised
lappet-like projections at the side of the orifice, as seen commonly in speci-
mens from Massachusetts (Osburn, 1912, pi. xxvii, fig. 65a, 66). In spite
of all these variations in secondary characters there is a remarkable con-
stancy in the primary orifice, denticle, oral spines, and the primary zocEcial
walls.

Another variation, and one which I am at a loss to explain, is this:
Some colonies consist of zooecia, the secondary characters of which are all
of one type or with little variation, while other colonies will show not only
zooecia representing two or more of the so-called varieties, but all sorts of
intergradations as well.

Genus Lepralia (part) Johnston, 1847.
Lepralia audouinii (d'Orbigny).

D'Orbigxy, 1850-2, p. 401 {Cellcpora audouinii). — Smitt, 1873, p. 56 {Escharella
audouinii).

This species, which is known from the Red and Mediterranean Seas,
occurs at the Tortugas from low water to 10 fathoms. Colonies taken from
the bottom of a skiff which had been in the water from May i to June 23
were an inch in diameter, with ooecia containing embryos. In life the
colonies are translucent white, but the ooecia often appear light pink when
embryos are present.

Smitt's record was based on a single colony growing on a nullipore at 37
fathoms, west of Tortugas.

Lepralia porcellana Busk.

Busk, i860, p. 283, pi. 31, fig. 3. — Smitt, 1873, p. 62 (Lepralia deidostoma). — Norman,
1909, p. 305, synonymy.

Taken at 5 to 15 fathoms on shells and on other bryozoa, such as Steg-
anoporella magnilabris and Holoporella turrita. The key-hole form of the
orifice, the pointed avicularia w^ith elongate mandibles, and the absence of
pores except occasional ones at the margin serve to distinguish the species
readily. Ooecia present. In one case there is a vicarious avicularium, some-
what larger than the usual type, mounted on a smooth, swollen chamber.
A low, smooth, umbonate process is sometimes present below the orifice.

Recorded by Smitt from Pourtales's collections from 30 to 120 fathoms.

15

2IO Papers from the Marine Biological Laboratory at Tortugas.

Lepralia uvulifera n. sp. (Figs. 19 and 20.)

Zooecia small, the surface shining and covered by small, smooth knobs.
Pores are wanting except for 3 to 5 large, marginal pores. Usually these
are placed as follows: One on either side near the orifice, one on either side
about half-way back, and a single pore at the posterior end. The orifice
is slightly broader than long, evenly rounded anteriorly, nearly straight
on the posterior border, hinge teeth scarcely noticeable. Peristome slightly
raised anteriorly and beset with about six slender spines. Posteriorly the
peristome rises into a strong, mucronate process, which in complete calci-

FiG. 19. — Lepralia uvulifera, n. sp.

Fig. 20. — The same. Portion of colony showing at the right
the orifice of a young zooecium.

fication is divided at the tip into three sharp spines. A small avicularium
is present on one zooecium, situated somewhat behind the orifice near the
lateral border, with the pointed mandible pointing backward and outward.
The ooecium is comparatively large, heavily calcified, and covered with
knobs like those of the zooecium; a high umbonate process rises on the top,
and there is a single large pore on either side of the ooeciostome near the
base; the ooecial orifice is wide and high, with a calcified projection extending
nearly vertically downward into the orifice and bearing a close resemblance
to the uvula of the human palate, except that it is a little more truncate at
the tip.

One small colony of perhaps three dozen zooecia incrusting the under
side of a Cupularia guiniensis at 10 fathoms.

This species is certainly closely related to L. water si Calvet (1906, p.
412, pi. XXVII, fig. 11) from the Cape Verde Islands, but it differs in having
the umbo strongly trifid instead of merely pointed ; in the more open form
of the ovicell, which partially incloses the anterior part of the peristome,
instead of being placed in front of it as in L. watersi; and also in lacking
the complete lower border of the ooeciostome shown in the figure of L. watersi.
Calvet's figure also shows the oral spines situated slightly outside of the
oral border, while in uvulifera the spines are on the border.

The Bryozoa of the Tortugas Islands, Florida. 211

Lepralia cucullata Busk.

Busk, 1854, p. 81.— Waters, 1909, p. 150, synonymy and references.— Jullikn et

Calvet, 1903, p. 141 {Schizoporella cucullata).
Common in shallow water at the Tortugas. In Hfe the color of the
colony is dark brownish-purple, sometimes almost black, and the polypides
in the young zooecia at the edge of the colony have a fine carmine color.
A number of colonies more than an inch in diameter were found on the
bottom of a skiff that had been in the water only from May i to June 23.
The generic position of this well-known species is very unsettled. A
majority of authors have agreed in placing it in Lepralia, others in Smittia,
and still others in Schizoporella. The reason for this difference of opinion
lies in the general simplicity of the zooecia, which lack characters usually
considered as diagnostic. The form of the orifice is somewhat intermediate
between that of Lepralia and Schizoporella. There is a very shallow curve
on the posterior border, which may be interpreted as a sinus, as Calvet has
done (1. c), or as merely the space between the hinge denticles. In the
absence of any means of settling the controversy, I place the species tenta-
tively in Lepralia, in which genus it was described by Busk.

Distribution: Mediterranean, Red, and Arabian Seas; Azores Islands;
Cape Verde Islands; South Africa, and Cape St. Lucas, Lower California.
Not before noted on the Atlantic side of America.
Lepralia rostrigera (Smitt).

Smitt, 1873, p. 57 {Escharella rostrigera).— Waters, 1885, p. 298; 1887, p. 61.— Jelly,

1889, p. 126 (under L. depressa).
The Escharella rostrigera of Smitt has been made a synonym of L.
depressa Busk by Miss Jelly (1. c), but in the opinion of the writer it repre-
sents a distinct species. Smitt recognized a relationship between the
species, but indicated that the differences are constant. The study of
material taken by me at the Tortugas confirms this view. The form of the
zocecial aperture is quite different, being well rounded at the posterior mar-
gin, as figured by Smitt, and there is never more than a faint indication
of a constriction of the sides. The avicularia are characteristically placed
far forward, alongside of the anterior part of the orifice, or frequently
entirely in front of it. Busk's figure of L. depressa (1854, pi. 91 , figs. 3 and 4)
shows no indication of pores. In rostrigera there are numerous small pores
in the frontal wall in younger stages, and large, marginal pores are present
in all stages. The larger zooecia with the broad aperture described and
figured by Smitt (pi. X, fig. 205) are occasionally present in my specimens.
Smitt's specimens were taken by Pourtales in 35 and 43 fathoms. Mine
were taken at 10 and 15 fathoms on coral and shells. A fine colony at 10
fathoms incrusted the base of a living Cladocora arhuscida.
Lepralia contracta Waters var. serrata Osburn.

Waters, 1899, p. 11.— NoRiLA.N, 1909, p. 306.— Osburn, 1912, p. 242 {Lepralia

serrata).
Several colonies incrusting shells at 5 to 18 fathoms. This variety was

212

Papers from the Marine Biological Laboratory at Tortugas.

described by me as Lepralia serrata n. sp., from the Woods Hole region
(1912, p. 242, pi. 26, fig. 57).

After a careful study of the Florida specimens I am of the opinion that
this can be only a variety of L. contractu Waters, described from Madeira
(1899, p. 11). Waters's description is brief and his figures small, but
Norman (1. c, pi. 41 , fig. 56) has refigured the species. The differences seem
to be largely in the secondary calcification, which appears to be much greater
in American specimens; also there are six oral spines in Madeira specimens,
while usually only four are to be observed in American material. The
avicularia in Florida specimens are much more abundantly developed and are
frequently raised high above the front of the zooecium on small mammillary
processes. Also, the lateral oral denticles are much more prominent than
they appear in the figures of Madeira specimens and are strongly bifid;
however, Norman figures one such bifid denticle. In one of my speci-
mens the spatulate avicularia are extremely elongated, fully as long as the
zooecium, and, while usually directed forward, are sometimes reversed in
position. The zooecial wall appears to be much more thickened than in the
Madeira specimens, to judge by the figures of Waters and Norman. Not-
withstanding these differences I believe they are closely related. The form
of the ooecium with the peculiar membranous area and the serrated char-
acter of the primary zooecial orifice are identical and are so unique that I
believe them to indicate the same species. Specimens from southern New
England are much more highly calcified than Florida specimens, which
seem in some respects to be intermediate between those from New England
and Madeira.

Lepralia edax (Busk).

Busk, 1859, p. 59 (Cellepora edax).—Sum, 1873, p. 63 (L. edax formce typica and
ca/carea).— HiNCKS, 1880, p. 311.— Verrill, 1909, p. 54.

One colony taken at 18 fathoms, forming a cylindrical mass more than
0.25 inch in diameter and many layers in thickness, which originally sur-
rounded a branched structure of some sort, and one colony taken at 8
fathoms incrusting a shell fragment. Smitt's figures and description of
this species are very complete. I have compared these specimens with one
from the English Channel sent me by Dr. S. F. Harmer. English speci-
mens are said to be found only on certain species of gastropod shells. While
Florida specimens have not been taken in such situations, I can distinguish
no differences of any importance. My Tortugas specimens belong to what
Smitt calls the forma calcarea, but it can scarcely be considered a distinct
variety.

The large, pointed, vicarious avicularia described and figured by Smitt
are abundantly distributed over the colony. The ovicell has a peculiar
membranous area on the upper surface, a feature also possessed by L.
contracta Waters, though in the latter species this area is much nearer the
zooecial aperture.

The Bryozoa of the Tortugas Islands, Florida. 213

Smitt records the species from Pourtales's collections from 49 to 79
fathoms. Verrill records it for the Bermudas in shallow water.

Lepralia janthina (Smitt). •

Smitt, 1873, p. 63 {Lepralia edax forma janlhina).

This, I believe, must be separated from edax as a distinct species. The
form of the orifice, a character to which Smitt gave undue importance, is
similar, but in janthina the orifice is not only much larger, but relatively
wider and shorter, with less space behind the hinge denticles. The avicu-
laria are all pointed and are situated on a prominence behind the orifice with
the mandible turned forw^ard and frequently slightly outward. There are
no vicarious avicularia. The color is a deep blue-black or violet, entirely
different from any specimens of L. edax I have seen. The zooecia are consid-
erably larger than those of L. edax. Ovicells are wanting.

One colony 0.25 inch in diameter taken at 6 fathoms incrusting a shell.
Smitt records one specimen from 13 fathoms on coral.

The Cellepor a janthina of Waters (1899, p. 14) from Madeira can not
be the L. edax var. janthina of Smitt, as Norman (1909, p. 311) points out
in renaming Waters's species C. rotundora.

Genus Phylactella Hincks, 1880.
Phylactella labrosa (Busk).

Busk, 1854, p. 92 {Lepralia labrosa). — Jelly, 1889, p. 204. — Norman, 1909, p. 308.

Dredged at 22 fathoms on shells and hard sponges. Ovicells are lacking.
There is no median denticle on the proximal margin of the orifice, nor are
the zooecia disposed in radiating lines, but form a solid crust. However,
Norman (1. c.) has shown that both of these characters are variable.

Phylactella collaris (Norman) var. aviculifera nov. (Fig. 21.)

Norman, 1866, p. 204 {Lepralia collaris); 1909, p. 309. — Jelly, 1889, p. 203. — Robert-
son, 1908, p. 307.

This species has not heretofore been recorded from the Atlantic side of
North America, though Miss Robertson records it from southern Cali-
fornia, and it is found on the Euro-
pean side of the Atlantic Ocean as
far south as the Madeira Islands.

At the Tortugas, growing on
shells at from i to 15 fathoms,
occurs the variety which I have
here called aviculifera. In all re-
spects this appears to agree fully _^ p^. , , „ „ • ,Kr v • ...

. , . ° ^ Fig. 21. — Phylaclella collaris (Norman) avicuhfera, var.

With F. collaris, except that there "°'*'- viewed partly from in front and showing avic-

'■ ularium.

is a rather large avicularium situ-
ated within the collar immediately below the orifice, with a triangular
mandible pointing upward, or very rarely sidewise. This appears on every
zooecium of the colonies studied. Norman has described a rounded avicu-

214 Papers from the Marine Biological Laboratory at Tortugas.

larium on the lower lip of occasional zooecia of L. labrosa from Madeira
(1909, p. 309)-

Genus Cellepora Linne, 1767.
Cellepora dichotoma Hincks.

HiNCKS, 1862, p. 304. — Smitt, 1873, P- 53. pl- IX, fig. 193-198 (C. aviadaris). — Jelly,
1889, p. 51, gives synonymy and also, p. 46, includes Smitt's references in
C. avicularis Hincks. — (?) Verrill, 1878, p. 305 (C. avicularis); (?) 1901, p.
54 {C. avicularis).

Two well-developed specimens, 0.5 inch or more in height, taken at a
depth of 10 fathoms; one was attached to a hydroid stem. There can be
no doubt that these specimens belong to C. dichotoma instead of C. avicularis,
as I have compared them with authenticated material from England.
Smitt's figures, which he refers to C. avicularis, undoubtedly are C. dicho-
toma, as Hincks (1880, p. 404) has already pointed out. Smitt indicates
the bathy metrical distribution as from 9 to iii fathoms.

Verrill's records for C. avicularis at Fort Macon, North Carolina (1878),
and Bermuda (1901) in all probability apply to dichotoma, though it is
possible that the former should be C. americana Osburn (1912, p. 238),
which Verrill at another time (1879B) listed as avicularis. C. americana and
C. dichotoma have been taken at Beaufort, North Carolina, by the writer.

Cellepora verruculata Smitt.

Smitt, 1873, p. 50. — Busk, 1884, p. 150 {Escharoides verruculata). — Jelly, 1889,
p. 60. — Calvet, 1906, p. 444.

A number of colonies taken on shells and in similar situations from low
water to 15 fathoms. Smitt described the species from a single specimen
taken by Pourtales west of the Tortugas in 42 fathoms. Recorded by Busk
(1. c.) off Heard Island at 75 fathoms, and by Calvet (1. c.) for the Mediter-
ranean Sea, "Naples, c6tes de Corse, Cette."

The zooecia undergo a remarkable change in appearance with advancing
calcification.

Genus Lagenipora Hincks, 1877.

J_agenipora ignota Norman.

Norman, 1909, p. 309, pi. 42, figs. 10-13.

A colony taken at Tortugas in 12 fathoms on a dead colony of Mucronella
bisinuata and another at 10 fathoms on a Vermetus shell seem to be identical
with Norman's species. They differ in not rising into rounded branches,
but the small size of the colonies easily accounts for this. The character of
the zooecium, with the pair of oral avicularia, is identical, and the ooecia
also. The large spatulate avicularia, said by Norman to be present in
extraordinary numbers, are but sparsely represented.

Hitherto known only from the Madeira Islands, where it was taken at
70 fathoms (Norman).

The Bryozoa of the Tortugas Islands, Florida. 215

Genus Holoporella Waters, 1909.

Holoporella albirostris (Smitt).

Smitt, 1873, p. 70 (Discopora albirostris). — Busk, 1881, p. 346 (Cellepora albirostris). —
Jelly, 1889, p. 45 (references under Cellepora albirostris). — Waters, 1885,
p, 304; 1887, p. 68 {Cellepora albirostris); 1909, pp. 159-161 {Holoporella
gen. nov.).
This species, which is very characteristic of the region, occurs in abun-
dance from low water to 15 fathoms. It grows on coral-rocks, corals, shells,
and on the firmer sponges. Recorded by Smitt from Florida waters in 25
to 35 fathoms. It is readily distinguished from other species of this region
by the presence of a towering, pointed white spine situated at one side
of the proximal margin of the orifice. Except in the youngest stages the
colony has a dark grayish or blackish color, against which the white spines
stand out in sharp contrast. A very small avicularium is situated at the
base of the spine, large spatulate avicularia are distributed over the colony,
and vestigial, minute avicularia are occasionally present at the side of the
orifice.

Waters in 1909 (1. c.) established the genus Holoporella to include, for
the most part, those species of the old genus Cellepora in which the sinus
is lacking and which have a rounded or straight posterior margin of the
orifice, and in which also the opercular muscles are attached near the
border. Levinsen (1909, p. 347) erected a new family, Holoporellidae, to
include this genus, thus farther removing it from Cellepora. This separa-
tion is based on the fact that in Holoporella the operculum is attached,
sometimes by hinge teeth, to the lateral margins of the orifice, while in
Cellepora the operculum is suspended at the edges of the sinus. Oral spines
are often present in Holoporella, but are wanting in Cellepora. Rostra, with
avicularia, may be present in both genera. Waters includes the following
American species of the Florida region in Holoporella: H. {Discopora)
albirostris (Smitt), H. {Lepralia) turrita (Smitt), and H. {Discopora) pertusa
(Smitt). Levinsen adds H. {Discopora) advena (Smitt), which apparently
completes the list of the described species of this region.

Holoporella pusilla (Smitt).

Smitt, 1873, p. 70 {Discopora albirostris forma pusilla). — Busk, 1884, p. 436 {Cellepora
pusilla).

In the "Floridan Bryozoa" Smitt describes this species as a form of
his Discopora albirostris, stating that their relationship is " incontestably
proved by the very same form of the zooecial aperture." Busk (1. c.)
objects to placing pusilla as merely a variety of albirostris, giving it as his
opinion that the two seem to him to be quite distinct. Although Busk
possessed no material of pusilla, he was evidently correct in forming this
opinion. If the single specimen taken by me at the Tortugas is identical, as
I believe it to be, with Smitt's pusilla, it differs from albirostris in the entire
absence of dark pigment; in the smaller size of the colony, the individual,
and the orifice; in the presence of radiating lines of depressions and small

2i6 Papers from the Marine Biological Laboratory at Tortugas.

wart-like prominences; in tiie smaller size of the ooecia; in the much slighter
development of the rostrum, which never rises much above the avicularium ;
and in the more procumbent position of the rostrum, which projects forward
over the orifice. The form of the orifice is, as stated by Smitt, very similar
to that of albirostris, but (as Busk points out) this form is common to a
number of species. Small hinge denticles are sometimes present. Smitt
figures the species as having from 3 to 6 oral spines. There is scarcely any
indication of these in the specimen in my possession. The oral avicularium
is usually rounded, but occasionally the mandible is obtusely triangular.
No other avicularia are present.

One colony 0.25 inch in diameter incrusting a shell at low tide.

Smitt records it from Pourtales's collections at 9 to 60 fathoms.

Holoporella magnifica n. sp. (Figs. 22 and 23.)

A very coarse species with very large zooecia and orifices large enough
to be distinguished readily by the naked eye. Taken at 10 fathoms incrust-
ing a sponge and a very small colony on a dead shell at the same depth.

s

rh

Fig. 22.— Holoporella magnifica n. sp. Young zooecium at edge of colony, showing manner of growth.

Rostrum and rostral avicularium not yet developed.
Fig. 23. — The same. Highly calcified portion near center of colony, showing one deeply immersed and three

erect superficial zooecia, independent avicularia of two sizes, and the rostral avicularium.

Zooecia large, procumbent at the growing edge, erect and irregular at
the center of the colony. Imperforate except around the edge, where there
are large pores, often coarsely cancellated between the cells in the middle
of the colony. Very heavily calcified, even in young cells. The orifice is
much as in H. albirostris, except that it is much larger. The peristome is
very thick and broad, slightly raised all around the orifice (when ooecia are
present it extends over them in later stages of growth). The rostrum is
small and low, projecting so slightly over the orifice as to scarcely hide the
posterior margin in any case. On one side of the rostrum is placed a large

The Bryozoa of the Tortugas Islands, Florida. 217

avicularium with an elliptical mandible, black in color. Sometimes the
rostrum is so slightly developed that the avicularium appears to be placed
transversely behind the orifice. The operculum is heavily pigmented with
dark brown, especially at the edge. Large, spatulate avicularia with
brown or blackish mandibles are abundantly distributed over the colony.
There is a slight amount of the brownish pigment in the zooecial wall.
In all the avicularia the pigment is confined to the mandible, which is thus
strikingly contrasted with the colorless basal membrane. In the operculum
also the color is limited strictly to the hinged portion and does not extend
upon the base. The ooecia are, as in other species of this genus, shallow and
widely open, and they are deeply immersed in the crust, except in the very
young stage.

In addition to the specimens mentioned above, a large flat specimen, 2
inches across and many layers in thickness, w^hich was apparently growing
unattached, was taken by Dr. Paul Bartsch off Biscayne Key, Florida.
It has not been possible to identify these specimens with any known species.
The large size of the zooecia and the orifice, the great thickness of the walls,
even In young specimens, the straight proximal border of the orifice, the
absence of a tall rostrum, and the large size of the oral avicularium, together
with the elongate form of the mandible with its round tip in the vicarious
avicularia, form a set of characters which serve to distinguish it.

Holoporella turrita (Smitt).

Smitt, 1873, p. 65 (LepraUa turrita). — Waters, 1883, p. 438 {Smittia turrita); 1909
p. 161 (includes in Holoporella, new genus). — Ridley, 1881, p. 55 {Cellepora
turrita). — Jelly, 1889, p. 253 (Smittia turrita).

A very common species at 12 to 15 fathoms, growing on shells, corals,
and the firmer sponges. Recorded by Smitt from Pourtales's collections
as not frequent, only a few specimens taken, at from 26 to 44 fathoms,
growing on corals and nullipores.

Color in life bright pink to brick-red. Above the spreading incrustations
rise knob-like projections, usually 0.25 inch or so in diameter, sometimes
much larger. The younger zooecia are separated by delicate, raised, white
walls which are very conspicuous against the red color of the colony. The
white points of the blunt spines are also strongly contrasted with the ground
color.

Genus Petralia MacGillivray, 1869.

Petralia bisinuata (Smitt).

Smitt, 1873, p. 59 {Escharella bisinuata). — Levinsen, 1909, pp. 350-51-
This well-marked species, which has not been noted outside of the
Florida region, is common at the Tortugas on shells and sponges, from 10 to
18 fathoms. The color in life is bright vermilion. Smitt described the
species from Pourtales's collections at 9 and 19 fathoms.

Levinsen (see above) has placed this species in MacGillivray's genus
Petralia and has erected a new family, Petraliidse, to include this genus.

2i8 Papers from the Marine Biological Laboratory at Tortugas,

CTENOSTOMATA.
Genus Bowerbankia Farre, 1837.
Bowerbankia gracilis Leidy.

Leidy, 1855, p. 142. — HiNCKS, 1877, p. 215 (Valkeria caudata); 1880, p. 521 {Bower-
bankia caudata). — Verrill and Smith, 1873, p. "jog {Vesicularia gracilis),
and p. 710 {V. fusca). — Verrill, 1880, p. 28 {Vesicularia gracilis). — Lev-
INSEN, 1894, p. 82 (Bowerbankia caudata). — OsBURN, 1912, pp. 253-245
(synonymy); 1912a, p. 287.

Specimens of this species, which do not seem to differ from Woods Hole
material, were taken in the moat at Fort Jefferson and on the piles of the
docks, attached to shells, or creeping over the sea-weed, and covered to such
an extent by coral mud as to almost obscure them. No erect branches
were noted. Many individuals show the caudate process of the variety
caudata (Hincks).

On the North American coast the species has been identified by the
writer from Nova Scotia, Maine, Massachusetts, Connecticut, New Jersey,
North Carolina, and Florida. Otherwise it is known from England and
Denmark.

It is quite probable that the genus Bowerbankia, as Waters (1910, pp.
241 and 244) suggests, must be fused with Zoobotryon, but as Waters himself
keeps them separate pending more thorough studies of the ctenostomes,
the writer has not the temerity to combine them. The only marked dis-
tinguishing characters are the much greater softness and hyalinity of the
stock and the absence of a creeping stolon in Zoobotryon. These characters
will readily separate the one species of Zoobotryon from any Bowerbankia
with which the writer is acquainted.

Genus Zoobotryon Ehrenberg, 1831.
Zoobotryon pellucidum Ehrenberg.

Ehrenberg, 1831, pi. iii, fig. 10.— Sonder in Coll. Binder {Ascothamnion trinitatis). —
Waters, 1910, p. 243.

This species, first described from the Mediterranean, is common at the
Tortugas, floating and attached, especially in the Fort Jefferson moat and
about the piles of the dock. One colony 8 inches across was observed. The
zooecia in all cases were much obscured by a layer of white coral mud
adhering to the ectocyst.

Not previously recorded for the Florida region. Sonder (see above)
recorded it as a plant from the island of Trinidad, and according to Waters
(see above) it is known from the Isle of Pines, South Australia, Zanzibar,
the Red Sea, and the Cape Verde Islands.

Genus Cylindrcecium Hincks, 1880.

Cylindrcecium giganteum (Busk).

Busk, 1856, p. 93 (Farrella gigantea).— Hoicks, 1880, p. 535.

This species, which has been recorded from the British Isles, from various
localities in the Mediterranean Sea, from the Red Sea, and from the Queen

The Bryozoa of the Tortugas Islands, Florida. 219

Charlotte Islands, has not hitherto been noted on the American side of the
Atlantic. It is abundant at the Tortugas from low water down to several
fathoms, growing profusely over piles, shells, seaweed, and sponges. The
zooecial wall in all cases was thickly covered with a layer of coral mud.
Occasional individuals showed the swelling at the side of the zooeciostome
which forms the ooecium. Embryos in various stages of development were
present.

Genus Anguinella van Beneden, 1844.

Anguinella palmata van Beneden.

Van Beneden, 1844, p. 58.— Busk, 1856, p. 95. — Osburn, 1912, p. 253, pi. xxviii,
fig. 78.

This species occurs rather sparingly among algae on the reefs in shallow
water. It is very easily overlooked because its irregularly branched form
gives it a marked resemblance to some of the smaller algae. The species is
now known to occur on the American coast from the Tortugas to Buzzard's
Bay, Massachusetts (Osburn, 1912). At Beaufort, North Carolina, it is
very abundant and grows much larger than at the localities mentioned
above. It has not previously been noticed south of Charleston, South
Carolina, where it was recorded by Busk (Harvey, coll.).

Genus Amathia Lamouroux, 18 12.
FAmathia goodei Verrill.

Verrh^l, 1901, p. 329. — (?) Smitt, 1872, p. 4 {Serialaria sp. undet.).

Several colonies, growing attached to piles, seem to belong to this
species, which Verrill described from the Bermudas. Verrill described the
branches as "thick, soft and flaccid" and the zooids as "numerous, arranged
in large, dense elongated clusters composed of several close rows, which
often nearly or quite surround the stem, but are scarcely at all spiral."

BIBLIOGRAPHY.

AuDOUiN, J. V. 1826. Explication sommaire des planches des mollusques, des annelides,
des crustaces, des arachnides, des insectes, des echinoderms, des ascides
de L'Egypte et de la Syrie, par G. C. Savigny. Paris.

Beneden, p. J. VAN. 1844. Recherches sur I'anatomie, la physiologie et le developpe-
ment des bryozoaires qui habitent la c6te d'Ostend. Nouveau Memoires
Academic Royale Belgique, t. xvill, 1845. Bruxelles.

Bosc, L. 1802. Histoire naturelle des vers, vol. 3.

Busk, G. 1852-4. Catalogue of the marine polyzoa in the collection of the British
Museum. Chilostomata, part i, 1852; part 2, 1854.

1856. Zoophytology. Quarterly Journal of Microscopical Science, vol. 4, pp.

93-96, pi. 95-96.

1859. The polyzoa of the Crag. Palaeontographic Society.

i860. Catalogue of the polyzoa collected by J. Y. Johnson at Madeira, etc.

Quarterly Journal of Microscopical Science, vol. 8, pp. 280-285, pi. 30, 31.

1881. Polyzoa collected by Capt. H. W. Feilden in the North Polar Expedition;

with descriptions of new species. Journal of the Linnean Society,
Zoology, vol. 15, pp. 231-241, pi. 13.

1884. Report on the polyzoa collected by H. M. S. Challenger, pt. i, Cheilo-

stomata. Vol. x, pt. xxx, pp. (xxiv) 1-2 16, pi. i-xxxvi. London.

1886. Idem, pt. 2, Cyclostomata, Ctenostomata, and Pedicellinea. Vol. xvir,

pt. L, pp. (viii) 1-47, pi. i-x.

Calvet, L. 1906. Expedition Scientifique des Travailleur et du Talisman. Pp. 355-
495, pi. 26-30.

Couch, R. Q. 1844. Cornish Fauna. Vol. 3.

Dawson, J. W. 1859. Geological Survey of Canada for 1858. Polyzoa, pp. 255-257.

Desor, E. 1848. Ascidioidian polyps or bryozoa (from Nantucket). Proceedings of
the Boston Society of Natural History, vol. 3, pp. 66-67.

Ehrenberg, C. G. 1831. Symbolae Physicae, pt. 2, Evertebrata. Pages not numbered,
pi. 3, fig. 10.

Fleming, J. 1828. A history of British animals, exhibiting the descriptive characters
and systematical arrangements of the genera and species of quadrupeds,
birds, reptiles, fishes, Mollusca, and Radiata of the United Kingdom.
First ed., 8vo, Zoophytes, pp. 505-554. Edinburgh.

Gabb, W. M., and Geo. H. Horn. 1862. Monograph of the fossil polyzoa of the second-
ary and tertiary formations of North America. Journal of the Academy
of Natural Sciences of Philadelphia, vol. 5, pt. 2, pp. 111-179, pi. 19-21.

Gray, J. E. 1848. List of British Animals in the British Museum, part i.

Harmer, S. F. 1891. On the British species of Crisia. Quarterly Journal of Micro-
scopical Science, n. s., vol. 32, pp. 127-181, pi. 12.

1900. A revision of the Genus Steganoporella. Quarterly Journal of Micro-

scopical Science, n. s., vol. 43, pp. 225-297, pi. 12-13.
HiNCKS, T. 1862. A catalogue of the zoophytes of South Devon and South Cornwall.
Annals and Magazine of Natural History, sen 3, vol. 9, pp. 303-310. pl- 12.

1877. On polyzoa from Iceland and Labrador. Annals and Magazine of Natural

History, ser. 4, vol. xix, pp. 97-112, pi. x, xi. London.

1880. British marine polyzoa. Vol. I, 601 pages of descriptive matter; vol. il,

83 pi. to accompany text. London.

1880A. Contributions toward a general history of the marine Polyzoa. Annals

and Magazine of Natural History, ser. 5, vol. 6, pp. 69-91, pi. 9-II.

1880B. Same title, pt. 2 and 3, same journal, ser. 5, vol. 6, pp. 376-384.

1881. Same title, pt. 4, same journal, ser. 5, vol. 7, pp. 147-162, pi. 8-10.

1881A. Same title, pt. 8, same journal, ser. 5, vol. 8.

1887. Critical notes on the polyzoa. Same journal, ser. 5, vol. 19, pp. 150-164.

Jelly, E. C. 1889. A synonymic catalogue of the recent marine br>'Ozoa. 322 p.

London.
Johnston, G. 1838. History of British zoophytes. London. Bryozoa confused with
other groups under Ascidioida. Pp. 238-324, pi. xxix-XLiii.

1847. Idem, 2d ed. Polyzoa, pp. 253-406, pi. xlvi-lxxiv.

The Bryozoa of the Tortugas Islands, Florida.

221

JULLIEN, J. 1881. Note sur une nouvelle division des Brj'ozoaires Cheilostotniens. Bul-
letin Societe Zoologique du France, vol. 6, pp. 271-285.

1888. Mission Scientifique du Cap Horn. Br^^ozoaires. Tome 6, Zoologie,

pp. 1-92, pis. 1-15.
JuLLiEN, J., et L. Calvet. 1903. Brj'ozoaires provenant des campagnes de L'Hirondelle.

Resultats des campagnes scientifique du Prince de Monaco, fasc. xxiii,

pp. 1-188, pi. i-xviii. Monaco.
Lamarck, J. B. 1816. Histoire naturelle des animaux sans vertebres, vol. 11, 1" ed.,

Paris. Bryozoa scattered among "Polypes."
Landsborough, D. 1852. A popular history' of British zoophytes. London. Polyzoa,

pp. 265-386, pis. 15-20.
Leidy, J. 1855. Contributions toward a knowledge of the marine invertebrate fauna

of Rhode Island and New Jersey. Journal of the Academy of Natural

Sciences of Philadelphia, ser. 2, vol. 3 (Polyzoa on pp. 9-1 1).
Levinsen, G. M. R. 1894. Mosdyr (Polyzoa eller Bryozoa). In: Schiodte, J. C,

Zoologica Danica, 4de bd., iste afd., pp. 1-105, pi. l-ix. Kjobenhavn.

1909- Morphological and systematic studies on the cheilostomatous bryozoa.

364 pages, 24 plates. Copenhagen.
LiNNE, C, 1758. Systema naturae, ed. 10, vol. I. Lithophyta and zoophyta, pp. 789-821.
Holmiae.

1767. Idem, ed. 12, vol. i, pt. 2. Lithophyta and zoophyta, pp. 127-1337.

Holmiae.
MacGillivr.\y, p. H. 1869. Descriptions of some new genera and species of Australian

Polyzoa, etc. Transactions and Proceedings of the Royal Society of

Victoria, part 2, vol. 9, pp. 126-148.
MiCHELiN, H. 1 841-2. Iconographie zoophytologique . . . des polypiers fossiles de

France, . . . xii, 348 p., atlas of 79 pi. Paris.
Norman, A. M. 1864. On undescribed British hydrozoa, actinozoa, and polyzoa.

Annals and Magazine of Natural History, ser. 3, vol. 13.

1903. Notes on the natural history of East Finmark. Annals and Magazine of

Natural History, ser. 7, vol. 12, pp. 87-128, pi. 8-9.

1909. The Polyzoa of Madeira and neighboring islands. Journal of the Linnean

Society, vol. xxx, pp. 275-314, pi. 33-42.
Orbigny, a. d'. 1839. Voyage dans I'Amerique meridionale. \'ol. 5, pt. 4, Bryozoa,
pp. 7-23, pi. i-io. Paris.

1850-2. Paleontologie frangaise. Terrains cretaces. Vol. 5, pp. 1-1191, pi.

600-800.
OSBURN, R. C. 1912. The bryozoa of the Woods Hole region. Bulletin of the United
States Bureau of Fisheries for 1910, vol. 30, pp. 201-266, pis. 18-31.

1912A. Bryozoa from Labrador, Newfoundland, and Nova Scotia, collected by

Dr. Owen Bryant. Proceedings of the United States National Museum,
vol. 43, pp. 275-289, pi. 34.
Packard, A. S. 1863. List of animals dredged near Caribou Island (Labrador). Cana-
dian Naturalist and Geologist for 1863, pp. 406-412.

1867. Invertebrate fauna of Labrador and Elaine. Proceedings of the Boston

Society of Natural History, vol. i, pp. 66-69.
Pallas, P. S. 1766. Elenchus zoophytorum. Hagae Comitum.
PlEPER, F. \V. 1 88 1. Eine neue Bryozoe der Adria, Gemellaria (?) avicularis. Jahres-

ber. Westfalischen Prov.-Vereins, vol. 9, pp. 43-48, pi. I.
Pourtales, T. F. de. 1867. Contributions to the fauna of the Gulf Stream at great

depths. Bulletin of the Museum of Comparative Zoology, vol. i. No. 6

(Bryozoa on pp. 106 and iio-iii).
Reuss, A. E. 1847. Die fossilen Polyparien des Wiener Tertiar-beckens. Wicn.
Ridley, S. O. 1881. Account of the zoological collections made during the survey of

H. M. S. Alert. Proceedings of the Zoological Society of London.
Robertson, Alice. 1905. Non-encrusting chilostomatous bryozoa of the west coast of

North America. Ltniversity of California Publications, Zoology, vol. 11,

No. 5, pp. 235-322, pi. 4-16.

1908. The encrusting chilostomatous bryozoa of the west coast of North America.

University of California Publications, Zoology, vol. IV, No. 5, pp. 253-

* ^ 344. Pl- 14-24-

Smitt, A, F. 1872-3. Floridan Bryozoa, collected by Count L. F. de Pourtales. Kongl.
Svenska Vetenskaps-Akademiens Handlingar, pt. I, 1872, in bd. 10,
No. II, pp. 1-20, taf. i-iv; pt. 2, 1873, in bd. il. No. 4, pp. 1-83, taf. I-XIII.
Stockholm.

Solander, D. 1786. Natural history of many curious and uncommon zoophytes, col-
lected from various parts of the globe by the late John Ellis, syste-
matically arranged and described by the late D. Solander. London.

222 Papers from the Marine Biological Laboratory at Tortugas.

Stimpson, W. 1853. Synopsis of the marine invertebrata of Grand Manan. Smith-
sonian Contributions to Knowledge (Bryozoa, pp. 17-19, pi. i).

Verrill, A. E. 1875. Brief contributions to zoology from the Museum of Yale College,
No. 32. American Journal of Science and Arts, vol. 9 (Bryozoa on p. 414).

1875A. Idem, No. xxxiii. Results of dredging expeditions off the New England

coast in 1874. Ibid., vol. x. Bryozoa, pp. 41-42, pl- ni.

1878. In: Coues and Yarrow, Notes on the natural history of Fort Macon, North

Carolina, and vicinity. Proceedings of the Academy of Natural Sciences
of Philadelphia. List of polyzoa by Verrill on pp. 304-305-

1879. Preliminary check-list of the marine invertebrata of the Atlantic Coast

from Cape Cod to the Gulf of St. Lawrence. Pp. 28-31. (Published
privately, New Haven, Connecticut, April 1879.)

1879A. Brief contributions to zoology from the Museum of Yale College, No.

XLiii, Notice of recent additions to the marine fauna of the eastern coast
of North America, No. 6. American Journal of Science and Arts, vol.
XVIII, pp. 52-54. New Haven. , a 1 •

1880. Notice of recent additions to the marine invertebrata of the Atlantic

coast of America. Proceedings of the United States National Museum,
vol. 2, 1879 (published 1880). Polyzoa, pp. 188-196.

1900. Additions to the Tunicata and MoUuscoidea of the Bermudas. Zoology

of the Bermudas, vol. i (Bryozoa on pages 592-594. figs- 4 and 6 of pl. 20).

1901. Additions to the fauna of the Bermudas from the Yale Expedition of 1901.

Zoology of the Bermudas, vol. i (Polyzoa on p. 54).

190IA. Review of recent papers relating to the fauna of Bermuda. American

Journal of Science, vol, 11 (on page 29 in a footnote is a description of

Amathia goodei n. sp.).
Verrill, A. E., and S. I. Smith. 1873. The invertebrate animals of Vineyard Sound and

adjacent waters. Report of the Commissioner of Fish and Fisheries for

1871-2. (Br^^ozoa, pp. 707-714 and p. 747.) Washington.
Waters, A. W. 1883. Fossil chilostomatous bryozoa from Muddy Creek, Victoria, etc.

Quarterly Journal of the Geological Society, vol. 39, pp. 423-443, pl- 12.

1885. Chilostomatous bryozoa from Aldinga and the River-Murray Cliffs,

South Australia. Same journal, vol. 41, pp. 280-310, pl. 7.

1887. On tertiary chilostomatous bryozoa from New Zealand. Same journal,

vol. 43, pp. 40-72, pis. 6-8.

1898. Observations on Membraniporidae. Linnean Society Journal, Zoology, voL

XXVI, pp. 654-693, pl. 47-49- London.

1899. Bryozoa from Madeira. Jour. Roy. Micros. Soc, pp. 6-16, pl. 3.

1904. Resultats du voyage du S. Y. Belgica. Bryozoa, pp. I-103, pl. 1-9.

1906, Bryozoa from Chatham Island and d'Urville Island, New Zealand. Annals

and Magazine of Natural History, ser. vii, vol. xvii, pp. 12-23, pl. i.

1907. Tubucellaria: Its species and ovicells. Linnean Society Journal, vol. XXX,

pp. 126-133, pl. 15, 16.

1909. Reports on marine biology of the Sudanese Red Sea, etc., xii, Bryozoa

Cheilostomata, part i, Linnean Society Journal, vol. xxxi, pp. 123-181,
pl. 10-18.

1910. Same title, xv, Bryozoa Cyclostomata, Ctenostomata, and Endoprocta, pt. ir,

same journal, vol. xxxi, pp. 231-256, pl. 24, 25. r- ,

Whiteaves, J. F. 1901. Catalogue of the marine invertebrata of Eastern Canada,
Geological Survey of Canada, Ottawa. Polyzoa, pp. 9I-I44-

2(U

MBL WHOI LIBRARY

H IflEU I