HARVARD UNIVERSITY

Library of the

Museum of

Comparative Zoology

Numter 6

May 18, 194U

A STATISTICAL STUDY OF THE RATTLESNAKES

By Laurence M. Klauber

Curator of Reptiles and Amphibians,
San Diego Society of Natural History

VII* THE RATTLE, Part 1

Table of Contents

Introduction 3

Purpose of the Rattle and Method of Use 4

Character of the Sound 12

Method of Vibration 12

Speed of Vibration 13

Lengths of Rattle Strings 16

Chemical Composition 23

Method of Rattle Formation - The Prebutton 25

Interrelationshin of Rattle and Skin; Development

of a Rattle Segment 26

Coincidence of Shedding with Rattle Formation 29

The Rattle-Shifting Mechanism 31

Structure of the Rattle 35

Species Differences 4-1

Use of the Rattle by Man 43

Acknowledgments 4-3

Summary and Conclusions -43

Bibliography ^ 4'''

" Previ:>us sections of this study have appeared in the Occa-
sional Papers of this series as follows:

No. 1, Aug. 10, 1936 I Introduction

II Sex Ratio in Rattlesnake Popula-
tions
III Birth Rate
No. 3, Dec. 15, 1937 IV The Growth of the Rattlesnake
No. 4, May 31, 1938 V Head Dimensions
No. 5, Aug. 30, 1939 VII Fangs

A STATISTICAL STUDY OF THa RATTLESNAKES
By Laurence M. Klauber
l^, Lb Z VII THE RATTLE, Part 1

Introduction

The rattle is the peculiar attribute of the gr jup
of snakes of the New World which are the subject of this
study. This remarkable appendage, whose purpose has been
a matter of speculation and controversy since these snakes
were first observed witn astonishment approaching incredu-
lity by European naturalists, offers interesting research
problems. For we have in the rattle a mechanism (possibly
unique among the vertebrates) which, by the complete separa-
tion of successive elements from the body, renders such ele-
ments independent of succeeding bodily grovrth, or other
changes short of complete loss or destruction. Hence, if we
can determine the chronology of each separation, and the cor-
relation between rattle- and body-size, we have a measure of
the rate of bodily grov;th. The rattle, as corroborative evi-
dence in growth cxirves, has already been mentioned briefly
in this series (Occ. Pap. No. 3, p. 15) but much remains to
be added. For the rattle, in both relative size and growth-
trend, is found to show definite species differences. There
are not only such obvious differences in the comparative
sizes of the rattle and the body as that between Sistrurus
catenatus and S. miliarius . but different rattle-growth equa-
tions are found in species not greatly different in the adult
stage, as, for example. In Crotalus molossus and Crotalus
durissus; for the latter always has smaller rattles when
young, but rapidly overtakes molossus. Such growth trends
sometimes suggest relationships subsequently validated by
other characters — for instance that between C. durissus and
C . enyo.

These species comparisons, involving correlations
between three variables — body length, rattle number (i.e.
position in the string), and rattle dimensions — complicated
further by sexual dimorphism, involve a considerable volume
of data, too large for a single number of this series, if
included with other aspects of the rattle. It therefore ap-
pears necessary to delay the presentation of the studies of
dimensional differences between species, and to limit the
present publication to a discussion of those more general
features of the rattle cormnon to all species.

The literature on the rattle is quite extensive;
nearly every general work on reptiles mentions rattlesnakes,
which inevitably brings in the rattle. There is no pretense
that the bibliography at the end of this paper is complete;
it is primarily concerned with references devoted to the
rattle exclusively, or such as contain a novel or pertinent
discussion.

It is not my purpose to review exhaustively the
various descriptions and arguments which have been presented
respecting the formation and use of the rattle. Rather, I
shall emphasize the particular phases of development which

appear to me to have received inadequate, and, in some in-
stances, inaccurate treatment. In the classical descriptions
of the formation of the rattle, especially those contained in
the important papers of Garman (1883 and 1889) , there are
three items requiring modification or further clarification.
These are, first, the character and disposal of the birth-
rattle or prebutton; secondly, the relationship between the
rattle and the skin; and finally, a modified and more com-
plete description of the intricate mechanism whereby each
rattle is slipped backward, yet remains interlocked with its
successor, when a new segment has been perfected. But before
discussing these features I shall summarize the theories on
the purpose of the rattle and my own conclusions.

Purpose of the Rattle and Method of Use

The literature on the rattle is full of speculations
upon its purpose; these have been proposed in great variety
since tne earliest encounters between man and rattlesnake.
Some of those which have appeared most frequently conclude
that the rattle is. used:

As a mating call

As a call for help from other rattlers

To warn other rattlers of danger

To paralyze prey with fright, or startle them into
immobility

To decoy prey through curiosity

To charm prey with sound

To lure prey

by imitation of the cicada or other insects
by imitation of running water

As a warning to prey

As a warning to intruders

to avoid injury to the intruder by the rattler
to avoid injury to the snake by the intruder
to conserve venom and fangs

To prevent an attack on the tail by insects or
carnivores

As a substitute for hissing

As a religious warning

As a venom-carrying weapon

Most of these theories are quite old and have been
often repeated; it is difficult to determine with certainty
who originated any of them. And both their antiquity and
their character indicate that some were proposed by writers
who had little opportunity to observe rattlers in the wild,
or even under the artificial conditions of captivity. The
early compilers of natural histories had to depend on the
accounts of travelers in the Americas, or observations on
the few specimens which survived the long cold trip to

Europe. So it is not surprising that some of these theories
show slight knowledge of rattlesnake habits or their actions
under natural conditions, or even of the sound of the rattle;
the surprising thing is the continued repetition today of
theories and characterizations, many of which do not square
with the most elementary observations.

For example, desert rattlesnakes are as well equip-
ped with rattles as any others; yet desert prey would hardly
be lured by an imitation of water, nor does the sound of the
rattle much resemble that of a gurgling brook. Again, the
rattle cannot be considered a substitute for the missing hiss,
since rattlesnakes do hiss violently when alarmed.

In considering the many theories which h ave been
suggested to explain the fundamental purpose of the rattle,
it is at once apparent that a normal rattlesnake reaction,
which is almost universal, gives strong support to one theory
as compared to all the others. For what does any rattlesnake
do with his rattle? He sounds it when disturbed, as by some
movement or the approach of an intruder. Upon this there can
be no argument; it is the common experience of everyone who
has encountered a rattler in the field or startled one in
captivity, unless the snake has become accustomed to humans.
I have no desire to assume an irritating attitude of assur-
ance, but certainly I have seen this happen a thousand times;
and there are others of far more extensive experience who
corroborate this observation.

The rattle is used as a warning signal; upon this,
experience permits no argument; if it has other alternative
purposes the' burden of proof is upon those who advance them,
for they are not matters of daily observation. The "Keep
Your Distance" use of the rattle is so universal and instinc-
tive that one would suppose this problem of use to have been
long since settled; and I might well be accused of giving
undue importance to a nonexistent argument, were it not for
the fact that books continue to appear, attributing to the
rattle as\ much uncertainty of purpose, or diversity of use,
as the popular natural histories of 1300 (e.g. Berridge,
1935; Ditmar:, 1936; Gillespie, 1937). And it must be ad-
mitted that the nature and purpose of the warning have some-
times been so misinterpreted as to throw the theory into
doubt and disrepute.

The warning theory is very old (e.g. Piso, 164-3)
but from the first it has been, and still is, a matter of
argument because of a confusion, often unrecognized, between
three types of warning: that is, the warning of (l) intrud-
ers, possibly dangerous, warned for the protection of the
snake; (2) intruders warned for the protection of the warnee,
instead of the warner; or (3) prey, the last being a special
case of (2) . If we sharply distinguish between these three
and point out the manner in which the rattle is beneficial to
its possessor, by frightening away creatures which might
otherwise injure the snake, we at once eliminate the objec-
tions which have been so often advanced against the warning
theory. Certainly no animal will warn away the food upon
which it depends for subsistence, or develop so intricate a
mechanism for the altruistic protection of the innocent
passer-by. But it is equally clear that warning devices
which tend to safeguard their owners are common in nature;
and there is no more reason to question the purpose of the
rattle than of these other devices. It is only necessary to

show that the rattle is used for this purpose, is often ef-
fectual, and that its disadvantages are not as important as
have sometimes been supposed.

I shall give no further attention to those arguments
against the warning theory which are based on this confusion
between a beneficial warning, and one neutral or injurious to
the rattler. Darwin himself tried to correct this confusion
many years ago, with indifferent success; in fact, his posi-
tion is still occasionally advanced as an argioment against
the theory, notwithstanding his clear statement in advocacy
of it.*

As I have stated, the validity of the dangerous-
enemy warning theory may be presumed to rest with the answers
to these questions:

(1) Is the rattle used as a warning?

(2) Is it effective?

(3) Does it ever react against the rattler, and if
so, to what extent?

On the first point I think enough has been said
already. The rattle is so used — this is the reaction which
anyone familiar with these snakes expects invariably when he
approaches one.

Is it effective? The answer, of course, depends on
the circumstances, as it would were we discussing the growl
of a dog, the hiss of a gander, or the earth-pawing of a bull.
These are all warning reactions, and they may result in suc-
cess or failure, depending on the character and purpose of
the trespasser. They may be followed by more direct action
or by the retreat of the warner.

One of ..he mistakes made by early writers, and

♦ "It is admitted that the rattlesnake has a poison fang for
its own defense and for the destruction of its prey; but some
authors suppose that at the same time this snake is furnished
with a rattle for its own injury, namely, to warn its prey to
escape. I would almost as soon believe that the cat curls the
end of its tail when preparing to spring, in order to warn the
doomed mouse." Origin of Species, Fifth Ed., 1871, p. 196.

As a result of criticism, particularly from St. George
Mivart, Darwin clarified the statement in the subsequent edi-
tion of tne Origin: "Some authors suppose that at the same
time it (the rattlesnake) is furnished with a rattle for its
own injury, namely, to warn its prey .... It is much more
probable that the rattlesnake uses its rattle, the cobra ex-
pands its frill, and the puff-adder swells whilst hissing so
loudly and harshly, in order to alarm the many birds and
beasts which are known to attack even the most venomous spe-
cies. But I have not space here to enlarge on the many ways
by which animals endeavor to frighten away their enemies."
Origin of Species, Sixth Ed., 13B5, p. 162. But even this
statement has not prevented his being cited as opposing the
warning theory, when, in fact, he opposed only the misin-
terpretation of enemy warning into prey warning.

repeated in natiiral histories, is irl the general characteriza-
tion given of the rattler's warning response. Usually the
snake is portrayed in his resting coil with rattle sounding.
To the possible inadequacy of such a warning mechanism one
may agree, but this is not the entire story by any means.
Just as the cat arches its back, fluffs its tail, opens its
mouth, and spits and squalls, so the rattler has more than a
single warning reaction. It is true that a snake found in
his resting coil may first sound the rattle without changing
position. But if this preliminary warning fails to halt the
trespasser he quickly adopts more spectacular methods, which
concurrently place him in a better position either for de-
fense or escape. For now, still sounding the rattle furious-
ly, the anterior part of the body is raised above the ground
in an S-shaped spiral, with the head and neck held like a
poised lance ready for a forward lungej the posterior part
of the body is flattened to stabilize the anchorage or facil-
itate mobility; and the snake inhales and exhales with a
violent hiss.* Meanwhile, in an endeavor better to sense the
situation, the tongue is protruded to the fullest extent,
and is held alternately pendent and vertically erect, with
the tips widespread. Now he is ready for whatever may come;
he can strike, if the enemy comes within range, or he can re-
treat (still facing the intruder) toward the nearest rock or
bush which may serve as a refuge.

When a rattle is heard under such circumstances,
no prior experience is necessary for a realization of the
danger that lies behind it; both the stridency of the sound,
and the other actions of the snake to which the sound draws
attention, can leave no doubt in the mind of any creature
capable of tiie most elementary reactions of self-protection,
that retreat is advisable. Several of the early writers at-
tribute to the noise of the rattle an effect mysterious in
its action, in that animals having no previous experience
with it — such as the European horse — seem fully aware of its
purpose and recoil instinctively from the sound. While the
factor of experience may have had some bearing on the genesis
of the rattle, as is discussed hereafter, we need no longer
consider such experience necessary, especially when this
startling sound begins suddenly and unexpectedly under one's
very feet. But the effect of the warning is by no means de-
pendent entirely on the initial surprise, so strident is the
noise, and so alarming the other actions of the snake.

In judging the effectiveness of this warning as a
means of protection, one must have in mind the kinds of car-
nivores and birds of prey which would seek the rattler as
food, and the ungulates which might tread on it fortuitous-
ly. Against many of these, such as wolves, coyotes, bobcats,
and the like, the rattle would be a valuable protective ad-
junct— not unfailing, of course, but still of major impor-
tance in frightening these creatures into looking elsewhere
for a meal. One has only to watch the reaction of a dog or
cat with a rattler to see how effective this is, notwith-
standing the occasional newspaper stories of rattlesnake
killers. And the wild creatures are usually more wary than

* It is suggested that those who believe rattlers cannot hiss
(once the subject of controversy in the pages of a sporting
magazine) test this by removing all the loose segments of the
rattle, or by binding it up with cloth so that the noise of
the breath will not be masked by that of the rattle.

a dog backed by his master. Cowles, 1938, observed that
western skunks were driven from customary feeding grounds by
the sound of the rattle. Hawks, owls, and ravens would be
similarly frightened.

As to the larger herbivorous mammals, they are not
usually of such a disposition as to go out of their way in
search of trouble. Judging from observations on the reac-
tions of horses and cattle, it is quite evident that the
rattle has a definite protective value. In fact, with re-
spect to the ungulates, a specialization of the enemy warn-
ing theory has been suggested as indicating the conditions
which, coupled with a widespread warning reaction of other
snakes — the vibration of the tail when annoyed or alarmed —
led to the evolution of the rattle. This theory, which has
been accepted by many herpetologists, assumes the ancestors
of the rattlers to have lived in a prairie country Inhabited
by herds of hoofed animals akin to the modern bison. "The
hoofs of these ponderous animals, traveling over the plains,
must have been distinctly dangerous to snakes living in the
open, while the snake's bite would be distinctly unpleasant
to the bison, although death would probably have rarely en-
sued owing to the beast's great bulk. The bison would glad-
ly keep out of the snake's way, however, if warned, and this
warning the rattle gave." Barbour, 193^4, p. 42. This re-
finement of the warning theory was suggested by 0. P. Hay as
early as 1887 (p. 214-). It is to be noted that, in the early
stages of the development here pictured, recognition of the
sound, rather than its alarming stridency, would have been
the effective element. But, with the present highly develop-
ed instrument, recognition through experience is no longer
essential; tmgulates are likely to heed the harsh warning of
the rattle, preferring, as they do, to avoid disturbance and
take a peaceful course.

Garman (1889) describes the transition whereby the
rattle probably developed, through the modification and
specialization of the conical spine which is the caudal ter-
mination of all snakes, but relatively larger and heavier in
some than others. The covering of this cone is shed with
the rest of the skin. By a gradual enlargement of the cone,
a thickening of the covering, with a constriction at its an-
terior end, and a coordinating middle constriction, the suc-
cessively-mounted tail cones would be retained in a loose
chain, thus producing a sound when the snake vibrated its
tail in auger, as is the custom of so many snakes. The cul-
mination of such an evolutionary process would be the present
intricate and highly specialized mechanism. Garman also
makes the point that the warning serves as a means of conserv-
ing venom and fangs, a further advantage to the snake.

Summarizing this phase of the theory, we conclude
that the rattle is of definite value when sounded as a warn-
ing to animals which, with intention or unconsciously, might
injure the snake.

We come now to the third query: Is the rattle ever
a detriment by advertising the presence of the rattler to
potential enemies, when it might otherwise escape?

A niimber of the early writers assume that, like the
bell on a cow, the sound of the rattle is incidental to any
movements of the snake, thus continuously broadcasting its
presence. But, as a matter of fact, unless a snake is crawl-

ing through rocks or brush, so that the rattle strikes ex-
traneous objects, it makes no sound at all. Even when it is
so struck, the click or rasp can be heard only a few feet;
it is quite different from the shrill hiss produced when the
tail muscles deliberately vibrate the rattle. In fact, it is
impossible to simulate the true sound as m^de by the snake,
by rattling a string of rattles by hand, for the necessary
speed cannot be attained.

Others have pointed out that the use of the rattle
may serve only to invite destruction by certain animals, such
as deer and hogs, which are said to kill rattlers upon occa-
sion, i^'or example, Berridge, 1935, considers this a fatal
flaw in the warning theory. Now certainly this would be true
if the rattle served to invite destruction by advertising the
snake's presence, when inconspicuous inactivity, or a silent
withdrawal would be the safer policy. This, however, mis-
judges the rattler's ordinary response to the presence of an
intruder. No field collector knows how many rattlers he may
pass, which escape merely by lying quiet. But I have spied
enough of them, fully aware of my presence, yet making no
sound or movement, to realize that they will depend on this
method of escape whenever possible. The rattle is the reac-
tion of a suddenly startled snake, or one which thinks its
presence to have been discovered and proceeds to use the
warning as a last resort (short of the strike), in the hope
of frightening the trespasser away.*

I think this criticism also misjudges the reaction
of animals to the sound of the rattle. They are pictiired as
saying to themselves when they hear it, as a man would, "Here
is a dangerous creature which I had best destroy, for the
safety of others who may pass this way." But no v/ild animal
would reason thus; there is always a more specific purpose
in attack, such as a desire for food, or the protection of
young. Against such creatures, keeping in mind the fact that
the rattle is not sounded until procrypsis has failed, its
possession is definitely protective.

It is possible that, with man, an enemy v/ho can
destroy at a distance without endangering himself, the rattle
is a disadvantage, ^'or man is poorly equipped with senses
and the rattler may sometimes advertise himself too quickly
when he might otherwise lie undiscovered. But this is no
valid argument against the warning theory, since rattlers
long antedated man in the New World; and the rattle was de-
veloped without regard to the novel conditions involved in
this addition to the local fauna. "^ The same argioment holds

* A defense reaction of a different kind, probably reserved
for king snakes and other ophiophagus species, which are im-
mune to venom, and certain forms of skunks, has been describ-
ed by Cowles, 1933. ihe rattler raises the raid-section of the
body in a bow and strikes downward with it. Even very young
rattlers exiiibit this reaction in the presence of a king snake.
(^ Under certain circumstances the snake's use of the rattle
has been an aid to collectors. Thus J. R. Slevin located
specimens of C. 1. klauberi in the Huachuca Mountains, Ari-
zona, by rapping on stones on talus slides, which caused many
snakes underneath to souna their rattles, and disclose their
whereabouts. The same scheme was used by the Hancock i^^xpedl-
tion on Tortuga Island. Once I had the experience of having
a pyrrhus rattle at ray car as it passed the pile of rocks in
v/hich he v.as hidden.

in the case of the dotaestic hog and goat.

V*e may thus summarize our conclusions with respect
to the warning theory: '-Che rattle is used as a warning; it
is often effective; detrimental results are rare, particu-
larly in the case of the indigenous fauna.

l/ve nov.- pass to some of the other uses of the rattle
which have been suggested and examine the evidence for such
uses, and the arguments with respect to them. Of these the
prey lure, the mating call, and the call for help have been
the most persistently advanced.

The experiences cited in the literature to authen-
ticate them have been few and the evidence usually conflict-
ing, or subject to diverse interpretations (Dudley, 1723;
Shaler, 1872; Aughey, 1873).

My own field experiences and those of my associates
in this vicinity, have been completely negative. Certainly
the best argument against the mating call theory is the ab-
sence of rattling sounds in rattlesnake infested areas during
tne mating season, or when snakes mate in captivity. Fur-
ther, it is to be noted that young rattlers rattle vigorously
long before the acquisition of a second rattle makes the vi-
bration audible. This early use (immediately after birth,
in fact) is hardly to be explained as a sex call.

ihe same reasoning holds with all relationships be-
tween the rattle and prey; if such were the purpose, snakes
would often be heard rattling in the rocks and brush, es-
pecially in the spring; but they are not, except when there
is some source of disturbance to alarm them. I'he Ixire theory
is particularly difficult to justify on a practical basis,
having in mind the diversity of foods upon which the differ-
ent species of rattlers subsist, including a great variety
of mammals, birds, and lizards. It is hard to conceive what
kind of sound would act as a decoy, a lure by imitation, or
a charm, for so diversified a lot of creatures.

The experiments of Manning, 1923, tend to prove
that the rattlers' audient range does not permit their hear-
ing the sound of the rattle. This, if verified, would at
once eliminate the mating and call-for-help theories, or any
others depending on recognition of the sound by other rattle-
snakes.

I have noticed occasionally, when a number of nerv-
ous rattlers were kept in separate boxes, that excitement
upon the part of one sometimes started others rattling, but
the vibrations may have been transmitted through the floor
rather than the air. Or the artificial conditions incident
to the boxes may have caused an increase in audibility or a
change in frequency.

In any case, there is little evidence favoring the
call-for-help tiieory. On several occasions in the field I
have come upon two or three rattlers together. One, alarmed
at the intruder might rear up and rattle violently, yet the
others either did not hear the sound or appeared unimpressed
by it.

C. B. Perkins has observed that a rattler in cap-
tivity will sometimes rattle when a second snake approaches

10

food which the first has struck or is eating. This is a
rather natural extension of the warning, like the growl of a
dog, and shows that rattlers do not change the primary piir-
pose of the rattle when using it on their fellows. And cer-
tainly, if v/e concede that the rattle has at least a partial
use as a warning, we have difficulty, from a psychological
standpoint, in justifying the other uses. "It is hardly con-
ceivable that a creature of so low an order of intelligence
as the rattler, would Instinctively use the same reaction as
a threat to enemies and a mating call; or as a warning and
a lure for prey. Rattlers are equipped with the same sense
organs and special scents as are possessed by other snakes;
these are presumed to serve to bring the sexes together in
the mating season.* J. D. Mitchell, 1903, states that male
rattlers are aggressive in the mating season and will prompt-
ly rattle to challenge any intruder. This may have given
rise to the mating-call theory.

V»arning devices or mechanisms are relatively common
and of considerable diversity among snakes — for example, the
hiss of the gopher snake, the spreading hood of the cobra,
the scraping scales of the saw-scaled viper, and the inverted
tail-spiral of the ring-necked snakes. But the most wide-
spread is the vibration of the tail as a sign of alarm or
annoyance. This ancient and extensively used reaction is
generally interpreted as a warning mechanism; and there
seems no more reason to question the purpose of its more
highly specialized descendant, the rattle of the rattlesnake.
Although a harmless snake, vibrating its tail amongst dried
leaves, can simulate the noise of a rattler quite effective-
ly, this cannot be construed as imitative; the vibration of
the tail antedates the rattle and, besides, is widespread in
the Old V:orld, where rattlers do not occur.

It may be noted that at least one rattler (C.
ste.jnegeri) has a rattle so tiny that it must be virtually
inaudible. This rattler comes from a remote region in north-
western Mexico and I have had no opportijinity to observe a
live specimen, or to hear the rattle in use. In the adults
the rings are barely 3 mm. across as compared with over 20
mm. in some large species. The tail is proportionately much
longer than in other rattlers and is more tapered at the end.
It has been suggested that this rc^ttler is reaching the op-
posite end of evolution from the primitive creatures which
first warned the bison, and is now losing its rattle.

Does a rattler alv/ays rattle before striking? Cer-
tainly not; if a snake be stepped on he will turn and bite
with no coil, rattle, or any pause or warning. A snake close-
ly approached may be so suddenly alarmed that ne will lunpe
at the intruder with a simultaneous sound of the rattle; or
he may omit rattling entirely. But tiie threatening rattle,
as tne initial step, is the more usual procedure. Sometimes
a snake will endeavor to seek a nearby retreat, either facing
his enemy and with forebody raised in the striking pose, or
more precipitately by tne usual crawling method; in the
latter case he may stop rattling, thus abandoning his warning
threat for simple escape. If escape be denied he will alv/ays
coil and threaten; I have yet to see a rattler, not already
intimidated by captivity, which, when cornered, would not
turn and face the intruder, with rattle sounding furiously

* Noble, 1937

11

and with the forebody poised for a strike.

As has been stated, rattlers discovered in the
resting coll will sometimes sound the rattle without change
of position, until a nearer approach of the enemy causes
them to shift to the defensive or striking coll. Occasion-
ally they will give a characteristic warning click or two of
the rattle before sounding it continuously. If first found
in the ordinary crawling posture, a rattler will usually
draw itself up directly into the striking coll if the in-
truder is quite close. If it desires, a rattler can sound
the rattle while crawling.

To conclude: The threatening pose of the rattler
is aided by the rattle in two ways; the stridency of the
sound is in itself alarming; and attention is drawn by the
sound to a creature, whicn by every detail of deportment,
indicates the danger of nearer approach. There is very
positive evidence for the warning theory of use, but little
or no evidence for the others which have been proposed.

Character of the Sound

Many are the similes found in the literature for
the sound of the rattle: The buzzing of bumble bees or other
insects, peas shaken in a gourd, wh'eat rattled in the hand,
escaping steam, the cicada, running water, rustling leaves,
a child's rattle, an alarm clock, the folding of dried parch-
ment, the whirring of many spinning wheels, blowing through
loose lips, the rattling of seed pots, the wind in the trees,
the hissing of a hard rain. Of these probably the most ac-
curate likenesses to the noise made by a large rattler are
the sounds of certain species of cicada, and the hissing of
steam through a small jet.

The term "rattle" is to a certain degree a mis-
nomer when applied to the noise made by the snake; for
"rattle", by definition, implies discontinuous or discrete
sounds, whereas the separate sounds emanating from a snake's
rattle are much too close together to be distinguished by
the human ear. The result is a toneless* buzz, or, in the
case of zhe larger snakes, a loud hiss. The distance at
which it is audible depends on the kind and size of the
snake — from only a few feet in the case of S. miliarius to
200 feet or more with large specimens of such species as
C. adamant eus. C. durissus . or C. cinereous. When water-
soaked the rattle is practically inaudible. However, it is
not permanently affected by moisture, and recovers its
audibility as soon as dry.

Method of Vibration

As will be discussed hereafter, the rattle is
elliptical or subrectangular in cross-section, with the long-
er axis vertically perpendicular to the center line of the
body. In the crawling position the rattle is tipped up to
avoid wearing on the ground. When colled for defense, with
the rattle sounding, the tail and rattle are held in a posi-
tion approaching the vertical, or with a posterior slant.

* That is, dry and rasping, without hum, ring, or the quality
of metallic vibration. It is not toneless in lacking a sub-
stantially constant frequency.

12

In this position of almost perfect balance, it will deliver
the maximum noise for a given expenditure of energy and with
the least wear.

i'he vibration is transverse or lateral, that is,
in the direction of the shorter axis of the subrectangular
section, or horizontally at right angles to' the center line
of the tail.

Through the courtesy of Dr. Phillips Thomas and
Mr. A. E. Kitchener, of the Westinghouse Electric and Manu-
facturing Company, I was able to watch the operation of the
rattle under the light of a variable-frequency neon strobo-
scope. We had available a pair of C. tortugensis. with the
long rattle strings characteristic of this island species,
one of 13 and the other 11 rattles. It was most interesting
to see our previous assumptions concerning the nature of the
vibration borne out in this visual test. With the strobo-
scope light regulated to a frequency just below that of the
snake's tail, the waves could be clearly seen going out from
the tail to the end of the rattle string. The amplitude in-
creases as the wave moves outward, because of the cumulative
looseness of the interlock. The wave is much like that which
may be sent along a loose rope.

By moving the segments of a rattle slowly, with a
transverse motion similar to that seen under the stroboscopic
light, the method whereby the sound is produced may be ob-
served. Between each lobe of each segment, and the interlock-
ing lobe of the segment within, a succession of contacts and
separations is produced, especially on the deep groove sur-
faces. -No doubt each pair clashes at a number of different
points, so that several distinct sounds are produced, as the
wave passes each lobe. The total number of sounds per cycle
equals twice tne number of rings multiplied by the contacts
per ring. Assuming only one contact per ring, three lobes
per segment, and a 6-rattle string, at 50 cycles per second
(an approximate average number), we have 1000 contacts per
second. It is obvious that a frequency such as this will
result in a hiss rather than discrete sounds. A young rattler
v/ith only two rattles makes about 100 contacts per second;
the result is a much softer buzz than the hiss of an adult.
Siace the rattle is hollov/, dry, and the material somewhat
pliable, like heavy parchment, the sound is unmetallic; it
seems to have no musical quality.

In addition to the ordinary high-speed vibratory
motion, the snake, while rattling, often waves the tail slow-
ly from side to side.

Speed of Vibration

For the study of the vibration of the rattle, a
simple kymograph, driven by a synchronous motor on a con-
trolled-frequency circuit, was employed, with stylograph
paper on the drum. For registration a pin v;as run vertical-
ly through the proximal segment of the rattle, the point in-
serted downward between the rattle shell and the flesh of the
tail tip. The snake was held with the neck in the left hand
and the posterior part of the body in the right. The vibrat-
ing tail was then lowered until the pin touched the kymograph
drum. As it was later found that the speed of vibration was
characteristic of each snake, and its condition and disposi^
tion, rather than the weignt of the rattle string, it is be-

13

lieved that the pin weight did not affect the result. Sim-
ilarly, as tne scratch on the waxed surface v.-as usually so
fine as to require a magnifier for its examination, the drag
of the pin is not thought to have been a source of serious
error.

Altogether, 173 records were made on 31 snakes of
20 different species and subspecies. It was found possible
to get good records in most cases when large specimens not
long in captivity were used. The juveniles of the large
species, or adult snakes of the smaller forms, did not pro-
duce legible records. Specimens long in captivity seldom
reacted efficiently; and even fresh specimens often refused
to rattle properly, after being held for a few moments, in
spite of being vigorously pinched.

Although I suspect that there may be soue specific
differences in rattle speed — the longer tailed forms, such
as C. d. durissus. vibrating their tails at a slower rate
than C. V. oreganus. for example — I soon gave up any hope of
proving this statistically, for various other conditions,
such as the temperature and the degree of anger or fear
evidenced by the snake, make such wide variations as to mask
completely any species differences.

The kymograph records indicate that, for each
snake, under specific conditions, the rattle reaches full or
characteristic speed within a few cycles — there is no gradual
coming up to speed. The same conclusion is reached from ob-
serving snakes making the spasmodic or nervous clicks with
which they often begin or end a sound of the rattle; these
are really short runs of the rattle at full speed. However,
there are occasional variations within short ranges. For
example, a specimen of C. m. stephensi changed from a speed
of 54. cycles per second to 45 within two cycles.

The tail vibration seems closely to follow simple
harmonic motion. To amplify the kymograph record a needle
was run longitudinally down the center of the long rattle
string of a large snake, thus making it a rigid, but light,
extension of the tail. The end of the needle was allowed to
touch the kymograph drum and the resultant curves were photo-
graphically enlarged. The wave was analyzed as a Fourier
series by the twelve coordinate method, with the following
result:

y = 180.4 (x -t 92.58°) i 6.4 (3x + 61.8°)
+ 5.3 (5x - 176.7°) + 1.9 (7x +152.1°
+ 0.8 (9x 4 50.2°) + 0.8 (llx - 172.9°

This is not presumed to be highly accurate, owing to the un-
natural effect of the needle; but as the third and fifth
harmonics have an amplitude of only about three per cent of
the fundamental, and the higher harmonics one per cent or
less, the curve closely approaches a pure sine wave, and
simple harmonic motion is indicated.

I have stated that the wave, as seen under strobo-
scopic light, gains in amplitude as it travels out toward the
end of the rattle, because of the cumulative looseness of
articulation. Near the tip of the tail Itself the amplitude
is quite small — hardly more than a millimeter or so on each
side of the center line — even in a large snake. A young
C. adamanteus 830 mm. long, vibrating its tail at the rate

u

of 51 cycles per second, had an unusually large amplitude;
this was measured and found to be 3.4- nim. from right to left
peak, or a motion of about 1.7 mm. on each side of the center
line. This is certainly greater than the average motion,
which, in a snake a yard or so long, is found seldom to ex-
ceed 2 mm. betv/een extremes, and is usually between 1 mm. and
1-1/2 mm. But the rattle itself, particularly at the outer
end, has a far greater motion because of its looseness.

The proximal rattle, through which the motion is
transmitted to the remaining segments of the string, is
rather closely attached to the forming matrix subsequently
to be described. The terminal caudal bones are coosified
into a central reinforcement within the matrix; this forms
a short shaker (Fig. ^7) . The muscles producing the motion
are in the posterior part of the tail but anterior to the
matrix; there are six bundles of these, three on each side.
Viihile the amplitude of the motion is not great, the muscles
are powerful and are capable of maintaining the vibration at
full speed for a considerable time.

A study of kymograph records indicates that under
constant temperature conditions, if there be a change in
speed there is likely to be a change in amplitude as well;
a lower frequency accompanies a lower amplitude.

An elementary test is sufficient to prove the im-
portant effect of the temperature on the speed of vibration.
Eight runs on 5 specimens of C. v. viridis from South Dakota
at a temperature of 52° F. produced an average rattle speed
of 4-1-0 cycles, per second (max. 44«9, min. 28.1); ten tests
on 6 specimens at 73° F. an average speed of 57.9 cycles
(max. 66.9, min. 53.2). Thus an increase of 21° F. raised
the speed about 4.1 per cent.

The weight of the rattle seems to have no important
bearing on the speed, as shown by the following tests in which
segments were successively removed from the strings:

Female Nebraska C. v. viridis 980 mm. long:

Rattles in String Speed. Cycles per Second

6 45.5
A 51.5

2 50.8
1 50.2

Female Nebraska C. v. viridis 820 mm. long:

Rattles in String Speed. Cycles per Second

7 47.6
5 48.3

3 4-6.0
1 ^V.S

Only in one case was there an increase in the speed
when rattles were removed, and this may have been the result
of an increase in temperature from exertion.

There seems to be no appreciable sexual dimorphism;
nor was I able to find any consistent trend of rattle speed
with size. Of course, it is to be remembered that no suc-

15

cessful tests were made on juveniles or adolescents; such a
study could be made accurately only with a calibrated vari-
able-speed stroboscope. Of the species tested, at temper-
atures of from 65° to 75° F. (C. d. durissus. C. basiliscus,
C. envo. C. m. molossus, C. cinereous. C. lucasensis. C.
ruber. C. exsul. C. v. viridis. C. v. abvssus. C. v. lutosus,
C. v. or eg anus. C. a. pvrrhus . C. s. stephensi. C. 1. lepidus
and C. 1. klauberi)it was seldom that the speed was above 60
or below 40 cycles per second, with an average of about 48
cycles.

The lowest speed recorded was 28.1 cycles at 52°,
and the highest 66.9 at 73° F., both in C. v. viridis. No
tests were made at temperatures above 78° F.

To determine the amount of fluctuation in speed in
a continuous sounding of the rattle, I measured the lengths
of successive cycles and secured the following: Total cycles
147; mean speed 46.09, maximum 52.6, minimum 41.7, inter-
quartile range 44.67 to 47.64 cycles per second; coefficient
of variation 4.82 per cent. Probably the true variations are
somewhat less than these figures since the method of measure-
ment tends to exaggerate the fluctuations. This test was
made on a specimen of C. v. viridis at 73 F.

I find nothing to justify the suggestion of Hopley
(1882, p. 312) that a snake changes the speed of the rattle
deliberately to vary his signals. Besides the important ef-
fect of temperature, the fact that a higher amplitude ac-
companies a higher speed indicates that a greater degree of
alarm or anger tends to speed up the rattle. These seem to
be the most important factors affecting the vibratory rate.

Annoyed snakes can rattle for very long stretches
without interruption. Perry, 1920, mentions a horridus
which rattled for one-half hour. I have never made any ac-
curate records of long runs, but I have no doubt they some-
times exceed this period. In the laboratory it is frequent-
ly necessary to remove live specimens, so annoying is the
continuous rattling of specially nervous snakes not yet ac-
customed to captivity.

Lengths of Rattle Strings

The lengths and completeness of rattle strings
constitute an interesting subject of investigation. However,
in gathering such data it is virtually impossible to secure
statistics exactly representative of field conditions, since
it is inevitable that some rattles will be broken in captur-
ing the specimens, and, unless great care be taken, in their
preservation as well. It is not satisfactory to utilize
specimens which have been the subject of frequent examina-
tion in a muse'om, for, when the rattles are softened by
preservative, they are particularly subject to breakage in
removal from the containers. Also, one must be careful not
to utilize series of specimens in which there has been any
conscious selection with respect to the length and character
of the rattle strings. For example, my ov.n collection of
C. ruber would give an entirely erroneous idea of the pro-
portion of unbroken rattle strings occurring in that species,
since I have deliberately saved specimens with unbroken
strings in order to accumulate certain dimensional data.
Snakes long in captivity are unsatisfactory as their subjec-
tion to frequent disturbance may cause rapid rattle wear,

16

and skin changing periods are likely to be abnormal.

In Tables 31 and 32 data are presented on two
large series which were carefully handled in both capture
and preservation, and which are believed to be unbiased
samples. Only adolescents and adults are included.

In statistics of this character it is necessary to
take the season of capture into account as well as the spe-
cies of snake. The Platteville series was collected about
their dens at the time of going into hibernation in autumn
or coming out in the spring. At this time it will be ob-
served that most of those with complete strings have 5 rat-
tles. However, had the collection been made in the summer,
7 rattles would have been the mode. By the following hiber-
nating date, were the strings complete, the same snakes would
have had 8 or 9 rattles; but, as is seen, few retain such
strings entire, for breakage is rapid after there are 8 in
the string. Thus, in the table, these older snakes fall
into the broken-string group.

tiimilarly in the lucasensls statistics the table
represents an early spring series. Had it been made in the
summer, 8 and 9 rattle strings would have dominated the com-
plete strings. By the following spring these would become
tens or elevens. But unbroken strings of this length are
unstable, so that when next spring does arrive, few, if any,
of the two and a naif year old snakes still have unbroken
strings, and the 7 rattles of the one and a half year old
snakes of one year later would still set the mode at this
figure.*

Although the season of collection affects the mode
of the comnlete strings, it is less likely to affect broken
strings. For, if a rattler gets L, new rattles per year, it
is obvious that the modes in complete strings at any season
will be four units apart. But breakage occurs irregularly
and once a string is broken the seasonal modes or peaks
should largely disappear.

That there is some sexual dimorphism in the reten-
tion of complete strings is evident from the tables. For
example, 30.5 per cent of the male viridis had complete
strings, while 4-3.1 per cent of the female strings were en-
tire. This difference, by the use of a four-fold contingency
table, is found to be significant. In the lucasensls the
corresponding percentages are 12.6 per cent for the males,
and 17.6 per cent for the females. I am of the opinion that
the higher breakage shov/n by the males results from the same
cause as the predominance of males in collections, namely,
their greater activity.

The lower percentage of unbroken strings in
lucasensls . as compared with viridis, is a climatic effect.
For lucasensls has a longer growing season and averages
about 7 rattles (if the string is complete) in the early
spring, as compared to 5 in viridis . But a 7-3egment string
has a lov/er chance of survival without breakage than a 5-
segment, other conditions being equal.

* See Occ. Pap. J^o. 3, fig. 3, p. 16, for the correlation of
age with rattle segments in a typical species.

17

TABLE 31
PLATTEVILLE SERIES OF C. v. viridls

Number of

c

omplet
String

e

jbroken
Strings

Total

Rattles

M

I

"^

M

F

1

M

I

I

0

2

3

5

2

3

5

1

1

6

7

1

6

7

2

6

9

15

6

9

15

3

1

1

2

15

17

32

16

18

34

A

10

21

31

53

33

86

63

54

117

5

69

69

13S

53

42

95

122

111

233

6

8

20

28

46

37

83

54

57

111

7

9

12

21

A3

12

55

52

24

76

8

A

2

6

8

5

13

12

7

19

9

2

1

3

2

1

3

10

1

1

1

1

Total

101 125 226 230 165 395 331 290 621

(M, male; F, female; T, sexes combined)

18

TABLE 32

)

CAPE SERItS

OF

c.

lucasensis

Number of

Complete
strings

Broken
Strings

Total

Rattles

M F

1

u

F

T"

M

F

T

0

1

1

1

1

1

2

2

3

5

2

3

5

3

2

3

5

2

3

5

A

3

4

7

3

4

7

5

1 3

Ir

10

9

lo

11

12

23

6

6 6

12

15

21

36

21

27

48

7

9 7

16

31

17

48

40

24

64

8

6 5

11

38

17

55

44

22

66

9

1

1

28

14

42

28

15

43

10

2

2

22

7

29

24

7

31

11

15

4

19

15

4

19

12

1

1

1

1

13

2

2

2

2

14

1

1

1

1

Total 24 22 46 167 103 270 191 125 316
CM, male; F, female; T, sexes combined)

19

At this point mention should be made of the fact
that the button, being the thinnest and most fragile of the
rattles, and having only one gripping edge, is the most
easily lost (Table 33). The subsequent rattles are success-
ively strengthened, both by thicker and better reinforced
walls; and, because, beginning v^-ith rattle No. 3 or No. 4.,
the second transverse groove is serviceable in gripping the
succeeding rattle, so that the hold is doubled. At about
rattle No. i almost full adult strength and holding power
are reached (see p. 41). This fragility of the early rattles,
especially the button and Nos. 2 and 3, makes it clear why
extra long strings — say above 11 rattles — are so seldom un-
broken strings.

While there Is sexual dimorphism in the percentage
of specimens having complete strings, there is no significant
difference in the average length of the complete strings, the
following being the figures:

Average Number of Rattles. Complete Strings

Males Females Both Sexes

Platteville viridis 5.25 5.22 5.2-4

Cape lucasensis 7.16 6.78 6.98

Corresponding figures for the broken strings are
as follows:

Average Number of Rattles. Broken Strings

Males Females Both Sexes

Platteville viridis 5.24 5.04 5.15

Cape lucasensis 7.91 7.27 7.67

Hence, while more females retain complete strings,
they do not average more rattles; this indicates slower
female growth, or fewer skin changes at a given age.

It will be observed that there is little difference
between broken and unbroken strings as far as the average
number of rattles is concerned. I would expect to find a
somewhat higher average in the broken strings because of the
greater strength in the terminal rattles. There is some in-
dication of this in lucasensis. It must be recognized that
the broken strings are penalized by having among their number
several individuals with from 0 to 3 rattles; none with
these low numbers is contained in the complete-string cate-
gory, since they would be Juveniles and therefore eliminated
from our tabulation. If we begi« with 4-rattle strings then,
as expected, the broken strings have a higher average number
of segments than the unbroken.

One point does stand out in the figures which have
been presented — namely, the definitely higher average of the
lucasensis strings as compared with viridis . Differences
such as this may be attributed to a variety of reasons: the
comparative dispositions of the snakes; the prevalence of
enemies causing the use of the rattle; the character of the

20

country inhabited; and, finally, the relative strengths of
the strings.

An example of the first differential character may
be observed in C. ruber and C. v. or eg anus in San Diego
County. In a given period, 21 of the former were secured
with complete strings having in excess of 8 rattles; of
oreganus there were only 12. The total number of adults was
about the same in the two species. Ruber is probably the
most placid of all rattlesnakes; oreganus . on the other
hand, is of a rather nervous temperament. As these two
snakes live in the same surroundings, characterized by boul-
ders and chaparral, it is to be pres\imed that disposition
accounts, at least in part, for the differences in the lengths
of the rattle strings.

As examples of the second effect — the prevalence of
enemies — the island species tortugensis and exsul may be
cited. The former especially is characterized by long rattle
strings. Out of 270 lucasensis with broken strings, only L,,
or 1.5 per cent, had more than 11 rattles; out of 37 tortu-
gensis. 7, or 18.9 per cent, had more than this number, in-
cluding one with 15, one of the longest natural strings I
have seen. Similarly, of the available specimens of exsul.
8,7 per cent had more than 11 rattles.

With respect to the character of the country, it
has been noted, as might be expected, that snakes which in-
habit sandy areas, such as cinereous . in some parts of its
range, and cerastes . are often characterized by long strings.
For they are less subject to wear and breakage, than the rat-
tles of the brush and rock dwellers; the latter not only
siiffer from abrasion, but are more likely to have their rat-
tles caught in a forked branch or the cleft of a rock.

Finally, we have the matter of relative strength.
I think it probable that the difference between the average
length of viridis and lucasensis strings, to which attention
has already been directed, is largely due to this effect. It
is true, of course, that the larger rattles are subject to
greater amplitude of vibration and more 'friction, but they
seem to be even more than compensated for these requirements
by added thickness and strength. The smaller species, such
as S. mlliarius . S. catenatus . C. lepidus . and C. t. pricei.
particularly the first named, rarely have strings, either
complete or broken, exceeding 7 or 8 rattles. In general,
where the other variables are equal, larger species of rat-
tlers tend to have longer strings. Small rattlesnake species
with proportionately large rattles, of which tigrls is the
best example, often have long strings; for, in these, there
is an abundance of structural strength, compared to the forces
causing breakage.

It is of interest to note that in the Platteville
series, several of the effects which I have mentioned mili-
tate against long strings. In this series of 621 snakes
(adolescents and adults) only a single snake reached ten
rattles; this was a broken string on a male snake; there
were three snakes with nine. The longest complete strings
contained only eight rattles; of these there were six, four
males and two females. The correlation of body length with
number of rattles, which I hope to discuss in a subsequent
paper, indicates that many of these adults would have had
strings far exceeding these in number, had they remained un-

21

broken; in other words, it was not a size limitation that
cut the long strings to such small numbers— rather, it was
the natural wear and breakage.

Very short strings are also abnormal; thus out of
621 Platteville snakes (no juveniles included) five had no
rattles, and seven only one. these were noiseless and there-
fore useless strings. The five unfortunates had had their
rattles (including the matrices) cut completely off; they
showed by other scars that they had been "killed", and their
rattles removed as trophies, but had subsequently recovered.

I have never made a particular effort to secure
data on record-breaking strings, and the entirely fantastic
figures whicli have appeared in the popular literature on the
subject need little comment, for it is well known that
strings can be quite easily pieced together to deceive the
unwary. I can only say that, out of some ten thousand rat-
tlesnakes which constitute the bases of these notes, only
one (an island lucasensis) had as many as sixteen rattles,
and only two others (a cinereous and a tortugensis) fifteen.*
The longest unbroken string was thirteen; this was on a
Cape mitchelli. Of course, it must be understood that these
data are partly the result of studies of preserved specimens,
which, in the handling they received, frequently lose rat-
tles. But in any case, strings longer than these, although
they no doubt occur, must be highly exceptional.

Often, when too many rattlers are crowded together
in a shipping container, their tails and rattles will become
so tangled together enroute, that they can be separated with
the greatest difficulty, and only then at the expense of much
breakage. ViTiether anything of this kind ever happens in the
dens during hibernation I do not know.

However we may regret, from the standpoint of a
statistical investigation, the fact that so few long and com-
plete strings of rattles are at hand, we must admit that
from the viewpoint of the owner, breakage is both normal and
beneficial. Very long strings, say beyond 12 rings, cannot
be particularly serviceable from the standpoint of a sound
producing apparatus and might be a positive detriment, for
the very lengtn and weight would tend to damp out the vibra-
tions. Probably a string of from six to eight rings would
constitute the most efficient vibrator. Thus it seems normal
that rattles should be worn away, so that the long string is
exceotional. Both abrasion by extraneous objects and the
vibratory use of the rattle itself are effective in limiting
tne strings. In any case, it is to be remembered that string
averages in the field are certain to be somewhat higher than
those in collections because of inevitable breakage in cap-
ture, transportation, and preservation.

It has been noted that some snakes suffer consider-
able wear and breakage of rattles in captivity, particularly
if they retain enough animation to rattle at the approach of
strangers. Rattlers brought into a zoo with very long strings,
soon wear them out and lose many segments. On the other hand,
rattlers born in captivity, and therefore accustomed to man,
or others which are well treated and become inured to cap-

* This does not include records of the Wiley brood mentioned
on p. 30.

22

tivity, may develop very long strings v,'ithout breakage. This
indicates that the snake in the wild does not often have
cause to sound his rattle for long periods; in fact, in rat-
tlesnake country, one almost never hears a rattle sounded ex-
cept as the result of one's own intrusion uoon the privacy of
a snake.

A rattler, detained by holding the rattle, quickly
demonstrates how rings are lost in nature, should a rattle
be caught in the fissure of a rock or the fork of a twig;
for, under such circumstances, the snake will give a violent
twitch of the tail and the offending rattle will be snapped
off.

To the statements respecting some of the very long
strings mentioned in the literature, we must take definite
exception. A string of l^L, mentioned by an early traveler
has been cited unnumbered times by subsequent writers. A
string of this length would be broken by kinking or merely
by an attempt to pull it out from under the snake's own body.
But it may be presumed that this string will persist in
future natural histories at least until the last hoop snake
has rolled down the hill, to stab, with its venomous tail,
the defenseless oak tree at the bottom. The indestrucitibil-
ity of folk lore was never better illustrated than in tales
regarding the rattle. The story that the rattle is sounded
involuntarily, whenever the snake moves, was denied at least
as early as 17A7 (Mead), yet it still appears occasionally.
And that perennial favorite to the effect that the snake adds
one rattle to the string annually, giving an accurate record
of the snake's age (which would ignore breakage, even if the
annual addition were true) was denied by Lac^p'fede as early as
1789, and since then by innumerable other authors. Never-
theless, it is extensively believed today.

Chemical Composition

Chemical analyses made in duplicate, on rattles and
substances assumed to be somewhat similar in composition,
produced the following average results, all components being
expressed as percentages:

Rattlesnake

Cattle

Human

Rattle
4V.33

Skin

Horn
50.89

Hair

Carbon

48.17

48.14

Hydrogen

7.17

6.94

6.72

6.89

Sulphur

1.35

1.13

3.17

4.12

Nitrogen

U.13

13.08

16.66

16.06

Ash

2.53

9.12

0.40

0.74

Difference

(Oxygen,

etc.)

27.49

21.56

22.16

24.05

All determinations were made on a moisture-free

basis.

The high ash content of the skin is notable; it
was found to contain iron, calcium, magnesium, aluminum,
manganese, potassium, sodium, phosphates, sulphates, chlor-
ides, and silica.

It will be observed that the gross chemical com-
position of the rattle is not greatly different from that of
cattle horn, or human hair. Proportijnately the greatest

23

.PROXIMAL RATTLE

Fi^. 47

Fii. 51

Fii. 48

.WIDTH OF RATTLE

Fig. 49

terminal rattle
(not original button)

Fii. 52

Fii. 50

cEconniE

Fii. 53

Fii. 54

a

.47

II

48

II

49

n

50

II

51

11

52

n

53

II

54

The shaker within the matrix

The prebutton

The button

The button and No. 2 rattle

Rattle terminoloiu -- complete strini

Rattle terminoloiy -- broken strini
after loss of several seiments

Trend in rattle irouth as shown in
early natural histories

Actual trend in rattle irouth (This loni
unbroken strini is shown to illustrate the
trend; actually for a snake to retain a
strini of this lenith unbroken would be
quite phenomenal ) .

2A

divergences are in the low sulphur and high ash content of
the rattle.

These analyses were made through the courtesy of
Mr. C. E. White, of San Diego.

Method of Rattle Formation - The Prebutton

Having early been convinced of the accuracy of the
theory that a rattlesnake secures a new rattle each time it
sheds its skin (although there has not been entire concur-
rence by herpetologists on this point) , I was at a loss to
understand why young rattlesnakes, which, like other snakes,
shed their skins soon after birth, v;ere not subsequently
possessed of two rattles, one to match the original skin
with which they were born, and the second, the newly ac-
quired skin. This, however, is easily explained, for it is
found that the young rattlesnake does not retain, but, on
the contrary, sheds the rattle with which it is born. This
birth-rattle will hereafter be referred to as the prebutton.
Uarman noted the difference in shape between the prebutton
vFig. 4,8), and the true button or first permanent rattle
(Fig. 49), but thought the one was transformed into the other
by a stretching, accompanied by a constriction at the outer
end and a swelling of the proximal edge, thus acquiring the
characteristic shape of the button. This is found to be con-
trary to what actually occurs; for, from observations of re-
cently born broods, it is found that when the birth-skin is
shed the prebutton goes with it. This loss occurs for two
reasons; first, the thinness and fragility of the prebut-
ton; and, secondly, because its configuration, and that of
the true button beneath, are such that it cannot retain a
grip on the latter. The crimped edge of the prebutton is
too pliable and weak to prevent opening, and furthermore,
the central transverse groove in the button is too shallov/,
so that there is little for the prebutton to grip. As a re-
sult, the prebutton is shed just as the skin itself is lost.
Usually it is shed as a part of the skin; in fact, the pre-
button is hardly distinguishable from the skin except upon
close examination. Its very softness may cause it to be
overlooked in the shed skin of the baby rattler.

The fact that baby rattlers shed shortly after
birth is verified by observation on broods born in captivity;
this is a program characteristic of most snakes, whether
oviparous or ovoviviparous. For example, a brood of 1-4
C. h. horridus were born on September 20, 1934-> between 8:30
and 10:30 in the morning. The eyes began to turn blue on
September 25th; this cleared up by October first and second.
The first snake shed on the night of October 3rd; most of
the youngsters shed on the 5th, and all by the 6th. Of a
brood of 11 C. ruber born on October l6th, about half shed on
October 26th, and the others on the following day. Three, in
another brood born August 10th, shed on August 21st. Three
C. v. oreganus born August 24., 1937, shed on September 3rd.
These observations were made by C. B. Perkins. A. M. Jackley
has advised me that he observed that C. v. viridis in South
Dakota shed in from seven to ten days after birth; one
brood, however, shed in three days. I think that nearly all
broods shed between one and two weeks after birth, thereby
changing from the prebutton to the button. Barton, 1800, and
Curran, 1935, mention juveniles born with two- or three-seg-
mented rattles. Amongst broods totaling several himdred
youngsters, I have never witnessed anything of this charac-

25

ter. Often the second segment (Fig. 50) is acquired shortly
before nibernation, and I have seen broods in which sor.e in-
dividuals had attained their second rattles within a month
to six weeks after birth. Thus in the San Patricio series
of 139 C. cinereous, born in late August and preserved one
month later, three specimens had already acquired their sec-
ond rattles. In the Platteville series of C. v. viridis,
collected while entering or leaving hibernation, out of 226
young of the year, 44, or 19.5 per cent, had already achieved
a second rattle, and one had three.

Although I think there is only one proper use for
the term "button", this being to indicate the first rattle
permanently retained after the skin with which it was once
coordinated has been shed (Fig. 49), the word has been used
in other senses. Hardly a rattler is ever reported in the
newspapers unless it is stated to have had "blank rattles
and a button". But here button usually means the terminal
lobe of the last rattle, even though the string may not be
complete, the true button and additional rattles having been
lost. Others use the term to indicate the proximal lobe, in
the sense that this "buttons" the string to the body. But
the term button was originally used, and prooerly, I think,
as the single, soundless rattle of the young rattler (after
the prebutton had been shed) , which is more button-like than
the later segments. As soon as the string is broken, and
one or more segments lost, which is quite early in the life
of most snakes, the rattler no longer has a button.

The successive rattles are usually of sufficiently
different shapes, particularly with respect to the transverse
grooves and angularity, so that it is quite easy, after a
little experience, to be able to distinguish between the pre-
button, the button, and succeeding elements of a string, if,
in the latter case, one or two rattles have been lost (Figs.
51 and 52) . C. cerastes has a peculiarly shaped button which
sometimes makes it rather difficult to ascertain whether it
is, or is not, present. The same is also true of C. durissus
and its relatives, which have smaller buttons than other rat-
tlers of corresponding size.

Interrelationship of Rattle and Skin;
l-^evelopment of a Rattle Segment

The second divergence from the usual descrintions
has to do with the relation between the skin of the rattle-
snake and the rattle itself, for these refer to the rattle
as a thickened and corrugated terminal section of the skin.
It has been stated that when the skin is shed it tears away
from this thickened and horny terminus, which is then re-
tained by the corrugations of the succeeding rattle.

While this no doubt explains the genesis of the
rattle, the mechanism of formation has become so highly dif-
ferentiated that its relation to the skin can only be traced
witn difficulty, and there may be some question as to what
element of tne skin constitutes the ancestral analogue of
the rattle. For, as at present perfected, the rattle is not
merely a thickened terminus, but results from a separate and
discontinuous deposition. The skin of the rattlesnake ac-
tually extends to the tio of the tail, covering not only all
exterior portions of the body, but the concealed and corru-
gated rattle-forming matrix as well. The rattle is formed,
not by the thickening of the skin of the matrix, but by the

26

deposition upon its surface, of an entirely separate, clay-
like substance, which thereafter hardens and becomes a
rattle.

To trace the possible origin of the rattle it v^ill
be useful to recall sorae features of the skin. It will be
found that the skin has a duplex character. Those parts
which cover scale surfaces are somewhat thickened, and, if
shed prematurely, often retain considerable pigment. Sur-
rounding these thicker sections there is a network of thin-
ner elements, whicn comprise the folds of the skin betv/een
the scales — that is, they cover the interstices belov.- the
imbricatians. As the skin is shed the folds are flattened
out; thus the skin takes on a larger area and is easily re-
moved. If a snake be preserved a short time prior to shed-
ding and is subsequently dried, the scale tops will often
come off in the form of evenly-edged and separate flakes.
But if the skin is naturally shed the scale tops are like
clear lenses in the general texture of the continuous skin;
they are less pliable than the rest of the skin and are less
easily torn. Evidently, in the changes which take place in
a skin prior to its being shed, the scale tops are replaced
first, and may be separated from the skin, of which they form
a part, by the liquid of the preservative before the thinner
skin covering the interstices is ready for replacement.

It is believed that the rattle is analogous to a
scale top, at least as far as its original derivation is con-
cerned; and just as a scale top seems to be superimposed on
a general skin, from whicn at certain stages it can be sep-
arated, so also the rattle is separable from the skin of the
matrix upon which it has been deposited. This may be veri-
fied by a' study of the prebutton, which is intermediate in
character between the body skin and the perfected rattle.

The development of the rattle can be best described
by assuming that we are watching the formation of a new seg-
ment, as it v/ould be seen if v/e removed the previously formed
segment; for the latter ordinarily conceals, as a protective
roof, the posterior corrugations of the matrix upon which the
new rattle is to be formed. We shall consider in this out-
line a typical adult rattle having three lobes. The last two
would be disclosed by the removal of the prior rattle; the
anterior of the three will be formed under the tail skin
which is next to be shed.

V(e picture, then, the rattle-forming matrix with
its three transverse corrugations. It is covered with a
thin skin, which is apparently a continuation of the skin
which will be the snake's exterior covering after tne next
shedding. Uoon the surface of this skin the rattle is now
deposited, in the form of a laminated grayish, opaque, and
clay-like material, presumably extruded from glands within
the matrix. I'his material remains soft, and, if picked off
with a sharp instrument, comes away from the skin in irregu-
lar flakes. Tne laminations are easily seen and the separa-
tion from the skin below is quite clear. The anterior or
proximal edge of the soliciifying rattle is entirely definite
and even; there is no confusion witti, or adhesion to, the
new skin , such as would be expected if the rattle were simply
a tnickened patch of skin.

The grooves in tiie matrix are not rounded in out-
line; rather, tlie corrugations have a shrunken and angular

27

appearance; to these the soft and newly formed rattle close-
ly adheres.

When the deposition of the rattle has been com-
pleted, the material hardens. It becomes thinner, denser,
and quite transparent. The skin below acts as a parting
strip, or separating surface; and when the hardening is
complete, the rattle springs away from the deeply indented
and angular matrix beneath, and assumes the more rounded out-
lines which characterize the finished rattle. There is evi-
dence that the hardening progresses by laminations; at
least, specimens are found with a transparent surface layer
like cellophane, while below there still remains a thin
sheet of the amorphous clay-like material.

The shape of the new rattle is determined by the
matrix, not by the posterior or older rattle acting as a
mold, as some authors have suggested. Such would be im-
possible, for no mold would be available for the anterior
lobe. There is always considerable clearance between a new-
ly formed rattle and its predecessor, until the newcomer
springs away from the matrix upon completiou.

When the proper time arrives and the new rattle is
completely hardened and finished, the old skin is shed, thus
uncovering for the first time the anterior lobe of the new
rattle segment. The two posterior corrugations are, of
course, covered by the anterior and middle corrugations of
the previous rattle, and thus they are not disclosed. (This
is the normal condition; in our description we had assumed
the previous rattle removed, so that the formation of the
new segment might be observed.)

One point in this process remains to be described,
namely, the relation of the new skin, which now constitutes
the exterior covering of the snake, and the new rattle. As
previously stated, this skin is not attached to the new rat-
tle. On the contrary, it is first lapped, in a deep fold,
over the outside of the anterior edge, or bead, of the rat-
tle, and thence passes underneath, where it is apparently
continuous over the matrix, in the form of a soft white exu-
dation, which is probably the analogue of the skin in the
interstices between scale tops. Sometime during the life of
the new rattle this soft material must dry out and disappear,
either as broken flakes or otherwise. It cannot be reab-
sorbed into the matrix since it is found that the next rattle
has begun to form before it has disappeared. This soft
material is not only a possible analogue of the skin, but it
is also related to the transfer of the lobes of the matrix,
subsequently to be described.

The prebutton suggests a transition stage between
the ancestral skin and the final rattle mechanism, thus
serving to indicate the analogous elements. Here the true
rattle is so thin, compared with the skin itself, that it is
carried away with the skin; it is found to be but a thin,
transparent capsule over the skin. However, while the skin
is continuous over the body and to the tip of the tail, the
prebutton covers only the rattle portion. Thus the prebut-
ton seems to be an intermediate stage between the transparent
skin, which covers the scale tops, and the rattle; the skin
below, from which the prebutton may be readily separated,
corresponds to the lower layer of skin over the body, of
which it is seen to be a continuation. No doubt in the pre-

28

button we have a glimpse of an earlier stage In the evolu-
tion of the rattle. Also, we have a proof that the skin
changes on the matrix just as it does over the rest of the
body surface.

The fold of the skin over the bead, or outwardly-
curved, anterior edge of the rattle, is important, since by
stretching, as the rattle slips back, it makes available the
additional space under which the anterior lobe of the next
succeeding rattle is produced. In any snake (not a rattler).
Just prior to shedding its skin, each scale and fold of the
old skin exactly corresponds to the same scale and fold of
the new skin below. This is not true of the tail of the
rattlesnake. Just anterior to the rattle, for a short time
before a skin is shed, the completion of the proximal lobe
of the new rattle and the slipping back of the old, causes
the old skin to loosen and stretch, and become the protect-
ive covering of the growing lobe of the new rattle. This
stretching is permitted by the fold over the bead previous-
ly mentioned; and after the stretching has taken place, it
will naturally be found that, toward the end of the tail,
the scales of the old skin do not correspond in position with
the scales of the new skin, but have been longitudinally
displaced, relative thereto, for as much as a centimeter or
more in the larger snakes. During the stretching process the
end of the old skin is held in place by being caught in the
folded edge of the rattle which is slipping back.

Coincidence of Shedding with Rattle Formation

ITfiile most observers have agreed that the forma-
tion, or at least the disclosure of each new rattle, is co-
ordinated with' each change of skin, there have been some
dissenters, notablv Hopley, 1382, and Feoktistow, 1839
(translation, 189l) . Others have doubts, as for instance,
Storer and Wilson, 1932. Curran and Kauffeld, 1937, express
the opinion, said to be verified by laboratory observation,
that, while no new segment is added without the shedding of
the skin, the skin may be shed without producing a new rat-
tle. But it seems to me evident that these conflicting ob-
servations were largely made on captive or defective speci-
mens;* and I have already pointed out how seriously the
natural processes of these reptiles are interfered with by
the artificial conditions of captivity, if they are not
given great care, or if they refuse to eat naturally. Here
we see them sluggish, a prey to starvation and disease, their
skins shed at irregular intervals, and often in strings and
patches rather than entire as in a healthy snake. Small
wonder then, that the observer is unable to note v.'hether a
fresh rattle is disclosed coincident with each shedding, or
whether the total rings acquired in a given period equal the
number of moults in the same time. Even where the skin is
shed entire, the posterior end may pull away from the pre-
vious rattle and disclose the new segment several days or
weeks before it is shed. Under such circumstances, unless
the rattles are carefully counted, an addition will not be
noted as accompanying the skin change. Also, it is possible
that a segment may be lost at about the time the new rattle
is disclosed, and this would not be observed unless each seg-
ment were marked or numbered.

* Hopley' s was obviously a deformed freak.

29

Those who have postulated an independence of shed-
ding and rattle acquisitions, have failed to note the con-
tinuity of the skin, over the visible part of the body, with
that covering the hidden matrix. The anterior ring of the
nascent segment is always developed under the protection of
the posterior end of the visible skin. Were the two unre-
lated chrDnologically, the new rattle would either have to be
developed in tne open (which would give it no protection in
the lamellate stage), or the skin of the tail would have to
break away from the crimped selvage of the old rattle and re-
grip the new. That the dead skin can do this is obviously
impossible. It would seer.i equally impossible for a new skin
to be formed over the crimped selvage of an old rattle.

But in healthy snakes we know from a study of large
numbers of specimens in all stages of the process, that the
shedding of each skin uncovers the anterior lobe of a new
rattle; the fact that each skin really extends to the tip
of the tail and, therefore, every pair of rattles is sep-
arated by a skin, allows of no other interpretation than that
each exuviation means a new rattle.

As examples of the distortion of the life cycle
which sometimes occurs in captivity, I might mention a C. v,
oreganus v/hich was kept alive some tlriree years. At the end
of that time it had a complete string of ten rattles, but the
length of the snake was only about 60 per cent of normal.
Another even more striking instance, in which prenatal in-
fluences may also have been effective, is presented by a
C. V. viridis . born to a captive specimen September 21,, 1932.
This young rattler died a year later on October 2, 1933. At
that time it was oossessed of an unbroken string of nine
rattles where, had it led a normal existence, it would have
had five, thus showing the effect of the avoidance of a
winter hibernation. But, far from profiting by this increased
activity, at death it had a length only half that of a normal
snake and the final rattle had but 56 per cent of the width
of a normal nintn ring .

Mrs. Grace Olive V.'iley was kind enough to show me
one of her remarkable captive-bred litters of rattlers. This
was a brood of 9 C. cinereous, born June 17, 1932; I saw
them on September 20, 193-4. Four of these had complete rat-
tle strings, 3 with 15, and one with 1^^ segments; one had
lost only the button and had 15 rattles left. The others had
broken strings. The remarkable features of this group are
the high number of rattles at the age of 27 months, no doubt
due to forced feeding and the absence of hibernation; and
the large proportion of complete or almost complete strings,
resulting from the snakes being undisturbed. But the rattles
were not normally proportioned for cinereous . being smaller
than the usual size, showing either a more frequent ecdysis
than usual, or a less rapid body grov/tn. As 14 to 16 rattles
at the age of 27 months is no doubt greater than normal,
probably both causes distorted the rattle strings. Of course,
these are extreme instances of the irregularity of growth due
to captivity; other cases are cited in the literature, of
specimens bred, born, and grown in captivity, whicn more
nearly ap-oroach the normal. C. B. Perkins raised three cap-
tive-bred C. ruber whicn were born on August 10, 1937. On
February 17, 1940 (age 30 months) two had coraolete strings;
one 9, and the other 10. These are well within the limits
of the number of segments that wild rattlers of tne same age
would ordinarily have. Also, it may be noted tnat a record

30

was kept of the shedding of these snakes; the nurnter was
found to correspond to the niomber of rattles. However, it
is only too evident that the life cycle cannot be judged
with assurance from a study of captive specimens.

VVhether the rattle continuously increases in size
during the life of a snake, is closely involved with the
question whether adult snakes grow slowly, but continuously,
until their deaths, instead of reaching a definite growth-
limit rather early in life, as do mammals and birds. This
has been discussed in Part IV of this series (Occ. Pap. No.
3, p. 31, 1937) and will be mentioned again in the study of
rattle dimensions. Since the rattle size is determined by
the dimensions of the matrix, there is a high correlation be-
tween body size and rattle size during the early period of
rapid growth; this correlation is less evident in the time
of slower grovrth during the adult stage, since the size of
the matrix is affected by body nourishment more than by body
growth. Hence there is occasionally evident a cyclic varia-
tion in size, correlated with seasonal activity and the avail-
ability of food. And under extreme conditions, such as the
effects of brush fires on food supply, or self starvation in
captivity, there may be a marked diminution in rattle size
as new segments are added. Later, with a recurrence of nor-
mal conditions of nourishment, the dimensions may again re-
turn to normal.

The early pictures of rattles showed a continuous
growth and steady increment in the size of the successive
segments (Fig. 53). Such Pictures of unbroken strings of 32
and 37 rattles will be found in Seba, 173A, plate 95; and
Shaw, 1802, plate 91, shows one of uX segments. As a matter
of fact, the growth is rapid up to and including rattle No.
6, and then continues at a reduced rate up to rattle No. 12,
or L4 (Fig. 54), declining more rapidly in females than males,
Beyond this it is impossible to tell whether there are any
further increments, for, if present, they are so small as to
be completely masked by seasonal and other variations, as pre-
viously mentioned. At any rate, beyond rattle No. 12, the
rings are of substantially equal size, and the sides of the
string are parallel. There are certain indications that
shedding (and therefore the acquisition of new rattles) is
less frequent after full adult size is attained; however,
this is a matter requiring further study.

The Rattle-Shifting Mechanism

V.'e nov; come to the third element in the rattle
formation, which, it seems to me, has not been adequately ex-
plained. It has been merely stated that' when the time comes
for the formation of the new rattle, the old segment slips
backC one corrugation, thus leaving the new segment a corru-
gation in advance. This, however, is a mechanical operation
of considerable intricacy requiriig furtner explanation. It
must be remembered that, after the completion of an adult
rattle, we have a ratiier rigid object having thiree lobes, or
corrugations, which surround three similar lobes of the rat-
tle-foriaing matrix below. There is not close adherence here,
for, as has been stated, the rattle has sprung away from the
matrix in the drying or setting process. Nevertheless the
lobes of the matrix are sufficiently lai;ge so that they can-
not slip forv.'ard through the constrictions which constitute
the transverse grooves of the rattle. Vie have, then, the
problem of effecting a relative longitudinal displaceiaoat of

31

one lobe between the matrix and the rattle; that is to say,
either the rattle must slip back one lobe, or, what consti-
tutes the same thing, the matrix must travel forward one lobe,
so that its middle lobe will engage the anterior lobe of the
rattle.

This might be effected in several ways. For ex-
ample, the lobes of the matrix might be decreased in size,
until each would pass through the constriction of the rattle
immediately in advance of it. However, if this were done, it
is evident that the rattle would no longer be held to the
tail; it would slip off the now-tapering matrix and be lost.

It might be possible that, as the matrix becomes
tapered, as above suggested, the skin would grip the bead of
the rattle, and, by stretching, permit the rattle to drop
back one groove, although preventing a backward passage be-
yond this point. But, this leaves much to be desired as a
safe mechanism, since a tearing of the relatively fragile
skin during the process would cause the loss of the latest
and all prior rattles. This would frequently result in a
situation seldom met in nature; for although snakes often
have incomplete strings, they are rarely found with a single
rattle, such as would be the case if they were collected im-
mediately following an accident such as has been described.

It seems to me that the only satisfactory explana-
tion of the relative movement between rattle and matrix,
v/hich will comply with the mechanical necessities of the
case, is to visualize a sort of wave action in the corruga-
tions of the material constituting the tissue of the lobes,
relative to the svirface skin and the bones beneath. One
might observe this effect by picturing tv;o tight rings on the
finger of a glove. The rings are the rattle constrictions;
the bulges between are the lobes. Assume some internal force
to cause the bulges to move toward the end of the finger.
The rings will be carried along, slipping over the glove sur-
face (the skin) and always separated by, and held in place
by, the bulges. In the same way, the matrix skin formerly
covering the anterior lobe moves forward, relative to the
rattle, until it occupies the location where the anterior
lobe of the next rattle will subsequently appear; and a
swelling begins at this point which constitutes the reappear-
ance of the anterior section of the matrix in its new posi-
tion. Yet this has been effected while still maintaining a
lobe of the matrix continuously within the anterior lobe of
the displaced rattle. So it is that the second matrix lobe,
now somewhat decreased in size by the exudation of some of
its substance in the form of a white, jelly-like material,
occupies a position below the anterior lobe of the elder rat-
tle, i^imilarly the terminal or third lobe of the matrix is
now below the middle corrugation; and the original final
lobe is now represented only by the soft white Jelly; it is
without structure, for bones and skin have moved forward.
Hence the matrix is in position for the form£.tion of the new
rattle, which will be properly interlocked with the old.
This process is aided by the clearance between the matrix and
the rattle, the latter in the hardening process having sprung
away, as previously mentioned.

The query may be made as to the necessity for so
involved a description of so simple a process, for the old
rattle has nierely slipped back one lobe. The point is, that
during this process, it is necessary continually to maintain,

32

under each rattle lobe except the last, a lobe of the rattle-
forming matrix, for otherwise the rattle would be lost in the
transfer; and the only means whereby this can be effected is
by the relative surface movement of the skin, coordinated
with a wave movement and exudation of the matrix tissue as
above described.

The process will be more clearly understood by
reference to Figs. 55 to 58, inclusive, showing several
transition stages; this is a schematic, rather than an ac-
curate dimensional presentation. Attention is particularly
directed to the movement of P, a single element of the matrix
surface; especially is it to be noted that the motion of P
is wavelike, for it occupies at various times both the bottom
of a corrugation and top of a lobe. Note also how the point
P' on the rattle, originally opposite P on the skin, gradual-
ly drifts away from it. And finally the reader may concen-
trate his vision on the vertically cross-hatched matrix in
the four figures in quick succession, and see how it seems to
crav.'l forward, while always retaining a bulge behind P', thus
preventing the escape of the old rattle, until the new one
has been formed below.

By observing a large number of rattlers, especially
some just before, during and after exuviation, we should find
all stages in this scheme of transfer, thus verifying the
method by a serial picture of the orocess. This has been
done by baring the matrix on a large number of snakes; the
surface, or completed rattles, were removed to disclose the
conditions below, both in the matrix and the incomplete rat-
tle in process of formation, and specimens were found rep-
resenting all stages in the longitudinal method of transfer
above outlined.

Probably the most striking illustration of the
transition is to be seen in the change which takes place in
the third, or posterior lobe, of the rattle about to be
slipped back. For here we have the tip of the matrix and
the end of the rattle itself, and, therefore, two points of
reference whose relative positions may be easily noted.

First we find the matrix completely filling this
terminal lobe; the rattle, newly formed, clings closely to
it, following each sharp angle. Next we find that the rattle
has sprung away from the matrix. Then the lobe of the matrix
seems to shrink; a space is seen between it and the rattle,
which fills with a soft, white substance of the consistency
of jelly. This movement continues; the lobe of the matrix
diminishes in size, the tip of the tail moves forward rela-
tive to the end of the rattle, and the space thus vacated is
filled with the amorphous jelly. Finally the posterior point
of the matrix has advanced through the constriction between
the third and second lobes of the rattle until it comes to
rest in the second lobe; and the posterior lobe is vacated
save for the jelly.

Vkhile this change has taken place in the third
lobe, a similar transformation is effected in the first and
second bulges, but here the change is not so evident. Vie
note only that the lobes on the matrix decrease in size and
move backwards, relative to the skin which covers them and
the vertebrae which constitute their core. As they shrink,
the space vacated partly fills with the jelly; it is partic-
ularly evident on their posterior shoulders, but does not

33

Fii. 56

Fii. 5?

Fii. 59

Fii. 5t

Fii. 60

Fii

a

.55~>:

n

56.

II

57.

n

58^

II

59.

n

60.

It

61.

II

62.

It

63.

61 Fii. 62 Fii. 63

Schematic diairam of rattle trans ferrence .

(P is a point on the surface of the matrix

and P' a point on the rattle. The dotted

material is soft and amorphous) .

Rattle strini with the outer end raised to
its hiihest position, shouini the effect of
the asymmetry on the possible travel

The same rattle strini as in Fii. 59 u)ith the
outer end in its lowest position
Vertical section shouini the asymmetry of the
internal fit of three interlocking seiments

Lateral asymmetry of a rattle seiment
Forward slope of the upper lobe

34

completely fill the space between rattle and matrix, tending
to adhere to the former. The longitudinal movement is now
complete; the new rattle can be deposited on the matrix in
its new position; and the jelly, evidently the analogue of
the skin which is about to be shed over the rest of the body,
disappears, probably by dessication. (It vias found by ex-
periment that the jelly, if left in the air, quickly dried
to a colorless film not greatly different from shed skin.
Its condition, as examined, may be somewhat affected by the
preservative.)

We judge, from the comparative prevalence of speci-
mens in these several stages, that the presence of a complete
rattle pver the matrix is a relatively stable condition,
while the slipping back followed by the deposition of the
new rattle is a transitional situation occupying a shorter
period of time. No doubt this phase of the process corres-
ponds to the relatively short period during which the eye of
a snake is dulled by an incipient exuviation.

V»e may state, in further verification of the
mechanics of transfer thus outlined, that the length of tail
of the rattlesnake is virtually unchanged in this intermit-
tent process; that is, the anterior edge of the rattle-
forming matrix (not the anterior edge of the last visible
rattle; is always at the same distance from the anus. The
number of subcaudals and the dorsal scales on the tail re-
mains unchanged.

Vve know, also, that the corresponding lobes of suc-
cessive rattles are always cast upon the same area of the
skin covering the matrix, for, if this be damaged, thereby
causing a defect in the rattle, the same defect will reap-
pear at a corresponding point of the equivalent lobe in each
succeeding rattle segment. For example, in a specimen of
C. ruber having a complete string of eleven rattles, there
was, on each anterior lobe beginning with the fifth, a pair
of longitudinal striations. These were found on no other
lobe of any segment. On the surface of the matrix itself
there was a sharp ridge corresponding to these striations
and duplicated nowhere else on the surface; this, obviously,
is the surface indication of the damage to the rattle-forming
mechanism which must have been sustained at the time the
fifth rattle was being formed, and which, ever since, has had
its effect on each succeeding rattle.

In another instance a specimen of C. cinereous had
a scaly imbrication, or flap, over a corner of each anterior
lobe; this was found to be due to a similar flap of skin-
covered tissue over the corresponding lobe of the matrix,
similar defects are comparatively easy to find in large
series of rattlers and always they are repeated at the same
point of each succeeding rattle.

Structure of the Rattle

The rattle is too irregular an object to permit a
universally satisfactory descriotion; however, certain gen-
eralizations may be made. It is knowTi, of course, that, as
the rattle is carried by the snake, the greater dimension of
each ring is vertical and the lesser transverse; that is,
the rattle is carried with the broader, flatter sides as

35

lateral faces.* In addition to the deep transverse grooves,
or constrictions, which separate the lobes of each rattle,
there is a secondary longitudinal furrow, soaetimes narrow
and sharply indented, more often broad and shallow, along the
lateral faces of the exterior lobes. The longitudinal groove
of each superior lobe loosely engages the groove in the in-
ferior lobe beneath; thus it affords a certain stability, by
guiding movement, and, by its corrugation, increases strength.
In the hidden lobes, the longitudinal grooves are deep and
sharply angular, so that the engagement of the groove of the
second lobe with that of the third is quite restrictive.
Often these internal furrows are so deep and compressed as to
produce internal reinforcing fins.

The device which permits the snake to carry his
rattles clear of the ground is the result of an asymmetrical
development above and below the longitudinal furrow. Nec-
essarily there is a looseness of fit, or lost motion, in the
coupling of the successive segments; this is required for
the use of the rattle as a sound producer. Were it con-
structed symmetrically about the longitudinal center line,
this looseness would result in a downward curve, which would
allov/ the posterior rattles to drag. But the rattles are not
symmetrical; on' the contrary, each ring is so distorted that,
when the rattle string is in a position permitting equal de-
grees of freedom both upward and downward, the center line is
a curve upbending posteriorly (Figs. 59 and 60) . V.Tien the
limit of travel doviTiward, as permitted by the successive
interlocks, is effected by the weight of the rattle, the
center line is virtually straight. By this means a slight
upv;ard tilt of the tail end suffices to keep the rattle clear
of the ground and this is the snake's normal crawling pos-
ture.

The method whereby the asymmetry is produced is
ingenious. Each flat side of a rattle is grooved by the
central longitudinal furrow. This is rounded, and may be
wide or narrow in the anterior surface lobe, but in the con-
cealed lobes it is deep and sharply cut. This furrov; serves
both as a reinforcement of what would otherwise be a weak,
flat surface, and as an interlocking guide between the mesh-
ing lobes. Thus each lobe is virtually divided into two
halves; looked at endwise the halves appear like intersect-
ing cylinders. So deeply cut by the side grooves is the in-
nermost lobe, that, from the end, the two cylinders seem
connected only by a thin vertical web. The important point
with reference to the asymmetry is that the upper half of
each exterior lobe is smaller than the lower half, while the
upper halves of the interior lobes are larger than their low-
er counterparts. Thus the fit in the upper half between the
exterior and interior lobes is closer than in the lower
halves, and the limit of travel in any direction is first
reached in the upper half (Fig. 6l) . This is further ac-
centuated by a longitudinal compression of the transverse
grooves and lobes in the upper half as compared v/ith the low-
er; each separate rattle is therefore curved slightly upward
as if it were a segment of a circle of large diameter (Fig.
62). In large specimens a third, but related, tyoe of asym-
metry is clearly evident; this is a forv»-ard slope of the

* Rather surprisingly this was once the subject of contro-
versy in the correspondence columns of a magazine devoted to
hunting and fishing

36

upper lobe, which is at an angle with the lower (Fig. 63).

The effect of the asymmetry can best be visualized
by the following simple experiment: First, hold a long rat-
tle string in the position as it is carried by the snake and
it will be observed that it is straight. Now turn it upside
down and the whole structure sags; if the. string be long
enough, the last links will hang vertically downward. An-
other method of showing that the mid-position of a rattle
string is a curve, is to press the two ends of a long string
toward each other so as to reduce the lost motion to a mini-
mum. It will be found that the string assumes a curve, with
the end tilted upward.

Not only is this asymmetry important in procuring
a safe carrying position, but, also, it has a beneficial ef-
fect in sounding the rattle. When the rattle is to be vi-
brated, the maximum effect will be produced, with the least
wear and consumption of energy, if the several clashing
elements are balanced on their center lines. Were the rat-
tle symmetrical, the tail would have to be vertical to se-
cure this balance, and the rattles would tend to fall in a
curve either forward or toward the rear. But with the dif-
ference in weight and size between the halves, an angle of
some eighty degrees with the horizontal will suffice to hold
the rattles in a straight line, and this is the angle at
which the rattler in the striking coil usually sounds his
rattle. Occasionally, however, it will be held farther from
the vertical, but still the asymmetry keeps the string
straight instead of drooping.

The articulation of the rattle is necessarily
loose, yet it is surprising how difficult it is to remove
successive rings, particularly adult segments with a full
complement of transverse grooves, so effective is the clinch
or interlock. The fracture and complete removal of an ex-
terior lobe at the center of a string will not cause the
loss of the rattles posterior thereto, as they are still
coupled by the smaller and stronger lobes within. So rat-
tles, unless cut or broken clear through, must be worn off
and lost in succession from the posterior end forward, and
each lobe of a segment must be completely fractured before
detachment finally occurs. Hence we see how these relative-
ly fragile members are so tenaciously retained.

As an indication of the looseness of the articula-
tion It may be noted that a string of adult oreeanus rattles
12 mm. in width and 4.2 mm. in length when adpressed, measur-
ed 4.8 mm. when fully extended, thus showing a longitudinal
lost motion of 6 mm. per ring, or 13.3 per cent of the nor-
mal string length of 45 mm. A string of 12 adamant eus rat-
tles 78 mm. long had a lost motion of 12.8 per cent; and a
set of 15 tortugensis rattles 86 mm. long, 8.3 per cent.
All of these strings were fully adult, that is, the posterior
rattles were of substantially the same size as the proximal.

In all rattles the interior lobes are more sharply
angular than the larger, outer lobe. I have hitherto men-
tioned the bead, or narrow outward cutl, along the lateral
anterior edge of each rattle. Usually this bead is doubled
back and flattened against the rattle to form a sort of hem
or selvage, a device beautifully designed to prevent the
start of a tear. Also, in the striking of this folded edge
with the shoulder of the next anterior rattle, we have one

37

GRIPPING
CLAW

REINFORCING

Pii. 66

Fii

64.

Fii. 65

D D

Fii. 68 Fii. 67

Fii. 73 Fii. 72 Fii. 71

EXAMPLE RATTLE SEGMENTS; UNBROKEN STRING FROM C.cinereous,

Fii. 61.

n

65

II

66

II

67

n

P8

II

«9

II

70

II

71

II

72

II

73

Vertical cross section of interlockini rattles
(Two are dotted to facilitate identl fication) .

Borizontal cross section of interlockini rattles
(Throuih the center of the lower lobe).

Vertical cross section at the bottom of the
anterior iroove, showini the reinforcini fins

Prebut ton

But ton

So. 2 rattle

No. 3 rattle

No. 6 rattle

No. 10 rati le

A four-lobe segment

38

of the contacts producing the noise, the bead being a rein-
forced wearing surface (point A, Fig. 65). The other prin-
cipal point of contact is between the heavily stiffened pos-
terior lobe and the transverse groove of the outer rattle
(ooint is). These contact points produce the maximum noise
with the least friction and wear.

Another Ingenious development is the reinforcement
of the transverse grooves between lobes by tiny ribs at the
four corners, for these grooves, as viewed in section, ap-
proach a rectangular, rather than elliptical, shape. Between
these ribs, at top and bottom, the interlocking clasp of the
first lobe is particularly effective, the grip being almost
claw-like; at the same time it acts as a hinge, permitting
unimpeded side movement in the direction taken by the vibrat-
ing rattle (Figs. 64 to 66, inclusive).

An outline of the changes in size and shape, in the
successive segments of a string, can best be exemplified by
a descriotion of the elements of a typical unbroken string.
A ten-rattle C. cinereous was selected for this study.

Nj. a (Prebutton, Fig. 67). (This is a hypothet-
ical description of this unit, since it disappeared with the
first shedding after birth. It is described from specimens
of recently born young.) This is flat and single-lobed. The
vertical cross-section is elliptical. The central longitud-
inal furrow is evident but not deeply indented. The material
is thin, flexible, and transparent, and the entire rattle
rather easily removable because It is only slightly constrict-
ed at the clinching end. It does not tear easily. From a
side view, the posterior end is semicircular in outline. The
thickness of material is about 0.015 mm.

No. 1 (Button, Fig. 68). There are two lobes only,
and these not sharoly differentiated; both are smooth and
rounded in outline. The anterior constriction is adequate
for holding and a slight bead is evident on the edge. The
groove between lobes, while distinct, is shallow, particular-
ly on the sides, and not clearly defined. The longitudinal
fixprow is continuous from end to end and is deeper posterior-
ly, thus pinching the end into upper and lower sections. A
vertical asymmetry is already in evidence, the transverse
groove being deepest above.

No. 2 (Fig. 69). Of this rattle the second (or
posterior) lobe is completely developed and the first trans-
verse groove sharply cut. The third lobe is in evidence, but
has no holding power. The bead on the first, or anterior
lobe, has become folded into a hem.

No. 3 (Fig. 70). There is little change from the
previous rattle except that the third lobe is better devel-
oped; it takes the form of two parallel horizontal cylinders
connected by a web. The vertical asymmetry is more in
evidence.

No. A. Of this rattle, the posterior (or third),
lobe is considerably enlarged by the outward divergence of
the two cylinders; it now has considerable holding pov;er,
which is of importance in retaining rattle No. 3. The re-
inforcing ribs at the corners of the transverse grooves are
clearly in evidence, but more so in the anterior groove be-
tween the first and second lobes. Viewed from the interior,
these ribs are seen to be of importance in strengthening the

39

Table 33
DIMENSIONS OF A RATTLE-STRING
C. cinereous

SECTION A-A,

Rattle
Number

W

^2

s

\

^2

N

\

\

\

^2

Wall
Thick-
ness

1

64

51

-

47

-

82

29

31

40

45

0.23

2

75

50

29

39

27

87

38

34

46

53

0.25

3

88

63

31

44

25

97

43

40

56

60

0.28

4

103

75

39

55

31

113

45

44

60

68

0.31

5

117

90

54

66

41

117

48

43

69

75

0.36

6

127

100

63

73

49

123

50

46

73

81

0.^3

7

133

106

73

77

56

122

48

43

76

84

0.36

8

139

116

78

85

60

120

49

44

75

83

0.43

9

UA

123

88

38

69

133

50

46

78

85

0.46

10

156

126

93

94

71

146

55

50

82

92

0.51

All dimensions in mm./lO

40

structure. The clinch of the first lobe Is particularly ef-
fective at the vertical ends, that is, at top and bottom,
comprising a veritable claw, which grips the groove between
the corner reinforcements of the second lobe of the No. 5
rattle.

No. 5. The third lobe Is now structurally of con-
siderable Importance. Viewed from the posterior, so that
each lobe appears as a small circle, concentric with its
larger anterior fellow, it can be seen that the anterior or
outer lobe has a larger lower than upper segment, while the
contrary is true in lobes two and three.

No. 6 (Fig. 71). The third lobe is still further
enlarged. The posterior surface of the middle lobe is deep-
ly indented.

No. 7. The transverse grooves are still deeper
and more sharply indented.

No. 8. There is little change beyond an enlarge-
ment of the parts.

No. 9. Now for the first time the divergent cyl-
inders which constitute the posterior lobe are extended into
little cones as if a fourth lobe were in contemplation. The
transverse grooves are deep set and angular, and the asym-
metry more than ever evident.

No. 10 (Fig. 72). This rattle, the proximal or
attached rattle of the present string, surrounds the matrix
from which it has only partly sprung away. The tentative
conical extensions from the rear of the posterior lobe have
become more prominent.

Dimensional data on this typical string are set
forth in Table 33.

Whether, had this string grown further, a complete
fourth lobe would have been developed, is not known. Four
full lobes are occasionally found ([Fig. 73), particularly in
the larger species, and sometimes the beginning of a fifth
has been observed. I have been unable to determine whether
the development of a fourth lobe is characteristic of a par-
ticular age, sex, or species — the indications are rather
against consistency on this point. I have seen some long
strings of twelve or more rattles in which the eldest showed
by its form that it was an adult ring, thus indicating that
the youngest was at least rattle No. 20, yet no fourth lobe
had begun to form. So it appears that three-lobe segments
are typical of most adult rattles. The second lobe is evi-
dent, but not serviceable, in the button; it becomes ef-
fective in rattle No. 2. The third lobe normally is evident
in rattle No. 2, and first becomes truly effective in rattle
No. 4' By the time rattle No. 7 is attained, the adult shape
is reached. Thereafter the increase in the size of succes-
sive segments is much slower; this increase continues at
least until the twelfth or fourteenth rattle, and possibly
beyond; however, after the twelfth it is so small as to be
masked by seasonal fluctuations.

Species Differences

The Intraspeclfic dimensional irregularity of the

41

rattle generally Inhibits its use as a key character, except
where the differences are considerable; for example, that
between S. millarius and S. catenatus. However, there are
definite specific and subspecific differences, not only in
rattle size proportional to body size, but also in the growth
equation. Data on this phase of the rattle are quite ex-
tensive and will be the subject of a subsequent report.

While specific differences in form and color are
obvious, they are not sufficiently striking or constant to be
always in evidence. It is, in fact, surprising to note the
morphological similarity of the rattle, in forms so widely
different in other characteristics as are some of the rattle-
snake species. The general scheme of the interlock, the sup-
port effected by vertical asymmetry, the longitudinal side
furrows, the bead or selvage, the fin-type reinforcements of
the grooves — all of these intricate perfections are found in
every species. The two incomplete lobes of the Juvenile
button are, in all forms, succeeded by the three fully inter-
locked lobes of the adult.

There are differences in the angularity and radii
of the curves of the lobes, the rattles of some species hav-
ing a more rounded appearance than others. This is particu-
larly characteristic of Sistrurus and the smaller species of
Crotalus . such as triseriatus. willardi. lepidus. cerastes,
and polYstictus. Of the larger species having this rounded
form, horridus is the most outstanding.

Some of the smaller snakes, especially polvstictus.
cerastes, and triseriatus. have smaller transverse dimensions
proportionate to the vertical (the ratio P2/W in Table 33)
than other snakes; also some, particularly cerastes and
triseriatus. are compressed longitudinally, i.e., the ratio
N/W is reduced.

When a rattle still surrounds the forming matrix,
it is practically transparent and therefore seems to have the
color of the skin below. However, when it has been cast off
it loses its transparency (although remaining translucent)
and through oxidation, or surface wear, assumes, in most
species, a characteristic straw color. Some, however, take
on a darker brown hue; this is particularly noticeable in
adamanteus. horridus . polvstictus. lepidus. tieris. willardi,
and all species of Sistrurus.

The color of the matrix (in reality a part of the
body pattern of the snake) is quite consistent and distinc-
tive. Usually it assumes the color of the final tail ring.
Thus, in the members of the cinereous and virldis groups, and
in such forms as horridus. polvstictus. molossus. and durissus.
it is usually black, especially anteriorly. Some forms are
partly colored; scutulatus is almost invariably black dor-
sally and light below, while cerastes has a black anterior
lobe and light posterior. Sistrurus and enyo are likely to
be mottled. Lepidus. triseriatus (including pricel) . and
willardi are pink, the latter with distinct black dots on
the matrix. Tigris has a grayish or pinkish matrix; this
character is so consistent as to constitute a key character,
in comparing with the subspecies of mitchellii. which have
black matrices .

Once having been slipped back, so that the matrix
is no longer visible within, it will be found that the color

of the rattle is entirely independent of the matrix upon
which it was originally formed; thus ruber rattles, formed
on a black matrix, are light, while tigris rattles are dark,
although the matrix is light gray or pink.

In discussing rattle and matrix colors, it should
be understood that only normal processes are considered. If
a skin about to be shed, be removed before its time, as fre-
quently happens In the course of preservation, the uncovered
rattle (or the matrix below) appears gray-blue in color;
this is the effect of the clay-like deposit which has not yet
been transformed into the transparent rattle substance, and
is entirely independent of the true matrix color. It is an
artificial condition and should be neglected in judging the
true color of the matrix.

Use of the Rattle bv Man

The rattle has not proved particularly useful to
man. It has occasionally been used as a decoration, es-
pecially with snakeskin belts and hat bands. There is some
sale for rattles in curio stores. Mr. C. B. Perkins tells
me they are placed in guitars in Mexico to make a sound ac-
companying that of the strings.

Most people, having killed a snake, carry the rat-
tles home as trophies. There have been accidents when the
rattles have been cut from the tail of a supposedly dead
snake.

Certain Indians are said to have held the belief
that the number of rattles on a snake indicated the number
of his victims. Other tribes used the rattle for medicinal
purposes.* Some ancient stone carvings incorporated the rat-
tle as a decorative motif.

Acknowledgments

I wish to voice my appreciation of many pertinent
suggestions from Mr. C. B. Perkins, and editorial supervision
by Mr. C. G. Abbott. Figures 4.8-50 and 55-58 were prepared
by Mr. L. C. Kobler, and 51-52 by Mr. Norman C. Bilderback.

Summary and Conclusions

1. The many purposes attributed to the rattle in
the past are listed, including the warning signal, mating
call, call for assistance, and lure for prey. That the rat-
tle is used as a warning cannot be questioned, as such use is
the normal and almost universal rattlesnake reaction observed
in the field. However, it is essential to differentiate be-
tween three types of warning: the warning of intruders,
possibly dangerous to the snake; an altruistic warning,
solely for the protection of the intruder; and the warning
of prey. A confusion of these has led to doubts respecting
the warning theory; when confined to the first — the warning
of possibly injurious enemies — it is logical and in conform-
ity with field experience. The warning is effective in re-
pelling enemies or unconscious intruders, both because of the

* Long's Expedition to the Rockies, Vol. 1, p. 236, 1823

A3

stridency of the rattle suid the other threatening actions of
the snake. The argument, sometimes used, that the rattle as
a warning is disadvantageous to the snake, by calling it to
the attention of enemies is not important . The rattle does
not sound unless the snake wills it; dependence is placed
on procrypsis, and the rattle is noc used until this fails.
While probably disadvantageous in the case of man, this is
not a valid argument against the warning theory, since the
rattle was developed long before the appearance of man in
the western hemisphere,

2. A specialization of the warning theory favored
by many herpetologists assumes the ancestral rattle to have
been developed as a protection against being tramped on by
large ungulates similar to the bison. Many genera of snakes
vibrate their tails in anger or as a threat. I'his habit, to-
gether with a gradual modification of the normal ophidian
tail-spine to conserve successive sheddings, would increase
the noise and aid in protecting the owner; for the bison,
Injured but not killed by a bite, would learn to avoid the
sound.

3. The warning serves to conserve venom and fangs,
also an advantage to the snake.

4.. Some of the other uses for the rattle which
have been suggested are inconsistent with rattler habits;
others are possible, but there is little field evidence to
validate them, and such as there is, may be subject to al-
ternative interpretations.

5. Experiments by one investigator indicate that
rattlers cannot hear the sound of the rattle; if verified,
these at once eliminate any theories such as the mating-call,
call-for-help, and other uses dependent on an exchange of
signals between snakes.

6. The rattle vibration is at too high a fre-
quency to permit recognition of discrete sounds. It is a
hiss rather than a rattle. Certain species of cicada or the
hiss of a small jet of steam most closely approximate the
sound of the rattle.

7. The rattle is operated by transverse vibrations
of an amplitude of a millimeter or so, produced by six muscles
in the tail. Viewed under stroboscopic light, the vibration
Is seen to have the effect of sending transverse waves out to
the end of the rattle string. Because of a cumulative loose-
ness of interlock, the waves enlarge as they progress out-
ward. Analysis shows the vibration to approximate simple
harmonic motion.

8. The sound is produced principally by the strik-
ing of the adjacent segments against the inner and outer
surfaces of the anterior transverse groove.

9. As determined by kymograph records, the speed
of vibration depends more on temperature than any other
single factor. There probably are species differences but
they have not been definitely determined. Frequencies of
from 23 to 67 cycles per second were noted. A speed of about
4.8 cycles is normal at about 65° F.

10. The loss of rattles to maintain the string at

from 5 to 9 segments is normal and beneficial, since very
long strings are inefficient vibrators. Strings above 14
segments are quite exceptional.

11. The more placid species tend to have longer
strings; the same is true of snakes which live in areas
where enemies are infrequent or absent, as" is the case with
certain Island forms. Snakes in rocky or brushy country
have shorter strings than those which live in sandy areas.
The larger species tend to have longer strings, on the aver-
age, than the smaller. The first rattles acquired are the
most fragile and hence the longest strings seldom are un-
broken strings.

12. Gross chemical analyses are given. The com-
position of the rattle is not greatly different from that
of the snake skin.

13. The prebutton, or rattle with which the snake
is born, is always lost with the first shedding, ^rtiich usual-
ly takes place within a week or two after birth. The pre-
button, by its structure, gives important clues to the gen-
esis of the rattle.

14.. The probable interrelation between the rattle
and the skin is discussed. While the rattle is analogous to
a scale top, it has widely differentiated therefrom. It is
laid down as a lamellate clay-like substance, on a basal
matrix of tissue which gives it its shape. Later it hardens,
becomes transparent, and springs away from the matrix. The
anterior edge is not Joined to the skin, from which it is
sharply differentiated. The rattle is deposited on the skin
of the matrix, which is coordinated with the skin of the body.

15. While some observers have doubted the synchro-
nism of rattle acquisitions with skin changes, it is thought
that they were misled by the erratic character of exuviation
in captive specimens. The nature of the process would seem
to indicate an absolute synchronism in normal growth. Prob-
ably most species add from three to five rattles per year.
This may be reduced when the adult size is attained and
growth is slower.

16. Increments in rattle size are rapid up to, and
including, rattle No. 6, but slower thereafter, although ap-
parent up to rattle No. 12 or 14 . Subsequently, individual
variations and seasonal fluctuations mask growth; the sides
of the string become substantially parallel.

17. The rattle-shifting mechanism, whereby the
first and second lobes of each rattle grip the second and
third lobes of its successor, without affording any oppor-
tunity for the string to drop off, is quite intricate. It is
effected through a wave action of the matrix, whereby the
lobes of tissue move backward (relative to the skin and verte-
brae) and gradually decrease in size by exudation. The same
element of matrix skin deposits a corresponding element of
each rattle, so that a defect, if one be present, always re-
appears on each rattle at the same point.

18. Structurally, the rattle is of beautiful and
intricate design, reinforced at possible points of weakness;
for example, there is a fold where a tear might start at the
anterior edge, the sides are furrowed for reinforcement, and

45

strengthening webs appear in the transverse grooves. Succes-
sive segments are separated with difficulty.

19. A longitudinal asymmetry is present and is im-
portant for two reasons: It prevents the end of the rattle
from dragging; and it holds the rattle in a straight line,
instead of permitting it to drop in a curve, when it is bal-
anced in the rattling position. The latter is at an angle

of about 10 degrees from the vertical; in this position of
balance the greatest efficiency — maximum sound with least
wear — is attained. The asymmetry is produced by a curve in
each rattle segment and a tighter fit, between interlocking
lobes, in the upper half of each segment as compared to the
lower.

20. A typical rattle is described. Three lobes
per segment are normal in the adult, but four lobes are oc-
casionally found.

21. Species differences in rattle color and lobe
conformation are discussed; also in matrix color, which is
really a part of body pattern. This paper omits discussion
of interspecific variations of rattle dimensions, a phase of
the investigation scheduled for subsequent publication.

46

BIBLIOGRAPHY

(The annotations do not purport to be complete sum-
maries of the references, but rather call attention to es-
pecially pertinent, or unusual, statements or experiences.)

Anon

1820 (circa) . The Natural History of Reptiles and

Serpents. Dublin, pp. VI + 178 (ref. p. 76).

1936 Poisonous Snakes of the United States. Bureau of
Biol. Surv., Wildlife Research and Management
Leaflet BS-70, pp. 1-19.

Gives mating-call theory (p. 13).

Audubon, John J.

1827 Notes on the Rattlesnake (Crotalus horrldus) ; in
a Letter Addressed to Thomas Stuart Trail.
Edinburgh New Philos. Journ., Vol. 3, pp. 21-31.
"The fact of their rattling, which tends to
alarm intruders, and to warn them of their
danger, is too well known to call for any
observations regarding it."

Aughey, Samuel

1873 The Rattle of the Rattlesnake. Am. Nat., Vol. 7,
pp. 85-86.

Cites experiences which, he feels, validate
the sex-call and call-for-help theories.

Barbour, Thomas

1922 Rattlesnakes and Spitting Snakes. Copeia, No. 106,
pp. 36-38.

The bison theory.

1926 Reptiles and Amphibians: Their Habits and Adapta-
tions. Boston and New York, pp. XX + 125 (ref.

pp. /^0-4.2) .

1934- Reptiles and Amphibians: Their Habits and Adapta-
tions. Revised edition. Boston and New York,
pp. XX + 129.

Method of rattle development is described.

An excellent statement of the bison theory

(pp. 41-43).

Barton, Benjamin S.

1794 A Memoir Concerning the Fascinating Faculty Which
Has Been Ascribed to the Rattle-snake and Other
American Serpents. Trans. Am. Philos. Soc,
Vol. 4, pp. 74-113.

1800 Supplements to a Memoir, etc. Separate, pp. 1-40.
While giving quotations citing the pre-
fascinating theory. Barton himself denies
its accuracy. He claims to have examined an
unborn young with 3 rattles.

47

Beal, W. J.

1872 The Music of the Rattlesnake. Am. Nat., Vol. 6,
p. 310.

The sovmd is like the singing of grass-
hoppers.

Beauvois, (P.) de

1799 Memoir on Amphibia. Trans. Am. Philos. Soc,
Vol. Ky pp. 362-381.

Rattlesnakes make no sound with the rattle
as they crawl along.

Berridge, W. S.

1922 Marvels of the Animal World. Boston, pp. 1-253.

1926 Marvels of Reptile Life. London, pp. 1-256.

Mentions the 4.4 often-cited rattle string.
Cites Darwin as against the warning theory
(pp. 55-58).

1935 All about Reptiles and Batrachians. London, pp.
1-270.

The use is still in doubt; it may be a sig-
nal to other rattlers or a sex call. The
warning of enemies is a doubtful theory as
this would lead to the destruction of the
snsJte (pp. 61-63).

Bingley, W.

1803 Animal Biography, Vol. 3, Amphibious Animals,
Fishes, Insects, Worms. London, pp. 1-580.
Makes the often-repeated statement that
Indians are afraid to go into the woods in
wet weather, since the warning rattle would
be soundless (p. 71) .

Boulenger, E. G.

1914. Reptiles and Batrachians. London, pp. XIV + 278
(ref. pp. 18^-186).

1927 A Naturalist at the Zoo. New York, pp. 1-206.

The rattle sounds like running water (pp.
174-175) .

Brainard, David

1855 On the Nature and Cure of the Bite of Serpents and
the Wounds of Poisoned Arrows. Ninth Ann. Rep.,
Smithsonian Inst., pp. 123-136.

The sound of the rattle is terrifying to most

animals (p. 124).

Brookes, R.

1763 The Natural History of Fishes and Serpents, etc.

(Vol. 3, of Brookes' Natural History). London,
pp. XXI + 408 + XII (ref. p. 368).

48

Bumpus, Hermon C.

1885 The Standard Natural History, Edited by John S.
Kingsley, Vol. 3, Lower Vertebrates. Boston,
pp. VII + 478.

In most particulars a very sound exposition
(pp. 397-398).

Burnett, W. I.

1854 Notes on the Rattlesnake, Proc. Boston Soc. Nat.
Hist., Vol. 4, pp. 311-315.

Carr, Archie F.

1940 A Contribution to the Herpetology of Florida.

Univ. of Florida Pub., Biol. Scl. Ser., Vol. 3,
No. 1, pp. 1-118.

The variation in the tendency of C.

adamant eus to rattle at intruders depends

on disposition (p. 95) .

Catesby, Mark

1731-43 The Natural History of Carolina, Florida and
the Bahama Islands, London, 2 Vols.

Questions the rattle-per-year theory (pp.
41-43) .

Cope, E. D.

1871 The Method of Creation of Organic Forms. Proc.
Am. Philos. Soc, pp. 89-123.

An increased use of the vibrating tail
would tend to increase nutrition, explain-
ing the development (p. 108). The rattle is
a disadvantage in the case of man (p. 120) .

1887 The Origin of the Fittest: Essays on Evolution.
New York, pp. XIX + 467 (ref. pp. 197-198).

1900 The Crocodilians, Lizards and Snakes of North

America. Rep. U. S. Nat. Mus. for 1898, pp.
153-1270 (ref. p. 1150).

Coues, Elliott and Yarrow, H. C.

1878 Notes on the Herpetology of Dakota and Montana.
Bull. U. S. Geol. Surv., Vol. 4» No. 1, pp.
259-291 (ref. pp. 263-264).

Cowles, Raymond B.

1938 Unusual Defense Postures Assumed by Rattlesnakes.
Copeia, No. 1 of 1938, pp. 13-16.

Describes the body-loop striking reaction of
rattlers, especially toward king snakes.

Curran, C. H.

1935 Rattlesnakes. Nat. Hist., Vol. 36, No. 4, pp.
331-340.

A rattle is not secured at every shedding.
Young are sometimes born with 2 rattles.

49

Curran, C. H. and Kauffeld, Carl

1937 Snakes and Their Ways. New York, pp. XVII + 285.
The skin may be shed without the addition
of a rattle, but a rattle cannot be formed
without shedding (pp. 47-53).

Czermak, Joh.

1852 Ziva, Jahrg. 1, No. 4, ?• 29.
Not seen.

1857 Ueber den Schallerzeugenden Apparat von Crotalus .
Zeit. f. wiss. Zool., Bd. 8, pp. 294--302.
An important early anatomical description
of the osteology of the shaker, its muscu-
lature, and the nature of the matrix. Pre-
sents, a theory of the rattle-transposing
mechanism.

Darwin, Charles

1871 The Origin of Species by Means of Natural Selec-

tion. Fifth edition. New York, pp. 1-447 (ref.
p. 196). (First edition, London, 1859).

1885 Same, sixth edition, London, pp. XXI + 458 (ref.
p. 162).

1877 The Expression of the Emotions in Man and Animals.
New York, pp. VII + 374 (ref. pp. 106-110).

Day, Walter de F.

1876 Music from a Snake's Tail. Forest and Stream,
Vol. 7, No. 5, p. 68.

Deering, T. W.

1872 The Rattle of the Rattlesnake. Am. Nat., Vol. 6,

pp. 490-491.

The rattle is more essential to the young to
provide food and protection than to adults,
yet they have fewer rattles.

DeKay, James E.

1842 Zoology of New-York. Part 3, Reptiles and Amphibia.
Vol. 1, pp. VII + 98; Vol. 2, plates Nos. 1-23.
"They (the Indians) assured me that whenever
a rattlesnake sprung his rattle it was a sign
that he himself was alarmed" (pp. 56-57).

Dltmars, Raymond L.

1907 The Reptile Book. New York, pp. XXXII + 472 (ref.
pp. 429-432).

1931 Snakes of the World. New York, pp. XI + 207 (ref.
pp. 114-115).

1936 The Reptiles of North America. Garden City, pp.
XVI + 476.

The use remains unknown (p. 337).

50

Ditmars, Raymond L. (cont.)

1939 A Field Book of North American Snakes. New York,
pp. XII + 305 (ref. p. 262).

Dudley, Paul

1723 An Account of the Rattlesnake. Philos . Trans.,
Vol. 32, No. 376, pp. 292-295.

"One other Circumstance of their rattling
has been observ'd, to wit, that if a single
Snake be surprised and rattles, and there
happen to be others near him, they all take
the Alarm, and rattle in like manner ....

"They are generally from three to five Feet
long, and do not commonly exceed twenty rat-
tles; and yet I have it attested by a Man
of Credit, that he killed a Rattlesnake some
Years since, that had between seventy and
eighty Rattles, with a sprinkling of grey
Hairs, like Bristles over his Body."

Dumeril, A. M. C. et Bibron, G.

185i4. Erpetologie Ge'nerale. Paris, Vol. 7, Part 2, pp.
XII + 781-1536 (ref. pp. 1457-U60) .

Duncan, P. Martin

1880 Reptilia in Cassell's Natural History, Vol. A,

pp. VIII +380 (ref. pp. 316-317). Also Edi-
tion of 1913, pp. XX + 432 (ref. pp. 258-259).

Feoktistow, A. E.

1889 Zur Physiologie der Klapper des Crotalus durissus.
Bull. Acad. Sci. St. Pe'tersbourg (n.s.) Vol. 1,
No. 1, pp. 1-4.

1891 Same. Mel. Biol. Acad. Imp. Sci. St. Pe'tersbourg,
Vol. 13, No. 1, pp. 1-4.

1893 On the Physiology of the Rattle of Crotalus duris-
sus. Ann. and Mag. of Nat. Hist., 6th Ser.,
Vol. 11, pp. 54-58 (translation of above).
He observed no synchronism between skin
changing and acquisition of rattles in cap-
tive snakes . A paper often quoted in support
of this contention.

Finney, Fred and Finney, L. L.

1931 Arizona Reptiles: Their Lives and Habits. Tuscon,
pp. 1-19.

Nine times out of ten rattlers do not sound
a warning until they are stirred up.

Funkhouser, W. D.

1925 Wild Life in Kentucky. Frankfort, pp. 1-385 (ref.
p. 139).

51

Gadow, H.

1908 Through Southern Mexico, Being an Account of the

Travels of a Naturalist. London, pp. XVI + 527
(ref. pp. 197-201).

Carman, Samuel

1884. North American Reptiles. Mem. Mus. Comp. Zool.,
Vol. 8, No. 3, pp. XXXI + 135 (ref. pp. XXV-
XXVI).

1888 The Rattle of the Rattlesnake. Bull. Mus. Comp.

Zool., Vol. 13, No. 10, pp. 259-268.

An important paper on the method of rattle
structure and use.

1889 On the Evolution of the Rattlesnake. Proc . Boston

Soc. Nat. Hist., Vol. 24-, pp. 170-182.

The second of Carman's fundamental papers,
the most important yet published. This one
deals particularly with the genesis of the
rattle.

1892 The Reptilian Rattle. Science, Vol. 20, No. 492,
pp. 16-17.

Geare, Randolph I.

1903 Venomous Snakes. Scientific American, Vol. 88,

No. 8, p. 137.

Gibson, Frank

1904 Superstitions about Animals. London, pp. 1-208

(ref. 102-103).

Gillespie, T. H.

1937 The Way of a Serpent (a Popular Account of the
Habits of Snakes). London, pp. 1-221 (ref.
pp. 129-132).

Goldsmith, Oliver

1774 An History of the Earth and Animated Nature.
London, Vol. 7, pp. 1-368.

The snakes make a loud rattling noise when-
ever they move. Young snakes have no rat-
tles (pp. 209-210).

Goode, G. Brown

1874 Rattlesnakes; A Query. Forest and Stream, Vol. 2,
No. 8, p. 123.

A query as to the largest rattler. The
largest in the USNM is 6 ft. 2 in. with 20
rattles. Mentions one from Tennessee said
to have weighed 117 lb. and had 87 rattles.

Grew, Nehemiah

1681 Catalogue and Descriptions of the Natural and Arti-
ficial Rarities Belonging to the Royal Society

52

and Freaerved at Gresham Colledge. Section III";

Of Serpents, pp. 1,^-52.

Denies that the rattle contains venom, as
has been stated by a previous author. The
rattle sounds at any movement of the snake.
The author does not think that the rattle in-
dicates the age of the snake, as has been
stated (ref. pp. 50-51).

Grinnell, Joseph and Storer, Tracy I.

1924. Animal Life in the Yosemlte. Berkeley, pp. XVIII +
732.

Mentions a specimen with 22 rattles (pp.
64,6-64.7), one of the longest probably-re-
liable records in the literature.

Giinther, A.

1838 On a Collection of Reptiles from China. Ann. and
Mag. of Nat. Hist., 6th Ser., Vol. 1, No. 3,
pp. 165-172.

Describes a peculiar tail modification in

Halys acutus .

Guthrie, J. E.

1926 The Snakes of Iowa. Bull. Iowa State College of
Agr. and Mech. Arts, No. 239, pp. 145-192
(ref. p. 190).

Haldeman, S. S.

1372 Vibrations of the Tail in Snakes. Am. Nat., Vol.
6, p. 304.

Hay, 0. P.

1337 The Massasaugua and its Habits. Am. Nat., Vol. 21,
No. 3, pp. 214-216.

An early expression of the bison theory of
acquisition.

1892 The Batrachians and Reptiles of the State of

Indiana. Seventeenth Ann. Rep. Indiana Dept.
of Geol. and Nat. Res., pp. ^09-602 (ref.
p. 537).

Henderson. J. G.

1872 Use of the Rattles of the Rattlesnake. Am. Nat.,

Vol. 6, pp. 260-263.

"In effect it amounts to this, and by ex-
perience its enemies learn to understand its
language, 'I am a rattlesnake, armed with
what will be death if you come too near;
give me a wide berth'.'" But the rattle is
not sounded until the snake is discovered.

Higgins, S. B.

1873 Ophidians, Zoological Arrangement of the Different

Genera, etc. New York, pp. 1-239.

As snakes with 4.O rattles have been found.

53

their living to an age of more than AO years
is established (p. 82).

Hoffmann, C. K.

1890 Reptilien (in Bronn's Klassen und Ordnungen des
Thier-Reichs) Vol. 3, pp. UOl-2039.

Quotes extensively from Czermak (pp. 1417-
U20) .

Holbrook, John E.

1838 North American Herpetology. Philadelphia, Vol. 2,
pp. 1-125 (ref. p. 85).

I84.2 North American Herpetology, Second Edition. Phil-
adelphia, Vol. 3, pp. 1-128 (ref. pp. U-15) .

Hopley, Catherine C.

1877 The History of a Rattle. Aunt Judy's Magazine,
Vol. 15, No. 135, pp. 527-536.

1882 Snakes: Curiosities and Wonders of Serpent Life.
London, pp. VIII + blJ^..

An interesting, but not always accurate,
• discussion of the rattle (Chapter 17,
pp. 29A-3U).

1887 The Rattle. Forest and Stream, Vol. 28, No. 24,
pp. 512-513.

Rattlers do not hiss.

Hornaday, William T.

1904 The American Natural History. London, pp. XXV +

"He rattles to save himself from injury, and
his persistent whirr has saved thousands of
persons and tens of thousands of domestic
animals from being bitten" (p. 34-8) .

James, Edwin

1823 Account of an Expedition from Pittsburgh to the
Rocky Mountains . . . under the Command of
Major Stephen H. Long. Philadelphia, 2 Vols.
Mentions use of pulverized rattle as a
medicine by the Indians (Vol. 1, p. 236).

Jones, Thomas R.

I84I A General Outline of the Organization of the
Animal Kingdom. London, pp. XV + 732.

Makes a noise at the slightest movement
(p. 559). Repeated in fourth edition,
London, 1871, p. 731.

Kalm, Peter

1752 Om Skaller-Ormen. Kongl . Vet. Acad., cp. 308-319;
1753, pp. 52-67, pp. 185-194.

1758 An Account of the Rattle-snake, and the Cure of its

54

Bite, as Used in North America. Medical, Chi-
rurgical and Anatomical Cases. London, pp.
282-293.

This is a translation (of the first only) of
Kalm's three papers listed above. The rattle
sounds like a mamber of spinning wheels.
Rattlers do not rattle when waiting for prey.

Klauber, L. M.

1927 Some Observations on the Rattlesnakes of the Ex-
treme Southwest. Bull. Antivenin Inst, of Am.,
Vol. 1, No. 1, pp. 7-21 (ref. p. 16).

1936 A Key to the Rattlesnakes with Svimmary of Charac-

teristics. Trans. S. D. See. Nat. Hist., Vol.
8, No. 20, pp. 185-276 (ref. p. 208).

1937 A Statistical Study of the Rattlesnakes. IV. The

Growth of the Rattlesnake. Occ. Pap. S. D.

Soc. Nat. Hist., No. 3, pp. 1-56.

A curve coordinating the size of oreeanus
with the number of rattles, p. 16.

Konig, C.

1906 Die Rassel der Klapperschlange. Naturwiss.

Wochensch. Vol. 21, No. 4, pp. 49-55. (See
also Jagerztg. Neudamm, Vol. 4.7; and Natur. u
Haus, Vol. 14) .

Lacepede, B. G. E.

1789 Histoire Naturelle des Serpens. Vol. 2, Paris,
pp. 20 + 527.

A new rattle is acquired at every molting,
and therefore more than one per year. For
a translation of Lacepede, see Robt. Kerr:
The Natural History of Oviparous Quadrupeds
and Serpents. Edinburgh, 4 vols., 1802;
Rattle, Vol. 4, PP. 252-262.

Le Conte, John

1853 On the Venomous Serpents of Georgia. Southern
Med. and Surg. Journ., N.S., Vol. 9, No. 11,
pp. 645-667 (ref. pp. 655-656).

Lewis, Ellas

1874 Concerning Serpents. Popular Science Monthly,
Vol. 21, pp. 257-275 (ref. p. 274).

Lombard, H. I.

1881 The Rattlesnake and the Copperhead. Forest and
Stream, Vol. 17, No. 5, p. 88.

Lydekker, Richard

1896 The Royal Natural History. Vol. 5, Reptiles,

Amphibians, Fishes, the Lower Vertebrates and
Their Allies. London, p. XVI + 584 (ref. pp.
241-243) .

55

Lydekker, Pichard (cont.)

1912 Reptiles (in Reptiles, Amphibia, Fishes, and Lower
Chordata; Edited by J. T. Cunningham). London,
pp. XVI + 510 (ref. 153-155).

Manning, F. B.

1923 Hearing in Rattlesnakes. Journ. Comp. Psych.,
Vol. 3, No. U, PP- 241-247.

An exceedingly important paper on the range
of vibrations audible to rattlesnakes. The
conclusion is reached that they cannot hear
the sound of the rattle.

Maximilian, Prinz zu Wied

1865 Verzeichniss der Reptillen, welche auf einer Reise
im nordlichen America. Nov. Act. Acad. Leopold.
Carol. Nat. Cur., Vol. 32, pp. IV + I4I (ref.
pp. 69-74).

Mead, Richard

1747 A Mechanical Account of Poisons in Several Essays.
Third Edition, London, pp. XLVIII +320.
The rattle is not sounded as the snake
moves (pp. 8I-84) .

Medden, Rheua V.

1930 Tales of the Rattlesnake: From the Works of the

Early Travelers in America. The Rattle. Bull.
Antivenin Inst, of Am., Vol. 3, Nn. 4, pp.
105-107.

Interesting quotations from old books of
travel.

Mitchell, J. D.

1903 The Poisonous Snakes of Texas, with Notes on Their
Habits. Trans. Texas Acad, of Science, Vol. 5,
Part 1, No. 2, pp. I-48.

"During the mating season the males are very
aggressive, and will promptly rattle a chal-
lenge to an Intruder; I have received such
a challenge when 50 feet away" (p. 36).

Mivart, St. George

1871 On the Genesis of Species. London, pp. 1-314 (ref.
p. 62).

1888 Rattlesnake. Encycl. Britan. Ninth Ed., Vol. 20,
p. 293. Also Eleventh Ed., Vol. 22, pp. 919-
920.

The noise is like a weak child's rattle and

can be heard 10 to 20 yards.

Mole, R. R.

1895 On the Formation and Disintegration of Segments of
Caudal Appendage in C. horrldus. Journ. Field
Nat. Club. (TrinidadT, Vol. 2, pp. 189-191.

56

Mole, R. R. (cont.)

Observed a rattler to sound the rattle when
swallowing a rat and therefore in a defense-
less condition.

Moran, G. H.

1878 Habits of Rattlesnakes. Forest and Stream, Vol.
11, No. 17, p. 34-1.

To an inquiry from Arizona as to whether rat-
tlers in the east carry their rattles narrow
edge down (as in Arizona), or broad side
down, the editor replied broad side downl

Nicholson, Hunter

1877 Rattlesnakes in Wet Weather. Nature, Vol. 16,
p. 266.

Remarks on the use of the rattle when damp,
and the ability of rattlers to indicate dif-
ferent feelings by means of the rattle.

Noble, G. K.

1937 The Sense Organs Involved in the Courtship of

Storeria. Thamnophis and Other Snakes. Bull.
Am. Mus. Nat. Hist., Vol. 73, Art. 7, po. 673-
725.

The importance of dorsal skin glands in sex

recognition.

Ord, George

1815 North American Zoology. Reprint (l89ii) from
Guthries' Geography, pp. 291-361.

The rattle makes no noise when the snake
crawls (p. 358) .

Ott, Isaac

1882 The Vibration of the Rattlesnake's Tail. Jour.
Nervous and Mental Disease, Vol. 9, No. 3,
pp. 5U-516.

The rate of vibration of the tail was found

to be 60 per second.

Owen, Charles

17^2 An Essay Towards a Natural History of Serpents.
London, pp. XXIII + 24.0 + (XII) .

"The tail is composed of several loose bony
Articulaments that make a roaring Soiind,
loud enough to be heard at a distance; and
therefore called the Bell-snake" (p. 91).

Owen, Richard

1866 Oil the Anatomy of Vertebrates. Vol. 1, Fishes and
Reptiles. London, pp. XLII + 650 (ref. p. 55>4) .

Pack, Herbert J.

1930 Snakes of Utah. Utah Agric. Exper. Sta., Bull.
221, pp. 1-32.

57

Parker, H. W.

1929 Rattlesnake. Encycl. Brltan., Mth Ed., Vol. 18,
p. 996

Perry, Armstrong

1920 Crotalus horrldus Rattles for Half an Hour.

Copeia, No. 86, pp. 85-86.

Piso, G.

184,7 Hlstoria Natiiralls Brasilae.

Not seen. Quoted by Tyson, 1683.

Pitman, Charles R. S.

1938 A Guide to the Snakes of Uganda. Kampala, pp.
XVIII + 362.

"All these warning devices afford conclusive
evidence that normally a snake is not ag-
gressive (p. 32)".

Pope, Clifford H.

193V Snakes Alive and How They Live. New York, pp.
XII + 238 (ref. pp. 34, 40-Al) .

Pope, Saxton

1919 Rattlesnakes: Their Venom and Antidotes. Calif.
State Journ. Med., Vol. 17, pp. 49-52.
Vibratory rate 120 per minute.

Pope, T. E. B.

1923 Facts and Fallacies Concerning Rattlesnakes. Year-
book Public Mus. Milwaukee, Vol. 3, pp. 132-
134.

Pycroft, W. P.

1905 The Story of Reptile Life. London, pp. 1-212 (ref,
pp. 159-160).

Quelch, J. J.

1891 The Rattlesnake - the Growth of the Rattle.

Timehri, Vol. 5 (n.s.), Pt. 1, pp. 1-11; also
pp. 170-171.

Important observations on rattle formation

in captive specimens.

Putnam, F. W.

1872 The Rattle of the Rattlesnake. Am. Nat., Vol. 6,
pp. 693-694.

The rattle is a detriment to the snake, ex-
cept as a sex call.

Rltter, W. E.

1921 The Rattling of Rattlesnakes. Copeia, No. 94,

pp. 29-31.

58

Hitter, W. E. (cont.)

A specimen of C. ruber could be heard 80
paces

Roberts, Morley

1930 The Serpent's Fang. Essays in Biological Criti-
cism. London, pp. 1-263.

The evolution of the rattle resulted from a
pathological condition. Once in British
Columbia the author heard rattler calls con-
tinuously all night on a still night.

Ruthven, Alex. G., Thompson, Crystal and Thompson, Helen

1912 The Herpetology of Michigan. Mich. Geol. and

Biol. Surv., Pub. 10, Biol. Ser. 3, pp. 1-161
(ref. p. 126). Also in Second Edition, Mich.
Handbook. Ser. No. 3, pp. IX + 228 (ref. pp.
132-133).

Schlegel, H.

1837 Essai sur la Physionomie des Serpens. Vol. 2,
Leide, pp. 1-606 (ref. pp. 556-561).

Schmidt, Karl P.

1929 The Truth about Snake Stories. Field Mus. Nat.
Hist., Zool. Leaflet 10, pp. 1-19. Also Sci.
Am., Vol. Ul, No. 2, pp. 134-136.

Seba, A.

1734-1765 Locupletissimi Rerum Naturalium Thesaxiri, etc.
Amstelaedami, 4 Vols.

Vol. 3, plates 45, 95, and 96, are interest-
ing in showing a steady and continuous growth
of very long rattle strings, as inaccurately
pictured by the illustrators of that day.

Shaw, George

1802 General Zoology or Systematic Natural History.

London, Vol. 3, Part 2, pp. IV, 313-615 (ref.
pp. 317-336).

Plate 91 shows a "wood rattlesnake" with

44 rattles.

1809 Zoological Lectures Delivered at the Royal Institu-
tion in the years 1806 and 1807. London, Vol.
1, pp. IV + 225 (ref. p. 36).

Shufeldt, R. W.

1900 Chapters on the Natural History of the United

States. New York, pp. 1-450 (ref. pp. I4I-I42) .

Smith, W. H.

1882 Report on the Reptiles and Amphibians of Ohio.

Rep. Geol. Surv. Ohio, Vol. 4, Zool. and Bot.,
Part 1, Zool., pp. 629-734 (ref. p. 673).

59

Stejneger, Leonhard

1895 The Poisonous Snakes of North America. Ann. Rep.
Smithsonian Inst, for 1893, pp. 337-4.87.

An excellent summary with quotations from
prior articles (pp. 380-392).

Stephenson, John

1838 Medical Zoology and Mineralogy. London, pp. 1-350
(ref. pp. 58-59).

Strecker, John K.

1929 Further Studies in the Folk-lore of Reptiles.

Baylor Univ. Contrib. to Folk Lore, No. 1,

pp. 1-16.

The myth of ascertaining a rattler's age by
the length of the rattle string is mentioned
(p. 9).

Storer, Tracy I., and Wilson, Beryl M.

1932 Feeding Habits and Molt of Crotalus confluentus

oreganus. Copeia, No. 4 of 1932; pp. 169-173.
Experiences with captive specimens raised
the suspicion that new rattles are not always
coincident with molting.

Surface, H. A.

1906 The Serpents of Pennsylvania. Month. Bull. Div.
Zool., Penn. State Dept. Agric, Vol. -4, Nos.
4 and 5, pp. 113-208 (ref. p. 191).

Taylor, W. E.

1892 The Ophidia of Nebraska. Ann. Rep. Neb. State Bd.
Agric. for Year 1891, pp. 310-357 (ref. pp.
353-354).

1895 Preliminary Notes on the Osteology of the North
American Crotalidae. Am. Nat., Vol. 29, pp.
281-285.

Terron, Carlos C.

1931 Los Crotalianos Mexicanos. Ann. Inst. Bioj..,

Vol. 2, No. 1, pp. 173-194 (ref. pp. 47-49).

Tyson, Edward

1683 Vipera Caudisona Americana, or the Anatomy of a

Rattle-snake Dissected at the Repository of the
Royal Society. Philos. Trans., Vol. 12, No.
144, pp. 25-53.

Gives the warning theory, quoting G. Piso,
1648.

Van Denburgh, John

1922 The Reptiles of Western North America. Occas.
Papers Calif. Acad. Sci., No. 10, 2 vols.,
pp. 1-1028.

60

Van Denburgh, John (cont.)

A note on catching C. leoldus by causing them
to rattle (p. 963).

Verrill, A. Hyatt

1937 Strange Reptiles and Their Stories. Boston, pp.
XIV + 195.

The rattle cannot properly be termed a warn-
ing since it is a nervous reaction beyond
control of the snake (p. 30).

Wallace, A. R.

1871 (Comments on Communications re Rattlesnakes) .
Proc. Zool. Soc. London, 1871, p. 2.

1890 Darwinism. Second Edition, London, pp. XIV + 4.9A
(ref . pp. 262-263) .

Werner, Franz

1913 Die Lurche und Kriechtiere (in Brehm's Tierleben,
5th ed.) Leipzig, Vol. 2, pp. XVI + 598 (ref.
pp. 557-558).

Wiley, Grace 0.

1929 Surprising New Facts about Snakes. The American
Weekly.

Notes on the Texas Rattlesnake in Captivity with

Special Reference to the Birth of a Litter of
Young. Bull. Antivenin Inst. Am., Vol. 3,
No. 1, pp. 3-14.

The acquisition of rattles by young is dis-
cussed.

Williams, Mabel C.

1920 Vibration Rate of the Tail of a Rattlesnake.

Science (N.S.) Vol. 51, No. 1305, pp. 15-16.
The pitch of the tone produced does not de-
pend on the speed or constancy of the tail
vibration, but on the natural resonance of
the rattles.

Wood, J. C.

1363 The Illustrated Natural History. London, Vol. 5,
pp. 1-810 (ref. 100-101).

This immensely popular work, first published
in London in 1851, was the standard natural
history in Great Britain and the United
States in the last half of the nineteenth
century, during which time many editions ap-
peared. While Wood did not originate the
rattle-per-year myth, his natural history
undoubtedly did much to spread and intrench
it.

61

Wyman, Jeffries

1862 On the Mode of Formation of the Rattle of the
Rattlesnake. Proc. Boston Soc. Nat. Hist.,
Vol. 8, p. 121.

62

cAcnie

Bookbinding Co., Inc.

.00 Cambridge SI.
Charleslown. MA 02129

3 2044 093 361 160