LI B RAR.Y

OF THE

UNIVERSITY
OF ILLINOIS

590.5
FI

v.32

•mow

Return this book on or before the
Latest Date stamped below.

University of Illinois Library

APR «

AUG 1 7 ;

364

L161— H41

DEVELOPMENT AND GROWTH

OF THE RATTLE
OF RATTLESNAKES

ARNOLD A. ZIMMERMANN

AND

CLIFFORD H. POPE

FIELDIANA: ZOOLOGY
VOLUME 32, NUMBER 6

Published by

CHICAGO NATURAL HISTORY MUSEUM
MARCH 29, 1948

DEVELOPMENT AND GROWTH

OF THE RATTLE
OF RATTLESNAKES

ARNOLD A. ZIMMERMANN

Professor of Anatomy, College of Medicine
University of Illinois

AND

CLIFFORD H. POPE

Curator of Amphibians and Reptiles
Chicago Natural History Museum

FIELDIANA: ZOOLOGY

VOLUME 32, NUMBER 6

Published by

CHICAGO NATURAL HISTORY MUSEUM
MARCH 29, 1948

THE LIBRARY OF THE

APR 1 5 1943

UNIVERSITY OF ILLINOIS

PRINTED IN THE UNITED STATES OF AMERICA
BY CHICAGO NATURAL HISTORY MUSEUM PRESS

J I

FI.

v,

CONTENTS

PAGE

INTRODUCTION AND LITERATURE 357

ACKNOWLEDGMENTS 360

MATERIAL AND METHODS 360

ANATOMY OF THE FULLY DEVELOPED RATTLE 363

GROWTH CHANGES IN THE END-BODY OF A LIVING TEXAS DIAMOND-BACK . . 367

DEVELOPMENT OF THE SHAKER OR STYLE 374

DEVELOPMENT OF THE RATTLE IN YOUNG OF THE FLORIDA DIAMOND-BACK . . 381

Morphological Features 381

Histological Features 386

Histology of the End-body 386

Phenomena of Degeneration and of Resorption 393

Phenomena of Accretion 398

GROWTH OF THE RATTLE IN THE MASSASAUGA 403

SUMMARY 407

REFERENCES 412

INDEX . . 413

355

INTRODUCTION AND LITERATURE

The rattle of rattlesnakes is a unique structure found only on
the New World pit-vipers (Crotalidae) of the genera Crotalus and
Sistrurus. Although this structure has been discussed in scientific
literature for more than two hundred years, there is still no adequate
account of its remarkable method of development and growth.

In 1940 Dr. L. M. Klauber, the noted authority on rattlesnakes,
published a 62-page paper carefully describing the rattle and discuss-
ing its growth and every aspect of its natural history. Because he
used neither histological techniques nor x-ray photography, Klauber's
account of growth failed to clarify the confusion that already existed.
His valuable paper is recommended to all who are interested in other
aspects than growth; it is especially useful for the mechanical inter-
pretation of the rattle's structure and action. Since his paper includes
an extensive bibliography with 145 titles, we have given only a brief
one.

There has been a vast amount of speculation on the function of
the rattle. Klauber (1940) discussed the various theories and con-
cluded that only the "warning theory" is reasonable. One cannot
doubt that the rattle is used to frighten intruders, and its sound,
therefore, may be compared with the growl of a wolf or the snarl
of a tiger. Any mechanism that helps the animal to avoid physical
struggle may be assumed to have selective value. We prefer to call
the sounding of the rattle a bluffing rather than a warning reaction,
since obviously the snake is merely trying to avoid damage to itself.

Certainly it is of special interest that the deaf rattler protects
itself by vibrations for which it has no sensory perception. This
reptilian bluffing is more appropriately directed against intrusions
by "higher" warm-blooded enemies with an acute sense of hearing.

The most notable description of the rattle's form and growth was
given as early as 1857 by Joh. Czermak, who studied scant material
of Crotalus durissus terrificus. Many modern German handbooks
of comparative anatomy (see Bolk and others, 1939) still quote
Czermak and reproduce some of his diagrams. Leuckart (1855) had
previously made some observations on this structure. He had
noticed, for instance, that a rattle evidently is not present at birth;

357

358 FIELDIANA: ZOOLOGY, VOLUME 32

there is merely a simple horny covering with a direct transition to
the scales of the skin. He had assumed that only three caudalmost
vertebrae were fused into a single flattened skeletal end-piece or
style.

Czermak corrected this point, after having counted seven or
eight full-grown elements in the "end-body of the vertebral column,"
as he called the style. However, he realized that differences might
be due to age or to species. Czermak compared this conical, bi-
laterally flattened style and its two rounded, more or less separate
points to a simple exostosis. He correctly observed that the vertebral
canal continues into the end-body and that it has rudiments of
intervertebral foramina "so that the spinal cord undoubtedly extends
also into the end-body." Obviously he had prepared the caudal
end of the vertebral column by maceration, as he also observed the
vascular canal below the centra and stated that "it probably con-
tains blood vessels."

Czermak recognized three main muscular masses associated with
the tail vertebrae: two dorso-lateral masses and one inferior mass,
each of which is subdivided into several strands and layers. He
noticed that this strongly developed musculature terminates at the
base of the style and produces the rapid vibrations of the caudalmost
vertebral column and of the end-body. The movements are trans-
mitted to the rattle, "whose individual segments vibrate and glide
on each other — whereby a very peculiar, hissing sound results."
Czermak observed the thickened skin covering the end-body, from
which each separate unit of the rattle is produced. He also noticed
that a deep furrow, covered by the last scales, separates the end-
body from the rest of the tail.

The connective tissue of the end-body was described by him as
a "spongy, whitish mass, but relatively firm, consisting of interwoven
fibers." He even made microscopic sections, evidently hand-made
razor cuts, and described ramified pigment cells everywhere within
the whitish connective tissue stroma and particularly toward the
surface, where some pigment cells mingle with "roundish compact
cells, so that the outermost layer of the cutis appears dark." He
actually observed pigment cells between the epidermal cells of the
cutis, an observation that our histological preparations have fully
confirmed some ninety years later. Besides pigment cells and col-
lagenous fibers, Czermak noted in the stroma of the end-body
numerous nerves and bloodvessels and an apparent absence of elastic
fibers.

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 359

Our own observations show that Czermak was a thorough and
astute observer, considering the little material and the simple
technique at his disposal.

This pioneer, however, went beyond the direct observations of
facts and attempted, "partly with certainty, partly with probability,"
to deduce the mode of formation of the rattle. In this mixture of
fact and fancy, of good observation and old-fashioned "logical"
deductions he failed. Parts of his explanation are such that a clear
picture can scarcely be obtained.

Nevertheless, much information was gained from the two speci-
mens at his disposal. "Both were," he pointed out, "in that phase
of development of the rattle, where the youngest or basal element
of the rattle had just been fully formed and was lying directly
adjacent, cap-like, to the thickened skin of the end-body." The
following points in Czermak's description are in agreement with our
own findings: Each segment is formed as an epidermal cap or cover-
ing of the thickened skin over the end-body. The segment is
separated later, like the stratum corneum of the remainder of the
epidermis, from the substratum. Each segment must be the exact
mold of that skin thickening. Differences in size and form of the
segments must be closely correlated with differences in the epidermal
covering of the end-body occurring during growth of the animal and
during the successive steps in the development of the rattle. Recent
segments cannot be mere duplications in form and size of earlier
ones. The new segments would burst open the older ones. They
form, instead, a partly telescoped row of cap-like segments. A new
lobe must be added to the proximal segment, this addition taking
place in the furrow covered by the last scales.

In his attempt to explain the shift and changes in form and size
of the epidermal thickening over the end-body, Czermak is vague
and his assumptions are difficult to follow. This part of his explana-
tion is pure speculation. He believed himself to have "correctly
sketched, in general, the mode of formation of the rattle in Crotalus
and to have discovered a new and interesting developmental process."
Czermak was aware that detailed observations of various steps in
this complicated growth process "are left to later, more extensive
studies." He specifically pointed out the need for finding and exactly
determining the several suggested (theoretical) stages in the develop-
ment of the rattle, for a study of the process of keratinization and
for determining whether a new segment is being formed with each
molting. In spite of many deficiencies, Czermak's stimulating paper
pointed the way for our own investigation.

360 FIELDIANA: ZOOLOGY, VOLUME 32

ACKNOWLEDGMENTS

This paper could not have been written without the generous
cooperation of the Department of Anatomy, College of Medicine of
the University of Illinois, and the Department of Zoology of Chicago
Natural History Museum. We are indebted to the following ad-
ministrators of these institutions: Dr. 0. F. Kampmeier, Head of
the Department of Anatomy; Colonel Clifford C. Gregg, Director
of Chicago Natural History Museum; and Mr. Karl P. Schmidt,
Chief Curator of its Department of Zoology.

We are grateful for assistance received from many other sources.
Dr. L. M. Klauber, Mr. D. D wight Davis, Dr. Rainer Zangerl, and
Mr. Leon L. Walters gave welcome advice and criticism. Technical
work was done by Miss Jane Bostrom and Mrs. Dorothy Foss at
the University of Illinois and at Chicago Natural History Museum.
Miss Norma Lockwood, also of the Museum, and Miss Nancy Joy
and Miss Zelma Oser of the Illustration Studios, College of Medicine,
University of Illinois, gave expert assistance in the preparation of
many illustrations from original camera lucida drawings made by
the senior author. The technical difficulties in obtaining good
x-ray negatives and appropriate enlargements for measuring purposes
were overcome with the skillful help of Messrs. Arthur L. Hesse
(x-rays) and Lawrence A. Toriello (photography), both of the College
of Medicine of the University of Illinois. Living or freshly preserved
snakes were supplied by Messrs. J. E. Johnson, Richard C. Snyder,
Ross Allen, and R. Marlin Perkins and the Lincoln Park Zoo,
Chicago. The rattlesnake from which we obtained an x-ray record
of one complete molting cycle was lent to us by the Chicago Zoological
Society through the courtesy and cooperation of Director Robert
Bean, Mr. Robert Snedigar, and Mr. Emil J. Rokosky.

MATERIAL AND METHODS

Our material was of three kinds: (1) Tails selected by careful
examination of hundreds of preserved rattlesnakes in the study
collection of Chicago Natural History Museum. These were not
suitable for histological work. (2) Well-preserved, fresh tails secured
especially for us. These, fixed in formalin, were suitable for histo-
logical study. (3) The living adult specimen of Crotalus atrox
studied by means of x-ray photography throughout one molting cycle.

The species studied were the Texas and Florida diamond-backs
(Crotalus atrox and C. adamanteus) ; the timber rattlesnake (Crotalus

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 361

horridus horridus); the prairie and Pacific rattlers (Crotalus viridis
viridis and C. viridis oreganus); and the massasauga (Sistrurus
catenatus catenatus). The living Texas diamond-back studied by
means of the x-ray came from New Mexico, although it was reared
in a zoo; the other specimens of this species were from Texas, except
a series of live juveniles of unknown origin. The Florida diamond-
backs came from Alabama. The timber rattlers were caught in
Wisconsin, the prairie and Pacific rattlesnakes in Nebraska and
southern California, respectively, the massasaugas in Wisconsin
and in the Chicago region.

The preserved tails were examined macroscopically as well as
microscopically. Camera lucida drawings were made of all important
specimens before they were dissected or sectioned. Gross dissection
and hand sectioning aided considerably in the early stages of the
work. Alizarine staining of bone and clearing by a modified Spalte-
holtz method was done on many specimens. X-ray photography
helped in working out details of growth of the rattle and of structural
relationships of the bony style. Tails were bisected longitudinally
before they were embedded and sectioned. Decalcification of the
bony structures within severed tail-ends was effected by immersion
in 2 per cent nitric acid for twelve hours. The slow embedding in
celloidin was accomplished by routine technique.

The sections were made at 25, 30, and 40 microns, stained
routinely with Delafield's hematoxylin and eosin and mounted in
Canada balsam. It might have been preferable to embed the young
tail material in one piece instead of first slicing it in half. Yet there
were compensating advantages: we gained, before embedding, mid-
sagittal orientation-views as well as better penetration of celloidin.
Although preparing snake material for microscopic study was a
novel experience to us, reptilian scales and hornified elements of the
rattle did not present special difficulties.

Investigation of rattle-accretion in the living, adult diamond-
back was accomplished by taking a series of x-ray photographs
throughout a molting cycle. We found ourselves confronted with
the difficult problem of proper exposure for structures of very
different radiopacity. According to the chemical analysis by C. E.
White, given by Klauber (op. cit.), the mineral ash content of the
rattle is only 2.53 per cent. The style and the vertebrae, however,
have a high degree of mineralization, and even the rattlesnake skin
contains almost four times as much mineral ash as the rattle itself.
The large, muscular masses within the tail, and the connective tissue

FIELDIANA: ZOOLOGY, VOLUME 32

in the end-body also produced surprisingly strong shadows in the
x-ray negatives.

This high differential between strong and very slight radiopacity
further necessitated special manipulation by the photographer in
making the enlarged positive prints. For measurements of growth
changes and for good reproductions in printing it was desirable to

CARTON

FIG. 57. A diagram of the equipment needed in getting the live Texas dia-
mond-back (Crotalus atrox) into a tube preparatory to making an x-ray exposure
of its tail. The snake was coaxed from a bag, supported by a carton, into a 5-foot
tube made of rolled sheets of cellulose acetate. The tube was then detached from
the bag and the snake safely carried to the x-ray table.

obtain two- and four-fold enlargements. We used the special Kodak
x-ray film, Industrial type M. This fine-grained, slow-exposure film
gave satisfactory results.

A novel method was used in bringing the living rattlesnake to
the x-ray table (fig. 57). The basic piece of equipment was a five-
foot tube made by rolling to a suitable bore several sheets of cellulose
acetate and holding them in position by rubber bands. One end of
this strong, transparent tube was closed by a loosely fitting, movable
stopper controlled by a piece of wire; the other end remained open.
We then placed a tapering sack in a carton so that the mouth of the
sack, held open by having its border turned outward over the rim
of the box, could be quickly closed by pulling a loop of string. One
bottom corner of the sack had a small hole into which the open end
of the tube, after being passed from the outside through a corre-

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 363

spending hole in one corner of the box, was fastened by adhesive
tape. The snake was gently hooked and dropped into the sack,
which was easily closed by tightening the loop of string. Sack and
tube were then pulled out of the box together.

Although it was usually easy to coax the snake from the sack
into the tube, patience was ever necessary, because an aroused snake
will bite through a sack. Once the snake had crawled in, the tube
was separated from the sack, and the snake safely carried to the
x-ray table. The stopper allowed us to keep the snake from crawling
too far into the tube; fortunately the reptile seldom tried to back
out, but any effort of that kind was readily controlled. On the
x-ray table the forward end of the tube was covered with leaded
rubber pads to darken and steady it. The x-ray exposures of the
live tail were made at a target distance of 30 inches with 32,000 V.
and 30 m.a., and best results were obtained with an exposure time
of seven seconds. Fine-grained Industrial type M Kodak films were
used in order to obtain satisfactory enlargements from the negatives.
The latter were developed for eight minutes at 65° F.

To release the snake we merely removed the stopper and quickly
placed that end of the tube in the cage. The snake then slowly
crawled out. During a period of three months we made thirty-two
exposures of this adult rattler with little difficulty or danger. More-
over, the valuable reptile was not in the least harmed.

ANATOMY OF THE FULLY DEVELOPED RATTLE

The general anatomical features of the rattle, well known to
herpetologists, have been described by previous authors. Klauber's
(1940) comprehensive paper on this structure makes a detailed
description by us unnecessary. He gives an adequate terminology,
and lays emphasis on structural features from the point of view of
mechanics. However, in view of our frequent references to the rattle
as a whole, it seems advisable to discuss and illustrate certain aspects
of gross structure.

Figure 58 shows the complete rattle of a relatively young Pacific
rattlesnake (Crotalus viridis oreganus) with seven articulated units
or segments, which are partially telescoped into each other like a
column of hats. The snake measured 765 mm. from snout to vent.
The same rattle is also shown disarticulated, the characteristic shapes
and proportions of the individual segments thereby becoming visible
in detail. Each segment was drawn by camera lucida, with the apex

364

365

366 FIELDIANA: ZOOLOGY, VOLUME 32

at the level that it occupied in the rattle. Eight hornified segments
are thus seen in a staggered arrangement from the oldest and terminal
segment or button to the youngest and largest unit. The latter is
most proximal and lies directly caudal to the last scales. This young-
est horny segment intimately caps the subjacent tissues of the end-
body, which consists essentially of living epidermis, derma, connec-
tive tissue, and the bony style.

Differences in size among the successive segments are readily
recognized. The total length of the button, in the specimen under
discussion, was 6 mm., whereas the length of the most proximal seg-
ment was 10 mm. The greatest width in the dorso-ventral direction
was 6.2 mm. for the button and 12 mm. for the proximal segment.
All segments are about as wide as long. The total length of the rattle
before its disarticulation was 4 cm.

There are striking differences in shape between the segments.
The larger proximal units consist of three distinct lobes. They may
be termed conveniently (a) the basal or proximal lobe, (6) the
middle lobe, and (c) the distal or terminal lobe. The last one shows
a more pronounced lateral longitudinal groove than the other two.
The button consists of only two lobes, which correspond to the
distal and middle ones of the other segments.

In young snakes the distal lobe of successive segments shows a
characteristic change in shape and a relative increase in size. The
rattle from a mature specimen of Crotalus viridis oreganus is illus-
trated in one piece as well as in its disarticulated form (fig. 59).
Obviously, a number of distal segments had been lost, and the
retained portion of the rattle does not allow an estimate of the age
of that snake, which measured 725 mm. from snout to vent. A
detailed comparison of the disarticulated segments reveals but little
change in size or shape among them, the distal lobes in particular
being nearly alike. Evidently the snake was a fully grown specimen.
The over-all length of the retained portion of the rattle was 4.3 cm.
It was composed of nine segments with an average length of 1 cm.
and a dorso-ventral width in the proximal lobe of 1.2 cm. The living
end-body obviously had reached adult dimensions before the latest
segments of the rattle were formed over its core. A comparison of
figures 58 and 59 clearly shows these differences between the segments
of young and old rattles.

Unquestionably the differences in size and shape of the successive
segments are due to differences in size and shape of the living end-
body. The younger the snake the smaller the end-body that con-

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 367

stitutes the mold or "matrix" for a segment forming over it. Details
of developmental processes involved will be described subsequently.

It is clear that differences between distal and proximal segments
of the same rattle are not due to a secondary reduction in size after
they have been formed. There is no evidence that drying or other
external influences were responsible for the smaller size of the seg-
ments that were established earlier. The button remains as it was
when first separated from its matrix. The gradual increase in size
of successive segments merely reflects the normal growth of the snake.

A degree of asymmetry has long been known to exist between
the dorsal and ventral halves of the segments. This asymmetry
keeps the end of the rattle from drooping, a condition that would
probably interfere with sound production.

In the complete, articulated rattle only the proximal lobe of each
segment and the original button are visible. Since one proximal lobe
is added to the growing rattle at each molt, the number of molts
can be ascertained by counting the proximal lobes. Adult snakes
rarely carry complete rattles and the number of segments, therefore,
indicates age little or no better than does size of the snake.

GROWTH CHANGES IN THE END-BODY OF A
LIVING TEXAS DIAMOND-BACK

(Observed by Means of X-ray Photographs)

To supplement our gross studies of the rattle and to obtain
confirmation of certain histological findings we used the x-ray
technique on a live snake. From March 25 to June 24, 1946, we
obtained thirty-two x-ray exposures from a specimen of Crotalus
atrox.

The snake measured forty-eight inches (exclusive of the rattle)
and weighed 4^ pounds. It was in excellent condition, having
grown to maturity in the Zoo of the Chicago Zoological Society at
Brookfield, Illinois, during the past 7% years. It was kept at rela-
tively constant temperatures (72° to 76° F.) and fed on freshly
killed rats. This valuable reptile passed through one molting cycle
while under our observation and was safely handled by the method
described on page 362.

The pertinent features of growth in the end-body are portrayed
in the positive prints from selected x-ray negatives (figs. 60-62).
The first print of this series shows the condition of the end-body

368 FIELDIANA: ZOOLOGY, VOLUME 32

and the style on March 25, 1946, four days after we took the animal
to our laboratory. At that time we failed to detect any outward
signs of an approaching molt or of new growth in the sulcus. The
x-ray negatives, however, furnished unmistakable evidence of such
phenomena.

This March 25 exposure shows the tail from the right side. The
faint outlines of the ten free segments (the rattle itself) are seen in
their normal relationships to each other and to the end-body. Since
they are previously established segments they are not further con-
sidered. The features of especial interest concern the end-body only.
It consists of three well-differentiated lobes whose connective tissue
stroma is outlined by a moderately dense x-ray shadow. In the
specimen the distal lobe extended 5.5 mm. beyond the end of the
style. This feature is the most significant sign of the approaching
molt. Soft tissues of the end-body begin to extend beyond the
style only at the termination of the long latent period. This exten-
sion characterizes, indeed, the relatively short pre-molting period
of accretion.

The basal projections of the style lie not in the proximal lobe
but deep to the caudalmost scales. This is the site of the widened
sulcus in which new growth takes place. It must be emphasized
that during this accretion period, as at all other times, two main
features of the anatomy of the tail remain fixed: (1) the plane of
articulation between style and caudalmost vertebra; (2) the number
of caudal vertebrae and the relative length of the muscular portion
of the tail. Growth in the sulcus, therefore, never encroaches upon
the tail itself but leads to the only other possible effect: a caudad
shift of soft tissues over the style.

The total length of the style, measured on its dorsal side, remained
constant throughout the period of observation. A new basal lobe
was formed in approximately fourteen days, a period that we
designate as the actual growth phase or period of accretion. The
first such period lasted from about March 20 to April 5 and the
second from June 12 to June 24. On April 13 and again on July 2
the molts were shed. During these periods of accretion the end-body
grew longer by the addition of a fourth distinct lobe in the sulcus.

Synchronously with this growth of the end-body there occurs the
caudad displacement or shift of the whole rattle. This is a natural
consequence of the pre-existing articulation of the last free segment
with the distal and middle lobes of the end-body. When these two
lobes shift caudad the whole rattle moves with them.

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 369

Up to the time of the actual molting there is no discrepancy
between the caudad shift of the end-body and that of the keratinized
epithelium covering it. In other words, there is no evidence in support
of Klauber's assumption of a "sort of wave action in the corrugations
of the material constituting the tissues of the lobes, relative to the
surface skin and the bones beneath." (Op. cit.) There merely occurs
a flow or shift of the end-body as a whole over the style.

Figure 60, c, illustrates the size and the relationship of the end-
body to the style near the climax of the period of accretion. A
comparison with figure 60, a, shows that in the earlier stage the
style extended well into the distal lobe of the end-body, whereas
in the later stage that lobe lay entirely beyond the style. The
maximal distance (6 mm.) between the tip of the style and the tip
of the end-body had been attained. This difference clearly is not
due to resorption of the tips of the style, which, as already stated,
retains a constant length, but to a caudad displacement of the
connective tissue of the end-body.

It is equally clear, however, that resorption of soft tissues in the
end-body itself must follow as a further sequence in the molting
cycle. Otherwise, the end-body would become larger with each
period of accretion. From the three-lobed condition the end-body
accrued to a four-lobed one between March 20 and April 5 and the
process was repeated between June 12 and June 24. If no reduction
took place the end-body would become five-lobed in the next growth
period, and the segments would become successively larger and more
complex. Instead, they remain constant in form and in their rela-
tionship to each other.

The resorption phenomena in the end-body are synchronous with
the molt and continue for a short time thereafter. Seventy-two
hours elapsed between the last exposure made prior to molting and
that obtained just afterward. The resorption of parts of the end-
body was complete at the end of that interval. In contrast to
accretion, which lasts for about fourteen days, resorption is relatively
short and occurs mainly during the actual process of molting.

The x-ray exposure shown in figure 61, b, was obtained soon after
the snake had molted. A comparison with the exposure made prior
to this (fig. 61, a) reveals striking differences in the end-body: From
a transient four-lobed condition it has changed to a three-lobed post-
molting one by the full resorption of the distal lobe. The previous
middle lobe thereby became the new distal lobe. Its connective
tissue core, likewise, was reduced and now lies immediately over the

370

FIELDIANA: ZOOLOGY, VOLUME 32

FIG. 60. Positive prints of x-ray photographs of the tail of a living adult
Texas diamond-back (Crotalus atrox). Dates of exposure: (a) March 25; (6) April
5; (c) April 8. These photographs portray the period of accretion. The base of
the style lies in the new basal lobe beneath the last scales. The distal lobe extends
from 5 to 6 mm. beyond the tip of the style.

distal tips of the style. This is the characteristic relationship of the
distal lobe to the style during the entire latent period, which follows.
In other words the previous second lobe has become the new distal
lobe in position and shape.

Resorption involves the connective tissue only, and it is evident
that the epithelial covering remains intact as the living stratum
profundum on the rapidly shrinking lobe. In contrast, the stratum

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 371

FIG. 61. Continuation of the series shown in figure 60. Dates of exposure:
(a) April 12; (6) April 15; (c) April 22. a, the climax of the accretion period ends;
b, the full resorption of the end-body, the stratum corneum remaining behind as
a new segment of the rattle; c, the relation of the end-body to the style at the onset
of the latent period. Molting occurred between the two stages shown in a and b.

corneum becomes detached and remains behind to constitute the
new horny segment of the rattle. It is a replica of the distal three
lobes of the transient four-lobed end-body.

There is a diminishing degree of resorption from the distal to
the proximal lobes of the end-body. Measurements of the dorso-
ventral diameters of the stroma in the various lobes were taken on
the x-ray negatives, which gave practically the natural dimensions

372 FIELDIANA: ZOOLOGY, VOLUME 32

in the living specimen. The differences in the pre- and post-molting
conditions are summarized in the following table:

DORSO-VENTRAL DIMENSIONS OF CORRESPONDING LOBES OF
END-BODY BEFORE AND AFTER RESORPTION

End-body after accretion End-body after resorption

(April 8) (April 15)

MM. MM.

11.5 Distal lobe .... Lobe resorbed

13.0 Middle lobe 10.5 New distal lobe

13.3 Basal lobe 13.8 New middle lobe

16.0 New basal lobe 15.0 New basal lobe

If the dorso-ventral diameters of the lobes in the pre-molting
condition and the same diameters of the newly formed horny segment
are compared they prove to be nearly the same. The newly detached
horny segment, therefore, reflects the shape and size of the distal
three lobes of the pre-molting end-body.

The resorption of the living soft tissues in the end-body is of
primary importance in the disengagement of the newly formed
segment of the rattle from its matrix, and in re-shaping the end-
body for the next growth period. There is definitely no flow or
longitudinal displacement of soft tissues within the epidermal cover-
ing and the newly forming horny segment. Instead, there is a with-
drawal or shrinkage after the soft tissues of the end-body with the
rattle have moved caudad. Before resorption takes place, the
stratum corneum of the newly forming horny segment, which is about
to become a functional part of the rattle, cannot be identified in
the x-ray negatives. After the resorption, the new horny segment
is clearly defined inasmuch as three lobes of the end-body have
sufficiently withdrawn to leave a clear space between the new horny
segment and the living tissues. That condition is illustrated in
figure 61, b. Before molting, the rattle consisted of eleven horny
segments and a four-lobed end-body; immediately after molting, of
twelve horny segments and a three-lobed end-body.

The size and morphological features of the end-body as newly
established during the active process of molting remained essentially
the same during the following two months (April 15- June 12, 1946).
This long phase of little or no change is the latent period. The
second molting cycle was then initiated by a new growth period.
This second period of accretion lasted from June 12 to June 24. At
the latter date a four-lobed end-body characteristic of the pre-
molting condition had again developed, and actual molting followed
on July 2.

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 373

FIG. 62. Completion of the series shown in figures 60 and 61. Dates of
exposure: (a) May 1; (6) May 14; (c) May 27; (d) June 24. Exposures a-c illus-
trate the relationship of the end-body to the style throughout the relatively long
latent period. In d the caudad shift of the soft tissues beyond the style and the
proximal growth of a new basal lobe have again occurred ; d represents a new period
of accretion. The second molt occurred on July 2.

374 FIELDIANA: ZOOLOGY, VOLUME 32

Figures 61, c, and 62, a-c, illustrate the condition of the end-
body and rattle during the latent period from April 19 to June 12.
Figure 62, d, shows the pre-molting condition as a result of the
second period of accretion. The anatomical features of the stage
represented by this figure are comparable to those in figure 60, b,
except that one segment had been added to the rattle. The interval
between these two comparable x-ray exposures was seventy-six days
and the addition of another segment to the rattle was imminent.
Based on the dates of shedding of these successive molts the cycle
was completed in eighty or eighty-one days. The table (p. 375)
summarizes the x-ray findings and correlates them with the observed
external conditions of the snake.

DEVELOPMENT OF THE SHAKER OR STYLE

The caudalmost vertebrae of rattlesnakes fuse to form the shaker
or style. This fusion takes place at different developmental stages
in different species. In Sistrurus catenatus, for instance, fusion occurs
postnatally, whereas in Crotalus atrox and C. adamanteus the con-
solidation of the last vertebrae is well advanced in late embryos.

The style was recognized in adult snakes as early as 1855 by
Leuckart, and described, as stated above, by Czermak (1857), who
named it the "end-body of the vertebral column." He compared it
with a simple exostosis in the form of "a single, conical, bilaterally
flattened piece with two rounded, more or less separate points" and
believed it to be composed of seven or eight vertebral elements.
He noticed that the vertebral canal continues far into this bony
structure and that it is provided with vestiges of intervertebral
foramina.

Garman (1888) stated that the caudal vertebral column in
embryos of rattlesnakes (specifically Sistrurus catenatus) is "similar
to that of other ophidians at the same stage : each vertebra is distinct,
movable." But he believed that later additions occur whereby the
end of the vertebral column becomes markedly changed. Shortly
after birth "the hinder seven or eight of the vertebrae are seen to
have coalesced into a single mass, showing a disposition to expand
so as to obliterate the processes and lines of demarcation of the bones
of which it is composed, but which are still plainly indicated. This
composite bone may be called the shaker."

Garman further stated that because of the expansion of the style
and the thickening of the skin "the muscles between the bone and

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 375

RECORD OF MOLTING CYCLE OF CAPTIVE TEXAS DIAMOND-BACK

Date Stage of cycle

March 20. Growth phase or period
of accretion already begun

March 25. Period of accretion con-
tinues

March 30. Period of accretion con-
tinues

April 1. Period of accretion con-
tinues

April 3. Period of accretion con-
tinues

April 5. Period of accretion con-
tinues

April 8. Period of accretion
reaches climax

April 10. Period of accretion still
at climax

April 12. Period of accretion ended,
period of resorption be-
ginning

April 13. Period of resorption con-
tinues

April 15. Period of resorption end-
ed, latent period begun

April 19- Latent period continues
June 12

June 24. Period of accretion well
advanced

July 2. Period of resorption un-
der way

Remarks

No exposure yet made; new lobe probably
began to develop at about this time. Snake
is kept at constant temperature (76° F.).

First exposure shows new basal lobe well
under way, but external examination reveals
no change; cornea clear. See fig. 60, a.

Cornea clear.

Exposure and external examination show
new basal lobe well developed; cornea
cloudy.

Exposure and external examination show
new basal lobe approaching maximum de-
velopment; cornea very cloudy.

Exposure and external examination show
new basal lobe at about maximum develop-
ment; cornea very cloudy. See fig. 60, b.

Exposure and external examination show
little or no discernible change in new basal
lobe; cornea has cleared. See fig. 60, c.

Exposure and external examination reveal
no further change; cornea clear.

Looseness of molt shows that ecdysis is
imminent; cornea clear. See fig. 61, a.

Molt found in cage; segment being added
to rattle; cornea clear.

Segment, now a functional part of rattle,
contains only remnants of the rapidly dis-
appearing stratum intermedium; cornea
clear. See fig. 61, 6.

Exposures made on and between these
dates were identical with that of April 15
except that remnants of the stratum inter-
medium were no longer evident; cornea
clear. See figs. 61, c, and 62, a-c.

New basal lobe had developed to a phase
comparable to that shown in exposure of
April 1, indicating completion of the 83-day
cycle; cornea slightly cloudy.

Molt occurs, completing cycle of 80 days
based on April 12 ecdysis.

376

FIELDIANA: ZOOLOGY, VOLUME 32

the skin are on the way to disappear." Our own observations on
the youngest postnatal stages do not confirm this view. There is
no evidence that muscle tissues ever extend farther caudad than
the base of the developing style.

Garman's figures obviously are diagrammatic and unconvincing.
We doubt that his material and technique were adequate for the
study of early stages.

FIG. 63. Camera lucida tracings of
three alizarine-stained and cleared tails of
young massasaugas (Sistrurus c. catenatus):
(a) of first, (6) of fourteenth, (c) of thirtieth
day after birth. In the younger stages the
vertebral elements of the developing style
are still discrete; in the oldest specimen
fusion has begun and extravertebral ele-
ments are being added to the style.

srnm.

We have studied the development of the style in numerous
specimens of Sistrurus catenatus and of the Crotalus atrox-adamanteus
group, using the method of selective staining with alizarine followed
by clearing. The material varied in age and size from tails of late
embryos to those of fully grown snakes.

Our observations on the development of the style were most
instructive in Sistrurus catenatus because in that species coalescence
of terminal vertebrae takes place postnatally and our material was
ample. A brood was born at the Museum, and specimens were
killed at intervals: one on the first, one on the fourteenth and a
third on the thirtieth day after birth. These young snakes were
218 mm., 211 mm. and 246 mm. long, respectively, and their buttons
measured from 3.7 to 3.9 mm. in length. The average length of the
fully formed button in this species being 5 mm., it is evident that
the end-bodies of these young specimens had not yet attained their
full size. Neither had a stratum corneum become detached from the
young epidermis. Further growth, therefore, takes place in the end-

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 377

body between birth and the time of detachment of the keratinized
button.

Figure 63, a, shows that on the first postnatal day separate
vertebral elements are present in the axis of the end-body. Nine,
possibly ten such units can be counted. Their contours are not
sharply defined since they are in transition from the blastemal to
cartilaginous and early bony stages. The caudalmost element is

FIG. 64. Camera lucida tracings
of the alizarine-stained and cleared
tails of (a) the advanced embryo of a
Texas diamond-back (Crotalus atrox),
(6) the advanced embryo of a Florida
diamond-back (C. adamanteus). The
original intervertebral foramina are
recognizable, and longitudinal trabe-
culae are developing on the ventral
side of the style.

very small and the entire group tapers. In our preparation the
caudal artery was seen to extend into the end-body.

During the first fourteen days of postnatal life little progress in
the establishment of a shaker or style is evident (fig. 63, 6). There
is no marked degree of fusion among the caudalmost vertebrae, but
osteogenic foci appear in the connective tissue surrounding the very
tip of the vertebral column. These secondary additions produce
nodular, irregular trabeculae, which change the previously tapering
end of the vertebral column to a broader and somewhat fan-shaped
mass.

The results of this change are more pronounced in the specimen
one month old (fig. 63, c), but vertebrae within the end-body are
still discrete. Fusion has begun distally between the caudalmost
vertebral units and the secondary bony contributions from the
surrounding connective tissue. Further osseous nodules are develop-
ing even beyond this fused portion of the young style.

The other material at our disposal for this part of the study
included embryos and early postnatal stages of Crotalus atrox and

378

FIELDIANA: ZOOLOGY, VOLUME 32

C. adamanteus. We stained and cleared the tail ends of two embryos
of Crotalus atrox (260 and 265 mm. in total length) and of one late
embryo of C. adamanteus that was obtained from within its amniotic
sac (total length 380 mm.).

The formation of the style in these late embryos of Crotalus was
more advanced than that in postnatal Sistrurus, a considerable

FIG. 65. Camera lucida tracings of the
alizarine-stained and cleared tails of three
Texas diamond-backs (Crotalus atrox) a few
months old. Further consolidation of the
style has occurred by the fusion of the caudal-
most vertebrae and by the addition of extra-
vertebral bony trabeculae at the ventral side
and tip.

fusion of eight to ten terminal vertebrae having taken place within
the end-body (fig. 64, a and 6). The original intervertebral foramina
are easily recognizable in the illustration and betray the segmental
portion of the style.

There are indications of extravertebral additions by osteogenic
centers near the tip. Furthermore, a pair of longitudinal trabeculae
develops on the ventral proximal side of the young style. Numerous
fine bloodvessels radiate from the style into the connective tissue
of the end-body. They are evidently the terminal branches of the
caudal artery.

The next older stage studied was represented by postnatal speci-
mens of Crotalus atrox. They were killed in October when a few
months old. Their average total length was 333 mm. Three
specimens of the twenty studied in this group are illustrated above
(fig. 65).

The end-bodies at this stage are definitely covered by a stratum
corneum that is detached or ready to be detached as the button.
Its length ranged from 7.1 mm. to 8.0 mm. in the cleared preparations
illustrated. The style has further developed by fusion of previously

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 379

discrete vertebral and extravertebral elements. The end -prongs
are forming from distalmost bony trabeculae arising in connective
tissue and the pair of ventral longitudinal bars gradually extend
distad. When well developed they look like runners on a sled. They
strengthen the ventral margin of the style. There are minor varia-

FIG. 66. Camera lucida tracings of the alizarine-stained and cleared tails of
two young Texas diamond-backs (Crotalus atrox). The segmental origin of the
proximal portion of the style is still recognizable. The extravertebral part of
the style extends into the distal lobe.

tions in the degree of consolidation of the various elements of the
style in specimens of about the same age. But the general form and
the structural characteristics of the style are attained within a few
months after birth.

Figure 66 shows the tails of two older specimens of Crotalus atrox,
455 mm. and 550 mm. in total length. The end-body of the smaller
snake was 9.4 mm. long and 7.6 mm. wide at the basal lobe; that of
the larger 10.4 mm. long and 10 mm. wide. In both specimens the
segmental origin of the proximal portion of the style is indicated by
successively overlapping bony ridges and trabeculae betraying the

380

FIELDIANA: ZOOLOGY, VOLUME 32

forms of various bodies and apophyses of former vertebrae; distally
no such segmental arrangement is discernible.

The proximal portion of the style extends into the middle lobe.
This is consistent with our observation on other material that the
spinal cord terminates within this lobe. The portion of the style
derived by extra vertebral ossification, therefore, occupies the re-
mainder of the middle lobe and extends somewhat into the distal
one.

The styles of fully grown specimens shown in figures 67 and 68
belonged to snakes (Crotalus atrox) 835 mm. and 1,000 mm. in

FIG. 67. Camera lucida tracing of the style and the caudalmost vertebrae of
an adult Texas diamond-back (Crotalus atrox). The highly trabeculated structure
of the style is evident.

total length. The long axes of their end-bodies measured 12 mm.
and 16 mm., respectively. The distinction between the originally
segmental and the non-segmental distal portions is less apparent in
old than in young snakes. Yet even in the largest specimens the
style is not a compact bony end-piece of the vertebral column.
Structurally, young and old styles are highly trabeculated, with
fairly smooth ventral and dorsal margins. Between the bony
trabeculae there are relatively large marrow spaces. From the style
emerge eight or nine segmentally arranged nerves, which course into
the soft tissues of the end-body. These nerves evidently are sensory
as there is no muscle within the end-body needing motor supply and
small ganglia are recognizable at their roots, just outside the vestigial
intervertebral foramina of the style. The dorsal and ventral muscular
masses of the tail are inserted on the corresponding proximal prongs
of the style. The rapid vibrations caused by muscular action are

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 381

secondarily transmitted to the rattle, the proximal segment of which
caps the soft tissues of the end-body.

DEVELOPMENT OF THE RATTLE IN YOUNG
OF THE FLORIDA DIAMOND-BACK

The histological study of the development of the first segment
with three lobes was a main object of the study. Previous workers,
failing to consider such details, merely speculated with the help of
theories and mechanistic schemes.

We used nine selected specimens of Crotalus adamanteus that
had been preserved shortly after birth. One specimen of this series

FIG. 68. Camera lucida tracing of the style and the caudalmost vertebrae
of an old Texas diamond-back (Crotalus atrox).

had a rattle consisting of a button and two segments, the youngest
of these segments having just been established. The other specimens
had only a button with a first segment in various stages of develop-
ment. Six of these specimens were studied histologically after the
tail ends had been embedded in celloidin and sectioned with the
microtome.

Morphological Features

The average length of the severed tail ends was 15 mm. and that
of their buttons was 7.4 mm. The average dorso- ventral diameter
at the proximal lobe of the button was 6 mm.

Figure 69, A-D, shows lateral views of four specimens of this
series. The original drawings were made with hand lens and camera

382 FIELDIANA: ZOOLOGY, VOLUME 32

lucida, before decalcification and embedding. The broad lateral
margins of the caudal plates indicate the ventral sides. Each button
has two lobes, which are separated by a transverse groove, and there
is a narrow longitudinal furrow on each side of the button. The
characteristic asymmetry of the button is readily seen, the ventral
side of the proximal lobe extending consistently farther caudad than
the dorsal side. This indicates either an earlier differentiation or a
faster rate of growth of the ventral side of that lobe.

In figure 69 the specimens are arranged in the order of develop-
ment, A representing the earliest and D the latest stage in the
development of the first segment. Outwardly, no marked features
indicate such development, but close examination reveals two im-
portant steps in the development of the new segment of the rattle,
one of resorption and one of accretion.

The feature of resorption is visible through the translucent
button. Even though the detached stratum corneum constituting
the button is heavily pigmented in Crotalus adamanteus, a shrinkage
of the inner core or end-body is easily detected, particularly with
the help of the binocular dissecting microscope. The progressive
degree of that shrinkage of the underlying end-body away from the
horny button is indicated in the figures by a broken line representing
the contour of the end-body.

The horny button becomes detached from the underlying epi-
dermis that gave it origin as the keratinized stratum corneum.
There is no expansion or change in form of the button after its
detachment; its invariable measurements in successive stages make
this fact evident.

Measurements were also taken of the distance between the
caudalmost tips of the button and of the end-body within. The
distances between these points were 0.3 mm., 1.4 mm., 1.2 mm.,
and 1.7 mm. for the successive stages portrayed (fig. 69).

In the latest stage shown the reduction of the end-body within
the button appears to be almost one-fourth of the length of the button
(1.7 mm. within the button of a total length of 7.2 mm.).

However, the processes involved in this lack of conformity be-
tween the button and the caudalmost extension of the end-body are
not clearly shown by our material. The obvious reduction is inde-
pendent of the growth proximal to the button. This growth estab-
lishes a new basal lobe and causes a slow caudad dislocation of the
end-body over the bony style.

383

384 FIELDIANA: ZOOLOGY, VOLUME 32

This feature of growth or accretion is detectable in figure 69,
B-D, immediately proximal to the button. Beneath the caudalmost
scales there arises a swelling in the region of the epidermal sulcus
between the button and the last scales. By the relatively rapid
expansion of the newly forming lobe the caudalmost scales them-
selves become separated from each other, their displacement being
in a longitudinal as well as in a transverse direction. Their inter-
squamal folds become erased or reduced by stretching.

In figure 69, A, the scales are seen to be closely imbricated,
whereas in D the scales of the last two rows are spread apart and the
intersquamal epidermis is visible. This spreading makes it possible
to ascertain whether a rattlesnake of any age is in the pre-molting
stage when growth of a new basal lobe occurs. We used this sign
of growth in checking over numerous preserved specimens in the
study collection of Chicago Natural History Museum. The sign,
always betraying proximal accretion of the rattle, is an indication
of imminent molting as valid as is the clouding of the cornea.

The stretched epidermis and the unimbricated scales, just proxi-
mal to the button, involve only that layer of the skin that forms
the molt. The caudal end of the molt is always attached to the
button, or, subsequently, to the horny segment about to be freed.
The area of attachment lies in the first transverse groove of the new
but still functionless basal segment. Consequently, the molt tem-
porarily hides the developing proximal bulge.

A comparison of figure 69, D, and figure 70 shows that similar
conditions prevail in very young and in mature specimens of different
species. Figure 70 was drawn from a specimen of the prairie rattle-
snake, Crotalus viridis viridis. It shows the tail end of a grown
snake in the pre-molting stage. The functional horny segments of
the rattle have been removed and the proximal segment has been
cut open lengthwise. One-half of this segment is covered by the
attached molt, whereas the other half has been freed of its molt in
order to expose the new basal lobe and the closely imbricated caudal-
most scales.

Further features are illustrated by figure 71, A-D, which was
drawn from longitudinal sections, made previous to embedding, of
the same specimens shown in figure 69, A-D. The figures with
identical letters portray the same specimens (only in the case of
D are they viewed from the same side). They were drawn with
identical magnifications. Three of the sections are slightly para-
sagittal but B and D show midsagittal sections. The extension of

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 385

the spinal cord into the end-body is clearly seen; it ends in the
distal lobe within the fused vertebral elements of the style.

In these hand-made sections the connective tissue of the end-
body had the characteristic white appearance with a bluish tinge
caused, mainly, by the interspersed pigment cells. As the sections
were unstained, the detail features and contours of the various tissues
are merely suggested.

FIG. 70. The tail of an adult prairie rattlesnake (Crotalus v. viridis) in pre-
molting stage. The functional segments of the rattle have been removed and the
end-body has been bisected longitudinally. The molt still covers the ventral half
of the newly formed basal lobe. Compare the widely separated scales of the
molt with the closely imbricated scales of the half from which the molt has been
removed.

The significant features in these figures are the sectional views
of the developing proximal lobe in the first rattle segment. The
region of special interest in each section is indicated by an arrow.
In A it merely marks the epidermal groove between the last scales
and the button. In B the formerly narrow sulcus had widened but
is still partly covered by the caudal plates and scales. This widened
groove indicates earliest growth in a newly forming lobe. In C a
young basal lobe is present in the identical location of the previously
widened groove, immediately proximal to the base of the button.
The new lobe is still narrow and covered by the future molt. In D,
finally, the newly added basal lobe of the first segment is well formed,
though still covered by the molt.

In all figures, the ventral and dorso-lateral muscular masses may
be seen extending caudad as far as the sulcus. This caudal muscula-
ture does not extend into the end-body proper at any stage of

386 FIELDIANA: ZOOLOGY, VOLUME 32

development. The several muscular bundles insert upon the base
of the style, which lies in the same transverse plane as the base of
the proximal lobe.

In D the space between the button and the end-body is shown
to be occupied by a peculiarly spongy tissue. It has signs of degenera-
tion, such as vacuolization and formation of irregular strands. This
tissue, derived from the stratum intermedium between two succeed-
ing generations of the stratum corneum, is destined ultimately to
disappear. Further details of this process are considered in the
following paragraphs.

Histological Features

The histological studies of growth in the developing rattle dealt
with the following main aspects of the problem:

1) Histology of the end-body: the connective tissue core; the
epithelium and its relation to the detached button; the process
of keratinization; the coalesced terminal segments of the
vertebral column (style) ; the extension of the spinal cord into
the end-body.

2) Phenomena of degeneration and of resorption: degeneration of
the stratum intermedium between successive generations of
the stratum corneum; bone resorption in the style; changes
in the connective tissue of the end-body.

3) Phenomena of accretion: evidences of new growth in the sulcus
between the last scales and the end-body; the formation of a
new proximal lobe.

Histology of the End-Body

A general orientation of the young end-body and its relation to
the button are shown (fig. 72, A and B) in the sections of a Florida
diamond-back (Crotalus adamanteus) not more than a few months
old. The original drawings were made as camera lucida tracings
from horizontal longitudinal sections, at a 20-fold magnification.
Figure 72, A, shows a section through the spinal cord. It is dorsal
to that shown in figure 72, B, which passes through the caudalmost
vertebrae and the style. The procoelous vertebral bodies and their
articulations can be readily identified (fig. 72, B). The ganglia of
the caudalmost spinal nerves are also shown.

The major portion of the connective tissue of the end-body
consists of densely interlaced collagenous fibers. They form a tough

^yb^aaMAhittfeft W

Wfr

387

388 FIELDIANA: ZOOLOGY, VOLUME 32

felt-work, intermingled with numerous pigment cells. This connec-
tive tissue bed is well provided with bloodvessels that emerge from
the style as the last branches of the caudal artery and vein. The
nerve supply in the connective tissue of the end-body is also rich.
The terminal branches of nerves emerge from between the fused
segments of the style. From seven to eight segmentally arranged
pairs were identified within the end-body proper. Their distribution
appears limited primarily to the connective tissue and skin of the
proximal lobe; only traces of nerve fibers were identified in the distal
lobe. However, no silver preparations were made. The difference
in the number and size of nerves supplying the two lobes of the end-
body, nevertheless, is striking. The relative paucity of nerves in
the distal lobe may be related to the resorption phenomena that this
lobe suffers during the development of the rattle.

The density of the connective tissue in the end-body is not
homogeneous. It is more uniform in the distal lobe where the col-
lagenous fibers are evenly interlaced. A distinction between the
derma of the skin and the underlying connective tissue of this lobe
cannot be made, whereas a definite dermic zone is easily identified
in the proximal lobe. Here the subcutaneous connective tissue shows
distinct zones. Surrounding the style there is a layer of coarse
collagenous fibers that form an open-meshed network with inter-
fibrous spaces. Farther peripherally, this tissue is connected to the
derma by a much finer connective tissue zone with greater homo-
geneity; coarse fiber bundles are lacking and there are no conspicuous
interstitial spaces. The coarse open-meshed zone of connective
tissue in the proximal lobe gives the impression of a fluid cushion
around the style. This arrangement may be related to rapid growth
changes that occur in the nearby sulcus and produce a shift of soft
tissues in the end-body over the stationary style. Further considera-
tion of this caudal displacement is given in the account of our x-ray
studies.

Pigment cells are numerous throughout the connective tissue
of the end-body. They are star-shaped, with numerous long processes
containing brown pigment granules. They are most numerous in the
derma directly beneath the basal layer of the epidermis, often lying
adjacent to the adventitia of small bloodvessels. In many instances
pigment cells were observed in the process of migrating from the
derma to the epidermis. Such migrated pigment cells are located
between the cells of the deeper layers of the squamous stratified
epithelium. Furthermore, both the detached button and the new

B

FIG. 72. Horizontal longitudinal sections of the tail of a young Florida diamond-
back (Crotalus adamanteus) : A, the termination of the spinal cord within the end-
body; B (ventral to A), the caudalmost vertebrae, style, and termination of lateral
muscle masses. At the tip of the end-body, between the button and the newly
forming stratum corneum, lies the degenerating stratum intermedium. Arrows
indicate the sulcus; the asterisk (*) refers to the details shown in figure 73.

389

390 FIELDIANA: ZOOLOGY, VOLUME 32

stratum corneum adjacent to the end-body contain characteristic
pigment deposits that in their mottled arrangement betray an origin
from infiltrated chromatophores.

The epidermis investing the end-body consists of squamous
stratified epithelium. Unfortunately, there is little uniformity in
the histological nomenclature and in the interpretation of the
various layers of the skin in Squamata. The same layers have been
variously designated, some discrepancies being due, undoubtedly,
to differences between genera or species.

The following description is based primarily on observations of
the epidermis in the young specimens of Crotalus adamanteus already
referred to. The detached button consists exclusively of a stratum
corneum (horny layer) and constitutes the first unit of the rattle
(figs. 69, 71, and 72). It is fully keratinized with variable amounts
of pigment granules deposited within the horny substance. The
latter has a natural orange-yellow color and is refractory to acid
or basic stains (eosin and hematoxylin). With the exception of the
pigment granules in it, the button is virtually an anucleated homo-
geneous layer. Traces of degenerated nuclei of once living cells are
found infrequently. In the thickened and reverted proximal margin
of the button there usually are a number of eosinophile granulocytes
that migrated into the outer stratum corneum before its detachment,
and died there.

The process of keratinization is of special interest. In the
epidermis of the end-body one recognizes a deeply staining basal
layer (stratum basale) of columnar or cuboidal cells. It is the
germinative layer (stratum germinativum) from which all other
epithelial cells are derived. In young specimens of Crotalus the
basal layer of the end-body is overlain by six to ten layers of polygonal
or slightly squamous cells, which may be referred to as a stratum
spinosum. Together, the stratum basale and the stratum spinosum
constitute the stratum profundum. There is no trace of keratiniza-
tion in this layer. In figure 72, A and B, however, a new generation
of a stratum corneum covers the very tip of the distal lobe of the
end-body. No such layer is present over the proximal lobe or over
proximal portions of the distal lobe. Therefore, the process of
keratinization does not occur simultaneously in the outer layer of
the entire epidermis covering the end-body but progresses from its
tip to the sulcus. There exists a narrow zone of epidermis in which
an intensely staining stratum granulosum can be identified (fig. 73).
In the specimen portrayed it lies around the middle of the distal

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 391

lobe and shows an abrupt transition into the yellow, fully keratinized
stratum corneum of the new generation. The deeply staining
granules of keratohyalin, characteristic of the stratum granulosum,
apparently are transformed into eleidin abruptly and in this narrow
zone only, which gradually extends proximad over the end-body.
Evidently there is a gradient in the progress of keratinization. The
transition between keratohyalin and eleidin strikingly resembles that
in the nail bed of a human fetal finger. Our selected material supplies
no evidence for the rate of progress of the keratinization from the
tip to the base of a newly forming horny segment.

Keratinization begins over the distal tip of the future three-lobed
end-body before a new basal lobe becomes established. During the
development of later segments, keratinization of the new stratum
corneum occurs in the final phase of the resorptive process, which
immediately follows formation of a new basal lobe.

The difference in these events becomes clear when one considers
that within the two-lobed button the distal lobe of the end-body is
not being resorbed to a considerable degree. If it were, there could
not result a three-lobed end-body by the addition of a new basal
lobe. The end-body within the button, therefore, remains essentially
two-lobed and keratinization of its new stratum corneum can begin
before new growth is under way in the terminal sulcus.

An additional feature is characteristic of this zone of keratiniza-
tion. Only a narrow cell layer immediately above the stratum
profundum differentiates into a stratum granulosum. The cell
layers above the latter quickly acquire entirely different charac-
teristics. They fail to accumulate keratohyalin granules but rapidly
show pronounced vacuolization. These outer layers in the thickened
band of cell proliferation become the stratum intermedium. It is
the outermost of the new generation of epidermal layers and provides
the zone of detachment of the old stratum corneum from the living
epidermis. Distal to the band of keratinization the stratum inter-
medium suffers such extensive vacuolization that its original epithelial
nature is virtually erased.

Here then, is an outer zone in the epidermal epithelium that
shows no trace of keratinization. It becomes detached from the
underlying stratum granulosum. In snakes with only a button it
expands into a spongy mass of spherical cells. This peculiar tissue
approximately fills the space between the old horny layer and the
newly forming one. In distal, older portions of the vacuolated
stratum intermedium many cell boundaries are lost. Through

392

FIELDIANA: ZOOLOGY, VOLUME 32

coalescence of cell spaces larger "vacuoles" are formed (see fig. 72, B).
Further degeneration of the stratum intermedium reduces it to mere
strands of former cell walls with still larger spaces between them.
Pycnotic nuclei are numerous in the older portions. This tissue,
shown macroscopically (fig. 71, D), is little more than a temporary

st. p. st. c.

st. i.

FIG. 73. Photomicrograph
of the epidermis and the derma
of a young Florida diamond-
back (Crotalus adamanteus) in
the area marked by an asterisk
(*) in figure 72, A.

This illustration shows
the transition from stratum
granulosum (st. g.) to stratum
corneum (st. c.) of the newly
forming horny segment. The
stratum intermedium (si. i.)
contains numerous granulo-
cytes; st. p. is the stratum pro-
fundum of the epidermis.

'

derma

st. g. st. i.

epidermis

"filler" between the horny button and the new generation of a stra-
tum corneum covering the end-body. In preparations obtained by
standard histological techniques the vacuoles of such a tissue appear
empty and the tissue looks much like a foam. It is more likely,
however, that the vacuoles were fluid-filled before dehydration with
alcohols. Probably this is the jelly-like tissue noticed by Klauber.
Eventually it disappears, presumably by desiccation.

In the process of degeneration the stratum intermedium suffers
an invasion of numerous leucocytes. These cells are eosinophile
granulocytes, the white cells in the blood of reptiles. They invade

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 393

the connective tissue of the end-body directly from their site of
origin in the bone marrow of the style. These round or ovoid cells
are two or three times the size of nucleated erythrocytes in the same
species. Their nuclei usually are eccentric and the cytoplasm con-
tains refractile eosinophile granules.

Figure 73 is a photomicrograph of the dermal and epidermal
layers in the zone of keratinization. A number of eosinophile granu-
locytes are seen in the stratum intermedium of the epidermis. Un-
questionably, they are functionally related to the rapid degeneration
of that layer. Their possible role will be referred to below.

The caudalmost extension of the vertebral column into the end-
body is clearly illustrated (fig. 72, A and B). The last articulated
vertebra reaches a plane passing through the sulcus. Distal to that
plane the successive elements of the caudal vertebral column are
fused to form the style. The segmental origin of the style is still
discernible through paired lateral projections and through the
segmental arrangement of the caudalmost nerves that appear from
within the style. In young specimens of Crotalus the style extends
nearly to the tip of the end-body. Its caudalmost portion consists
mainly of longitudinal trabeculae.

The style in these young specimens is not a dense bony mass, but,
as already indicated, contains marrow spaces. Connective tissue
lies between the delicate spicules or trabeculae of the distal part of
the style; ossification begins proximally. The development of the
style was described in a preceding section.

The caudalmost extension of the spinal cord (figs. 71, B-D, and
72, A) reaches into the distal lobe of the end-body, lying within the
larger portion of the style. The meningeal investments, particularly
a layer of connective tissue that corresponds to the mammalian
dura mater, extend farther caudad than the spinal cord. A con-
spicuous fibrous strand in the nature of a dural filum reaches the
tip of the end-body. This midsagittal region of the end-body shows
the interesting histological features of osteoclastic activity considered
below.

Phenomena of Degeneration and of Resorption

The earliest signs of tissue degeneration and of resorption appear
in the end-body of the specimens of Crotalus adamanteus not more
than a few months old.

The degenerative changes in the epidermis concern the stratum
intermedium, which comes to lie between the old and new generations

394 FIELD IANA: ZOOLOGY, VOLUME 32

of the stratum corneum. The invasion of this layer by a great
number of eosinophile granulocytes was previously indicated (fig. 73).
The presence of such conspicuous cells of non-epithelial origin within
an obviously degenerating layer of the epidermis deserves further
consideration. The relatively large leucocytes apparently remain
alive for a considerable time after infiltrating this epidermal layer.
The fact that their cytoplasm and nuclei retain their stainability
even in the caudalmost portions of the vacuolated stratum inter-
medium is good evidence of such viability.

What is the function of these cells in the process of ecdysis?
This study cannot give direct proof for a specific function but it
supplies circumstantial evidence that eosinophile granulocytes are
concerned with degenerative processes within the stratum inter-
medium. The functions of eosinophile granulocytes in reptiles
might well be the same as in man and mammals. Eosinophile leuco-
cytes in the stratum intermedium, therefore, may be concerned with
the elaboration of proteolytic enzymes (Naegeli, 1923). Hence, we
believe that the jelly-like vacuolated nature of the degenerating
stratum intermedium is the result of protein-dissolving properties
of the invading eosinophile leucocytes.

The distribution of numerous eosinophile granulocytes is by no
means limited to the stratum intermedium over the distal lobe of
the end-body. These leucocytes are equally numerous and presum-
ably equally important in the delamination processes that occur
in the epithelium of the sulcus at the time of formation of a new
proximal lobe. They are likewise encountered in the derma, in the
epidermis, and even in the detaching molt at the caudalmost inter-
squamal folds.

It would appear, therefore, that the dehiscence between two
layers of the stratified squamous epithelium does not occur simply
by a progressive hornification of the outermost cell-layers or by
simple physical factors such as drying and subsequent delamination.
The timing in the events leading to the development of such a rela-
tively intricate structure as the keratinized rattle evidently is pre-
pared and controlled by special histological events. The appearance
of large numbers of eosinophile granulocytes in zones of intra-
epithelial delamination undoubtedly constitutes one such controlling
factor.

An explanation seems necessary for the spatial relationship
between the horny button and the epithelium over the proximal lobe
of the end-body, as portrayed (fig. 72, A and B). It was pointed

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 395

out above that no stratum intermedium exists between these two
layers in the region concerned. Yet in the figures the stratum
corneum of the button is detached from the proximal lobe. We
believe this to be an artifact, due probably to some degree of shrink-
age of the end-body.

In vertebrates, the eosinophile granulocytes are derived from
the myeloid tissue in the marrow spaces of the skeleton. The marrow
of the style and of the articulated vertebrae contains, indeed, large
numbers of eosinophile myelocytes. In reptiles, as in birds, granu-
lopoiesis and erythropoiesis are distinct from each other within the

FIG. 74. Eosinophile myelocytes of the
bone marrow in the style and vertebrae of
a young Florida diamond-back (Crotalus
adamanteus).

myeloid tissue. Granulopoiesis, that is, the development of granular
leucocytes from myelocytes, occurs outside the bloodvessels in the
marrow, whereas the erythrocytes arise by an intravascular process
of cell proliferation. Figure 74 illustrates the characteristic type of
eosinophile myelocyte that constitutes the ancestral cell-forms of
eosinophile granulocytes. The cells portrayed are ovoid or polygonal
and have a slightly basophile cytoplasm. The nucleus usually is
excentrically located; it stains a dull blue and has no nucleolus but
contains a few indistinct chromatin granules. It is larger but less
dense than the nucleus of erythrocytes. The cytoplasm contains
relatively coarse eosinophile granules that usually are accumulated
in a limited area of the cell. In the fully developed and migrated
eosinophile leucocyte, these granules occupy the entire cytoplasm.

The diameter of the eosinophile myelocytes varies from 10 to 12
microns. Because of their round or polygonal shape they appear
considerably larger than the highly ellipsoid erythrocytes, although
the latter's longitudinal diameter measures about 10 or 11 microns.

Bone resorption occurs within the tip of the end-body distal to
the termination of the spinal cord. High power observation of the
mid-axis of connective tissue revealed the presence of osteoclasts.
The length of the distal lobe of the end-body shown in figure 72

396 FIELDIANA: ZOOLOGY, VOLUME 32

measured 4 mm., the caudalmost tip devoid of trabeculae only 2 mm.
Within this terminal zone, bone resorption extends through 1,200
microns. Here lie the meningeal filum and the last branches of the
caudal artery and vein. A high power field of this area is shown in
figure 75, A, originally drawn by camera lucida at a magnification
of 600 times. A number of multinucleated giant cells, readily
identified as osteoclasts, lie in a group, interspersed in the axial
connective tissue between the caudalmost bloodvessels. In some
sections as many as twenty-two osteoclasts were counted within a
length of 1.2 mm. of caudal tissue.

The shape and size of these giant cells vary greatly, in part be-
cause in any section mere fragments of the voluminous cell-bodies
may be encountered. Some of the cells measure 40-50 by 20 microns.
A whole cell or a fragment naturally shows a variable number of
nuclei, but from eight to ten nuclei were frequently observed. The
nuclei, characteristically, appear empty, vesicular, with a single
nucleolus. The cytoplasm of these osteoclasts is non-granular and
stains slightly purple with ordinary H and E stain. The c£lls are
stellate, their margins provided with processes that may be of
considerable length. These osteoclasts are strikingly like those of
mammals and man. Some osteoclasts are shown at a larger scale
(fig. 75, B) and in association with an isolated bony trabecula.

It is evident that a degree of bone resorption occurs at the distal
end of the young style and is part of the final shaping that involves
loss of its attenuated tip. This process of regression precedes the
development of a new proximal lobe. No comparable bone resorption
takes place during the formation of subsequent segments. Further-
more, the sequence of events here is the reverse of that observed in
x-ray exposures of the growing adult rattle. There, resorption of the
distal lobe follows accretion, as a temporary four-lobed end-body
becomes reduced to a three-lobed structure. In contrast, the main
event in the establishment of the very first three-lobed segment
consists in the addition of a third lobe to the slightly reduced two-
lobed end-body within the button.

It is evident (see fig. 72, A and B) that the end-body became
reduced in size after the button had been formed over it. Even
if a small degree of shrinkage is accepted as an artifact there remains
a considerable space between the distal lobe and the tip of the button.
The distal lobe previously was the direct mold for the button. There-
fore, some regression must have taken place. This may have been
due partly to a resorption of tissue fluids, which might explain the

.' sift
IP

'/''

.Ql mm.

FIG. 75. Osteoclasts (O) or multinucleated giant cells within the end-body
of a Florida diamond-back (CrotaZus adamanteus) : A, in the long axis of the distal
lobe, just beyond the terminal tip of the style; B, a single osteoclast (O) and several
granulocytes (G) adjacent to the bony trabecula of the style; C, three osteoclasts
(O) and, for comparison, three erythrocytes (E) situated near a resorbed bone
spicule.

397

398

FIELDIANA: ZOOLOGY, VOLUME 32

denser aspect of the connective tissue of the distal lobe as compared
with that in the proximal one.

Reduction of the underlying connective tissue in all probability
precedes bone resorption and is the earliest sign in the chain of
events that lead to the reduction in size of the living core within
the button. The essential phenomena are: (1) reduction of the con-

FIG. 76. Histological details in the sulcus of the section shown in figure
72, A. The thickened proximal margin of the button overlies the sulcus. The
arrow marks the beginning of an epithelial invagination, earliest sign of a newly
developing basal lobe.

nective tissue in the distal lobe beneath the keratinized button;
(2) bone resorption in the tip of the style; (3) degeneration of the
stratum intermedium between the old and the new generations of
the stratum corneum.

Phenomena of Accretion

We have already shown that the first three-lobed end-body is
formed by the addition of a new proximal lobe, and that growth
succeeds resorption in the distal portions of the young end-body
within the button. Figures 69, B, and 72, A, show distal resorption
under way when no noticeable accretion is evident in the sulcus.
Soon, however, relatively rapid growth processes do take place in
the sulcus. The general morphologic aspects were described in a

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 399

previous chapter and illustrated by figures 69 and 71. It should
be emphasized that most active growth of this region occurs in the
pre-molting stage and that the newly forming proximal lobe remains
hidden beneath the caudalmost portion of the molt. The molt is
engaged in the sulcus by its constricted end, which fits the sulcus
like a collar.

Histologically, the earliest and most important features of accre-
tion occur in the epidermis. The basic process is a simple epithelial
invagination in the sulcus beneath the proximal border of the button.
When a new lobe begins to develop, the ingrowing epithelial lamina
extends obliquely forward (see fig. 76, an enlargement of the area
indicated by an arrow in fig. 72, A). The thickened and reverted
margin of the horny button is seen to overlie the sulcus, the width
of which measures approximately 300 microns. The arrow (fig. 76)
marks the beginning of an epithelial invagination. The ingrowing
lamina lies adjacent to the last subsquamal fold. The overlying
tips of the last scales may be noted farther proximad.

The epidermis lining the sulcus is a relatively thin, deeply staining
squamous stratified epithelium. It consists essentially of the stratum
profundum. Unquestionably the deepest layer is in an active period
of proliferation, although the dense accumulation of pigment cells
prevented the detection of mitotic figures. The chromatophores
represent a characteristic feature of the growth zone in the sulcus;
they form here the darkest area in the derma and crowd directly
against the deepest cell layer of the epidermis.

Our sections of young end-bodies further show that differentia-
tion and growth on the ventral side of the sulcus precedes that on
the dorsal. We observed repeatedly that the epithelial lamina
separating a newly forming third lobe from the last scales was present
on the ventral side, whereas dorsally the second lobe was still growing
by expansion. This difference obviously is the cause for the dorso-
ventral asymmetry of the end-body and of the rattle as a whole.

The histological features of the developing basal lobe in further
advanced stages are illustrated in young specimens of the timber
rattlesnake (Crotalus horridus; figs. 77 and 78), which measured,
respectively, 366 and 373 mm. in total length. At the time of death
the cornea of the smaller one was clear, that of the larger slightly
milky. The histological findings confirm the fact established long
ago by reptile keepers, that clouding of the cornea precedes molting.

The newly forming basal lobe is the result of two growth processes
(see fig. 77) : the invagination of the epithelial lamina directly caudal

400

FIELDIANA: ZOOLOGY, VOLUME 32

to the last scales, and the vigorous expansion of the fibrous connective
tissue immediately distal to that lamina. The former process pre-
pares a new surface epithelium beneath which the growing connec-
tive tissue produces the bulging new lobe. Characteristically, the
expanding connective tissue is rich in chromatophores that crowd
beneath the basal layer of the covering epithelium. In the younger

FIG. 77. Cross section of a newly forming basal lobe in a young timber
rattlesnake (Crotalus ft. horridus). The opposite directions of the invaginating
epithelial lamina and of the bulging connective tissue of the new lolje are indicated
by arrows. The expanding lobe is rich in chromatophores.

stage (fig. 77) the total width of the base of the newly forming lobe
is 0.9 mm. and its height approximately 0.6 mm.

The epithelial ingrowth is bulbous at its deepest point, where
proliferation appears to be most active. This point lies close to the
caudal insertion of the dorsal musculature on the base of the style.

A clear delamination within this epithelial ingrowth has not
occurred at this early stage but signs of that process are unmis-
takable. The epithelium covering the young lobe is relatively thick
(about 100 microns) and consists mainly of a stratum profundum.
Its layers stain intensely blue with hematoxylin. The outer layer
of this young epidermis stains lightly with eosin, and includes many
cells of the squamous type. Chromatophores and eosinophile granu-
locytes, two interesting cell types foreign to the epidermis proper,
begin to migrate into the multilayered epithelium.

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 401

Eosinophile granulocytes invade the epidermis from the subjacent
connective tissue, where they can be observed in groups. Their
presence further confirms our opinion that delamination within the
epidermis, which here must occur in order to establish a lumen of
the newly forming sulcus, is not merely a physical process of fissura-
tion between cell layers, but is initiated by the highly probable
proteolytic action of these leucocytes.

The newly forming lobe is covered by the terminal portions of
the future molt, which is also invaded by numerous eosinophile
leucocytes that penetrate it in groups from the derma of the sub-
squamal folds. Leucocytic activity appears definitely related to the
process of ecdysis.

At a more advanced stage in the development of a new lobe
(fig. 78) the epithelial lamina separating the future basal lobe from
the last scales had grown deeper and, correspondingly, the developing
lobe had attained greater height (1.5 mm.). The base of the lobe
is still relatively narrow (1.2 mm.). Individual variability may
account for the differences in the shape of the new lobe of the speci-
mens illustrated and described. Pigmentation of the dermic portions
of the lobe is heavy. In the epithelial lamina a stratum granulosum
represents the preparatory stage of keratinization. As compared
with the conditions described for the smaller specimen (fig. 77),
the leucocytes here are more numerous in the connective tissue
adjacent to the invaginating lamina. They are likewise aggregated
in clusters. Some of these granulocytes had penetrated the stratum
germinativum of the epidermal ingrowth. They are also invading the
future molt from the connective tissue of the subsquamal folds on
the under surface of the last scales.

Very few leucocytes are present within the growing basal lobe
itself. This is well supplied with bloodvessels that radiate toward
the periphery, and some of the arterioles are accompanied by small
nerves. These vessels and nerves are terminal branches emerging
from the interior of the style.

Longitudinal sections of the specimen in the later pre-moiting
stage show interesting differences in the distribution of leucocytes
through the distal and middle lobes. Resorption in the style and
in the connective tissue of the soft parts had preceded the formation
of a new lobe. On the ventral side of the end-body the connective
tissue of the distal and middle lobes is rich in intercellular fluid
(edema). The small bloodvessels stain poorly and some appear

402

FIELDIANA: ZOOLOGY, VOLUME 32

ruptured. Associated with these apparently degenerating vessels
one again finds groups of eosinophile granulocytes.

On the dorsal side, in contrast, where growth changes lag, the
connective tissue is dense, with few signs of an edematous condition.
This is particularly apparent in the middle lobe. Here, the connec-

FIG. 78. Cross section of a farther advanced basal lobe of a young timber
rattlesnake (Crotalus h. horridus). The epithelial lamina of the sulcus is deeper
but still solid; delamination occurs later. The new basal lobe is richly vascular.
Chromatophores form a dense layer directly beneath the stratum germinativum.

tive tissue contains large numbers of leucocytes that apparently
have not begun their proteolytic activity.

The absence of leucocytes in a newly growing lobe, their scarcity
in the ventral portions of the end-body that have already degenerated
to an edematous condition, and, finally, their predominance in the
dorsal regions that subsequently undergo similar degenerative
changes, make it highly probable that eosinophile leucocytes are
concerned with such atrophic changes. Most likely their activity

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 403

in producing a proteolytic enzyme leads to a partial hydration of a
previously dense and tough fibrous end-body.

A thick stratum granulosum is present in any portion of the
epidermis where hornification soon follows. In unstained sections

FIG. 79. Parasagittal, longitudinal section of the tail of an adult massasauga
(Sistrurus c. catenatus). The snake was in the late accretion period and the molt
still covers the newly formed basal lobe. The space between the last formed horny
segment of the rattle and the end-body was filled with a spongy stratum inter-
medium.

that layer appears opaque and white; it stains deep blue, indicating
probably an acid condition of the cells that contain keratohyalin
granules.

GROWTH OF THE RATTLE IN THE MASSASAUGA

In order to compare growth of the rattle in Crotalus with growth
in the other genus of rattlesnakes we studied the end-body in Sistrurus
catenatus. Our comparative study of histological sections failed to
reveal any basic differences in mode of rattle formation. The fore-
going description of histological features in Crotalus, therefore, holds
in the main for Sistrurus.

Figure 79 shows the morphological aspects, in parasagittal
longitudinal section, of the tail end of an adult Sistrurus from which

404 FIELDIANA: ZOOLOGY, VOLUME 32

the free elements of the rattle had been removed. The snake had
a total length of 685 mm. When the animal was killed the corneae
were clear and the skin was peeling. The figure reveals a new basal
lobe largely established. It is covered by the molt still attached to
the proximal margin of the last-formed segment. At the time of
fixation a new generation of stratum corneum had already been
formed as the outermost layer of the epidermis covering the end-
body. The interval between the two generations of stratum corneum
was filled with a spongy stratum intermedium. The specimen,
therefore, was definitely in the late accretion phase.

The following histological features were observed in sections of
the same material: (1) Some degree of osteolysis on the trabeculae of
the style is evident. This affects the periphery of the main portion
of the style rather than the tips. The free ends of such trabeculae
appear to lose their affinity for stains, probably through demineraliza-
tion of the ossein. Eosinophile granulocytes and osteoclasts were
observed in the adjacent tissues. Such bone resorption is relatively
insignificant in comparison with that seen in very young specimens
of Crotalus and described in the preceding section. That the style
as a whole is not appreciably affected by resorption in adult specimens
is further proved by our x-ray findings on the live Crotalus atrox;
(2) the connective tissue of the distal lobe, and to a lesser extent of
the middle lobe of the end-body, was edematous and infiltrated by
numerous leucocytes, which had also invaded the vacuolated stratum
intermedium; (3) the dominant cellular elements in the bone marrow
of the caudalmost vertebrae and in the style were eosinophile
myelocytes, the precursors of the eosinophile leucocytes; (4) the
deepest layers of the epidermis over the end-body were being invaded
by the pigmented processes of numerous chromatophores from the
subjacent derma. As there was no stratum granulosum, keratiniza-
tion of a new stratum corneum evidently had not begun.

Two younger specimens of Sistrurus catenatus were likewise
studied histologically. One was in the latent stage. The other
was in the pre-molting or accretion phase, showing a thick stratum
granulosum as a characteristic feature of the epidermis during this
period, which comes late in the cycle.

A representative section of this end-body in the latent stage (see
fig. 80, A) is slightly parasagittal and does not reveal the caudalmost
extension of the spinal cord. (In other sections of the same specimen
the spinal cord terminates well within the style just distal to the
groove between the proximal and middle lobes.) The trabeculated

.5 mm.

B

FIG. 80. Representative longitudinal sections of the end-body of two mas-
sasaugas (Sistrurus c. catenatus): A, in the latent period; B, in the accretion phase.
In A the stratum corneum is seen as the outer layer of the relatively thin epidermis
of the end-body. (Its partial detachment is an artifact.) In B the epidermis has
thickened. Its characteristics are a well-developed stratum granulosum (heavy
black layer) and a degenerating stratum intermedium. The stratum granulosum
becomes the immediate mold for the new horny segment.

405

406 FIELDIANA: ZOOLOGY, VOLUME 32

nature of the style and its marrow spaces, the caudal termination
of the muscular masses, and the relative position of the last articu-
lated vertebra with respect to the sulcus are indicated in the figure.
The detachment of the stratum corneum obviously is an artifact,
due most likely to shrinkage of the soft parts in the end-body during
preparation. An epithelial ingrowth at the sulcus may be noticed.
It is undercutting the basal lobe and provides the earliest indication
of the future widening of the sulcus.

There is evidence of bone growth on the periphery of the proximal
portion of the style. This osseous framework of the end-body
naturally will increase in size with the rapid growth of the snake.
In the section, a fairly continuous single row of osteoblasts lies
adjacent to the margins of the style. In the distal portions bone
growth is much less marked.

Figure 80, B, shows a midsagittal section through the end-
body of the specimen in the pre-molting stage. The caudal termina-
tion of the spinal cord within the style and the course of the caudal
artery adjacent to its ventral surface are evident. The formation
of a new basal lobe was well under way at the time of fixation. In
section, the lobe measures 3 mm. in width on the ventral side.
Osteogenetic processes are identified by numerous osteoblasts that
line the dorsal margin of the style in an almost continuous layer.
Similar growth is noticeable also along the trabeculae in the distal
portions of the style. Appositional bone growth, therefore, appears
fairly active in the late accretion period.

In the pre-molting stage the stratum intermedium between the
old stratum corneum and the newly formed thick stratum granulosum
shows advanced degeneration by its highly vacuolar condition. Eosin-
ophile leucocytes within that layer have largely disappeared. Near
the sulcus and in the attached molt they are still numerous and stain
well. The connective tissue stroma of the soft parts in the end-body
shows no evidence of edema nor are there any signs of pronounced
infiltration by eosinophile leucocytes.

The most striking histological feature of the pre-molting stage
is the thick stratum granulosum of the epidermis. In the section
illustrated (fig. 80, B) this layer had been broken by fixation. During
the pre-molting phase the epidermis always is markedly thickened
over the end-body and appears white before embedding. Its stra-
tum granulosum is the immediate mold for the new horny layer of
the rattle segment, which differentiates during the final phase of the
active period. In the section shown, there is no sign of that final

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 407

keratinizing process. If it preceded the maturation of the new basal
lobe, the new horny segment would prematurely imprison the con-
nective tissue of the end-body and successive rattle segments,
necessarily, would become smaller.

SUMMARY

The development and growth of the rattle of the rattlesnake has
never before been adequately investigated. Having abundant
material and modern histological and x-ray technique at our disposal,
we have undertaken such an investigation.

Although the gross anatomy of the functional rattle had been
described by Klauber and others, we have illustrated a young,
tapering rattle and an old non-tapering one to show exactly how
the segments, each with three lobes, are interlocked, and to point
out certain differences due to age. After the relationship of one
horny segment to another is understood, the functional part of the
rattle may be disregarded and attention focused on the living tip
of the tail where actual development and growth take place. This
tip, the end-body, consists of (1) a bony core formed largely by the
coalescence of several terminal vertebrae and known as the style or
shaker; (2) fibrous connective tissue; (3) epidermis. The shape of
the end-body is like that of a segment. The end-body forms and
molds the successive segments, which become detached from it
synchronously with the sloughing of the molt, an event that takes
place from two to four times a year.

If the end-body as the living mold of each segment remained
constant in shape and in its relationship to the last-formed segment,
the successive units of the rattle obviously would be arranged one
inside the other. The horny segments are not nested but extend in
a linear series, the two proximal lobes of one capping the two distal
lobes of the next younger. The manner by which the segments
assume this special alignment had never been fully explained. Just
how does the basal lobe of each segment come to be exposed? It is
obvious that a shift of some sort takes place during the formation
of each segment.

This shift occurs as follows: First, a fourth lobe is temporarily
added to the end-body. This new element is formed just proximal
to the third or basal lobe. This growth dislocates caudad the last
horny segment and all preceding ones a distance equal to the length
of the new lobe. Second, the four-lobed condition of the end-body

408

FIELDIANA: ZOOLOGY, VOLUME 32

is suddenly reduced by resorption mainly of the distal lobe. This
re-establishes a three-lobed end-body but leaves behind the just-
detached stratum corneum, which itself becomes the latest segment
of the rattle. Thus there is growth at the base of the end-body
followed by reduction of its tip. The sequence of the histological
phenomena is the key to the problem. The caudad dislocation of a

FIG. 81. Diagram of the molting cycle of a Texas diamond-back (Crotalus
atrox) illustrating the relative duration of latent and active periods. Within the
period of resorption the molt is cast and a new segment is added to the rattle.

single functional segment and thereby of the whole rattle is fully
accounted for without invoking wave-action in soft tissues, slipping
of one horny segment partly off another, or any of the fanciful
theories that have been advanced.

There is a modicum of truth in the statement of early herpetolo-
gists that the segment of a rattle is nothing but a little of the molt
left on the end of the tail. That out-moded statement Fs useful
only in emphasizing the relation between the molt and a new rattle
segment: As the molt is released from the body as a whole the newly
keratinized segment being added to the functional rattle is released
from the end-body. The intimate relationship between molting and
segment formation is thus evident. In the newly forming rattle
segment the epidermal layers, which are truly homologous to the

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 409

molt, consist of the stratum corneum, which is retained, and of the
stratum intermedium, which degenerates.

In order further to elucidate this relationship we periodically
x-rayed the tail of a living Texas diamond-back throughout one
molting cycle. The complete cycle observed may be divided into
three periods: a long quiescent or latent period, a much shorter one
of active growth or accretion, and a brief one of resorption and
ecdysis. The approximate duration of each is indicated by the
diagram (fig. 81).

Individual and species variation in the duration of a cycle as
well as in its parts is to be expected; age, season, temperature, and
probably nutrition will have their effects.

Important histological changes that take place in the end-body
during each molting cycle concern the connective tissue and the
epidermis. The bony core or style, once formed, shows no charac-
teristic periodic changes.

During the latent period the connective tissue remains unchanged,
but when accretion begins, new connective tissue forms the bulk of
the developing fourth lobe. The expansion of this lobe forces the
soft tissues of the end-body caudad along the style. Then, during
resorption, a drastic reduction of connective tissue takes place in
the distal and, to a limited extent, in the median parts of the end-
body. Just prior to resorption, the connective tissue becomes
edematous and there is infiltration by a great number of leucocytes
derived from the marrow of the style. Their presence appears to
be functionally related to resorption. This process is comparable
to the reduction of the tail in a metamorphosing tadpole except
that, in the amphibian, resorption is complete, and no keratinized
shell, shaped like the original tail, is left behind.

The degenerative or absorptive changes that take place in the
epidermis of the end-body chiefly concern the stratum intermedium.
This cell layer is invaded by great numbers of eosinophile granu-
locytes, which are large leucocytes. They appear to advance the
processes of resorption by proteolytic action. This degeneration
allows a new generation of stratum corneum to become widely
separated from the older generation as the remaining layers of the
epidermis follow the rapidly shrinking connective tissue to its new
position. The stratum intermedium temporarily becomes a jelly-
like mass; its cells are highly vacuolated and, taken as a whole, lose
the characteristics of a squamous layer. The stratum intermedium
later dries out, the old generation of keratinized stratum corneum

410 FIELDIANA: ZOOLOGY, VOLUME 32

becoming free as a newly functional segment of the rattle. X-ray
photographs reveal remnants of this intermediate layer between
successive segments; evidently it contains traces of mineral salts.

The deeper layers of the epidermis and its supporting connective
tissue are continuous with the similar elements of the new basal lobe
and with them form an end-body that will lie dormant throughout
the latent period just beginning. Keratinization of the new stratum
corneum over this young end-body takes place immediately. It
begins at the tip of the distal lobe and progresses toward the basal
lobe. In tails covered by the loose molt, ready to be shed, the new
generation of stratum corneum of the end-body is firmer at the tip
than on the basal lobe.

The initiation of growth of the new basal lobe in the sulcus is
a process of simple epithelial invagination. Active cell proliferation
occurs in the deepest epidermal layer, the stratum profundum, and
the resulting ingrowing lamina extends obliquely forward. A con-
comitant feature of this process is the dense accumulation of pigment
cells or chromatophores that crowd upon the stratum profundum
and prevent detection of mitotic figures. This activity is hidden
beneath the caudalmost scales, which soon will be cast off as part of
the molt. The swelling of the growing lobe, produced by rapidly
proliferating connective tissue, noticeably stretches the skin and
spreads the overlying scales apart several days before the molt is
shed.

The significant events of development of the rattle take place
in the late embryo and during the first few postnatal months; there-
after the segments merely become larger as the snake grows. Devel-
opment in the three species investigated, the Florida and the Texas
diamond-backs and the massasauga, is essentially similar except for
an interesting temporal difference : In the massasauga, fusion of the
caudalmost vertebrae to form the style occurs after birth, whereas
in the Florida and Texas diamond-backs fusion is well-advanced in
late embryos. The tail of the massasauga, therefore, is more primi-
tive in showing at birth less differentiation. This is in agreement
with several other characters of Crotalus and Sistrurits.

The style of both species is formed chiefly by coalescence of the
terminal caudal vertebrae, about nine of them being involved.
Certain extravertebral additions occur: a few nodular trabeculae
from osteogenic foci in the tip of the end-body distal to the vertebrae,
and a pair of elongate trabeculae on the ventral side near the base
of the developing style. In young snakes the origin of the proximal

ZIMMERMANN AND POPE: RATTLE OF RATTLESNAKES 411

portion of the style is betrayed by regularly overlapping ridges and
trabeculae suggestive of the segmental arrangement of the vertebrae;
in old snakes the style becomes relatively homogeneous. The muscles
that vibrate the rattle are inserted in the base of the style, the last
normal vertebra articulating with it and acting as the pivot of
vibration.

In newborn rattlesnakes the outer covering of the end-body is
a stratum corneum soon to be detached as the keratinized button.
This simple structure terminates every complete rattle and has only
two lobes, both poorly defined. The first two segments that follow
the button are intermediate in structure between it and later three-
lobed units of the rattle.

There are two interesting differences between the earliest post-
natal growth processes of the end-body and later ones: (1) the stratum
corneum of the first segment succeeding the button is keratinized
before rather than immediately after the formation of a basal lobe;
(2) bone resorption occurs at the distal tip of the style only during
the initial molting cycle.

REFERENCES

BOLK, L., GOPPERT, E., KALLIUS, E. AND LUBOSCH, W. [EDITORS]

1931. Handbuch der vergleichenden Anatomie der Wirbeltiere. 1, pp. 375-448.
Berlin.

CZERMAK, JOH. N.

1857. Ueber den schallerzeugenden Apparat von Crotalus. Zeitschr. wiss.
Zool., 8, pp. 294-302.

GARMAN, SAMUEL

1888. The rattle of the rattlesnake. Bull. Mus. Comp. Zool., 13, pp. 259-268,
2 pis.

KLAUBER, L. M.

1940. A statistical study of the rattlesnakes. Part V: The rattle. Occ. Pap.

San Diego Soc. Nat. Hist., No. 6, pp. 1-63, figs. 47-63.
Contains an annotated bibliography of 145 titles.

NAEGELI, OTTO

1923. Blutkrankheiten und Blutdiagnostik. Lehrbuch der morphologischen
Hamatologie. 4th ed. Leipzig.

ZlMMERMANN, A. A., AND POPE, C. H.

1946. The development of the rattle in rattlesnakes. Anat. Rec., 94, p. 65.
[Abstract.]

412

INDEX

age, indication of, 367

bone marrow, 395, 404
bone resorption, 386, 395-398, 404
button, 366, 367, 378, 382, 383, 386-
388, 390, 399

chromatophores, 399, 400, 402, 410; see

also end-body, pigment cells
cornea, clouding of, 375, 399
Crotalus, 357, 393, 403, 410

adamanteus, 360, 374, 376, 378, 382,
383, 386, 387, 389, 390, 392, 393,
395, 397
atrox, 360, 362, 370, 374, 377-381,

404, 408

durissus terrificus, 357
horridus horridus, 361, 399, 400, 402
viridis oreganus, 361, 363-366
viridis viridis, 361, 384, 385
Czermak, Job., 357-359, 374

end-body, 358

accretion, 368, 369, 384, 386, 398, 403,

409

dorso-ventral asymmetry, 399
growth changes, 367
histology, 386, 399, 403, 404
latent period, 372, 404, 409
pigment cells, 358, 388; see also

chromatophores
resorption, 369-372, 382, 386, 393,

409

temporary fourth lobe, 369, 407, 409

epidermis, 384, 390, 399-401, 404, 405

invagination of an epithelial lamina,

399-401, 410

keratinization, 390, 391, 393, 410
stratum basale, 390
stratum corneum, 378, 390, 404, 405,

410

stratum germinativum, 390, 401, 402
stratum granulosum, 390, 392, 403-

406
stratum intermedium, 386, 391-393,

398, 403-406, 409
stratum profundum, 390, 399, 400,

410
stratum spinosum, 390

Florida diamond-back, see Crotalus
adamanteus

Garman, S., 374, 376

granulocytes, eosinophile, 392, 394, 395,

400-404
granulopoiesis, 395

Klauber, H. M., 357, 361, 363, 369, 392,
407

Leuckart, 357, 374

leucocytes, eosinophile, 392, 394, 401,

402, 409
live rattlesnake, method of handling for

x-rays, 362

x-ray photographs, 370, 371, 373
x-ray technique, 362, 363

massasauga, 361, 403-405, 410; see also
Sistrurus and S. catenatus catenatus
molting cycle

of live Texas diamond-back, 375
latent period, 375, 404, 408, 409
period of accretion, 372, 375, 398, 404,

408, 409
period of resorption, 372, 375, 408,

409
myelocytes, eosinophile, 395, 404

Naegeli, O., 394

osteoclastic activity, 393, 404
osteoclasts, 395-396, 404

Pacific rattlesnake, see Crotalus viridis

oreganus
prairie rattlesnake, see Crotalus viridis

viridis

rattle, development, 381, 410
rattle, fully developed; anatomy, 363
rattle, general, 357, 358

shaker, 374, 377
Sistrurus, 357, 403, 410

catenatus catenatus, 361, 374, 376,

403-405

style, 358, 368, 374, 377, 378-381, 393,
406, 410

Texas diamond-back, see Crotalus atrox
timber rattlesnake, see Crotalus horridus
horridus

warning theory, 357
White, C. E., 361

413

UNIVERSITY OF ILLINOIS-URBANA