I. Brown, Jr.
COLLECTION
THE EFFECTS OF PARASITIC AND OTHER KINDS
OF CASTRATION IN INSECTS
By
WILLIAM MORTON WHEELER
REPRINTED FROM
THE JOURNAL OF EXPERIMENTAL ZOOLOGY
Volume VIII No. 4
JULY, 1910
WILLIAMS & WILKINS COMPANY
BALTIMORE
THE EFFECTS OF PARASITIC AND OTHER KINDS
OF CASTRATION IN INSECTS'
WILLIAM MORTON WHEELER
WITH EIGHT FIGURES
I. THE EFFECTS OF STYLOPIZATION IN WASPS AND BEES
The perusal several years ago of a very interesting paper by
Perez ('86) on bees of the genus Andrena infested with Stylops
led me to undertake a similar study of our North American wasps
of the genus Polistes parasitized by Xenos. I began to collect
stylopized P. variatus during the autumns of 1898 and 1899, while
I was living in Chicago, but the wasps proved to be too scarce to
serve my purpose. During the summer of 1900, however, while
I was spending my vacation at Colebrook, in the Litchfield Hills,
Connecticut, I noticed many specimens of Polistes metricus Say
infested with Xenos (Acroschismus) wheeleri Pierce and I at once
began to collect them. 2
In ten days during the latter part of August I gathered one
thousand specimens of the Polistes from flowers of the golden
1 Contributions from the Entomological Laboratory of the Bussey Institution, Harvard University
No. 20.
2 There may be some doubt about the specific names of the host and parasite here mentioned. I
have called the wasp P. metricus as this is the name under which it is commonly known and because
our extremely variable species of Polistes are in a state of great taxonomic confusion. Miss Enteman,
who has studied them very extensively ('04), would probably refer my specimens to P. pallipes Le-
peletier, while others would be inclined to regard them as belonging to P. fuscatus Fabricius. Brues
('09 } and I had identified the parasite as Xenos peckii Kirby, but Pierce ('08), regards it not only as speci-
fically, but also as generically distinct. He has given it the name wheeleri and placed it in a new genus
(Acroschismus) because it has the cedeagus " considerably dilated at the base, arising between two claws,"
whereas Kirby's species is placed in another new genus, Schistosiphon. because it has the cedeagus
"cleft at the apex." The old genus Xenos of Rossi he restricts to the European species (vesparum
Rossi and jurinei Saunders). Although these generic distinctions may prove to be valid, I shall use
the old name Xenos in the present paper.)
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 8, NO. 4.
378 William Morton Wheeler
rod (Solidago canadensis) within an area of less than a square mile
and noted the sex of each individual and the number, sex and posi-
tion of the Xenos parasites which had protruded their heads
between the gastric sclerites of the wasps. A further study, of
the form and coloration of the hosts was undertaken in the hope
of detecting modifications, like those seen by Perez in stylo-
pized Andrenae. My observations, however, gave much less in-
teresting results than those obtained by the French naturalist,
and I therefore refrained from publishing them and awaited
an opportunity to continue them on additional material. This
opportunity, however, has not presented itself, so that I have de-
cided to give my observations for what they are worth, in the hope
that they may be amplified by some other more fortunate ob-
server. My preserved Xenos material was turned over partly to
Miss Enteman, who published a short paper on the genital ducts
of the females ('99), and partly to Mr. C. T. Brues who published
a brief account of the embryology of the parasite ('03). The
table on the page opposite contains the results of counting the
sexes of both host and parasite on the different dates of collecting.
From this table the following conclusions, valid only, of course,
for the particular summer and locality in which the insects were
collected, may be drawn:
I. Of the total number (1000) of Polistes metricus, 251 or fully
25 per cent were stylopized. This is a high percentage, though
as will be shown, it has been exceeded in the statistics of other
observers. It may be regarded as too great, first because the
parasitized individuals, being more sluggish, would be more
easily caught, and second, because niy interest in such specimens,
would lead me to exercise greater care in capturing them. I
would say, however, in answer to such objections, that I attempted
to collect the- wasps at random without noticing whether they
bore parasites or not, that a long handled net was used in captur-
ing them, and that the table con tains only specimens in which Xenos
had already protruded their heads between the gastric segments
of the wasps. A number of apparently unifested wasps were
dissected and were found to contain larval parasites, so that the
actual percentage of parasitism was even greater than that indi-
cated in the table.
Effects of Castration in Insects
379
.
15 n
flj
3
JJ
-v
s
t3
JJ
.
Number of
Collection
JU
rt
D
Total Number
Polistes Take
1
1
JH
"rt
'S
1
^CJ
rt
1
Total Number
Infested Polisi
Number of Ma
Infested
"o JS
u fl
J3
| J
* 1
<1J
PH
1
S o
3 B
2 #
*rt
3
S
"8 I
!*
i
Number of
Female Xent
August
i
14
60
4
56
33
33
85
?i
H
2
16
7^
3
69
3 1
3 1
6 7
58
9
3
'9
3 1
5
26
H
2
12
55
49
6
4
20
108
5
103
43
3
4
89
73
16
5
21
73
6
67
18
18
36
24
12
6
22
H3
6
137
12
3
9
"9
10
9
7
2 3
66
'5
5 1
2O
2
18
5
36
H
8
2 4 .
"37
36
101
21
5
16
4
3 1
8
9
27
167
5
117
2 9
8
21
55
34
21
10
29
'43
7
136
3
2
28
66
56
IO
Totals:
IOOO
"37
863
251
2 5
226
562
443
II 9
Aver.
1
and per
i
IOO
'3-7
86.3
25.1
2 -5
22.6
56.2
44-3
II.9
cent
J
2. The number of male Polistes increased very suddenly Aug-
ust 23 to 27 and then fell off still more abruptly. Apparently
these collections were made at the time of the emergence of the
male brood for the particular locality.
3. The greater difference in the ratio of male to female Polis-
tes (l : 6.3) is to be accounted for partly by this temporary
appearance of the males and partly, perhaps, by the fact that this
sex is much more wary and therefore more difficult to capture
than the females.
4. While the total number of females examined was somewhat
more than six times as great as that of the males, the number of
females stylopized was fully nine times as great as that of the
stylopized males. As the male brood of the wasp appears late
in the season this may be due to a partial immunity of this sex
from the attacks of the parasites, since Brues ('05) has shown that
the triungulin Xenos must enter the wasp larvae in the spring or
early summer (vide infra, p. 393.)
Effects of Castration in Insects 381
or two Xenos were present. The table shows that the average
number in all the infested wasps was about 2.4. These numbers
probably represent the few survivers of an originally much greater
number which had lived as larvae in the individual larval wasps.
Brues ('03) took as many as 31 larvae of X. pallidus of both sexes
from a single larva of the Texan P. annularis!
2. Both sexes of the Xenos may occur in the same Polistes, but
when the number exceeds 4, the Xenos are all males. In only one
case did I find as many as 3 female Xenos in the same host; in
all other cases there were only one or two. In. 45 of the 251 infested
Polistes, or in nearly 18 per cent, Xenos of both sexes occurred.
Hence while there is undoubtedly a tendency, as Brues has observ-
ed ('03), for the sexes to be the same in the same host, this is so
far from being a general rule, that the sex of the parasite cannot
be supposed to be determined by its host.
3. When more than one female Xenos is present in the same
Polistes, they are of the same size but each is smaller than the
females occurring singly in a wasp.
4. When both sexes infiabit the same Polistes the heads of the
females protrude between the more posterior segments, whereas
the cephalic ends of the male puparia may protrude between any
of the segments behind the first. The heads of the females there-
fore usually appear from under the posterior edges of the fourth
or fifth abdominal segments. This is obviously an adaptation to
the greater length of the female parasite, which has to he stretched
out in the abdomen of its host and could not protrude its head
between the more anterior segments without bending its body.
Sometimes both sexes protrude their heads side by side from under
the tergite or stermte of the same segment. Sometimes one sex
is on the dorsal, the other on the ventral side of the same wasp,
but protruding from the same segment.
5. When the female Xenos protrudes its head between two ter-
gites, it lies with its ventral surface uppermost, i.e., its dorso- ven-
tral orientation is the reverse of that of its host: when it protrudes
its head between two stermtes, it lies with its ventral surface down-
ward, i.e., with the same dorso-ventral orientation as the wasp.
This is obviously an adaptation to copulation with the winged
382 William Morton Wheeler
male,' for the latter must have to insert its penis along the ventral
surface of the head of the female and immediately under the over-
lapping sternite or tergite of the host.
That several of the conclusions drawn from the table on page
379 cannot have general validity is shown by comparing them with
the statistics of other observers. Home ('72) says that the speci-
mens of Polistes hebraeus which he observed in India were "ex-
tremely troubled with Stylops (Xenos), every fifth or sixth one
taken having a female of one under one of the segments of the abdo-
men." Theobald ('92) found that among 180 Andrena lapponica
taken in England during 1887, 1:05 or 58 percent contained Stylops;
of 60 bees of the same species, taken in 1888, 54 or 90 per cent were
badly stylopized. He believes that the female Andrenae are more
afflicted with the parasites than the males, and he records the num-
ber of Stylops found in the 54 bees taken during 1888 as comprising
33 females and 21 males; 2 females each contained 2, 3 males con-
tained 2, 25 females and 18 males i each. The corresponding
numbers for 40 stylopized specimens of Andrena nigroaenea
were 3 females each with 3 Stylops, i male with 3, 3 females
with 2, 5 males with 2, 16 females with I and 12 males with I,
making 22 females and 18 males. On the basis of these figures
Theobald differs from Perkins. ('92), who found the males of
various Andrenae and Halicti more frequently stylopized than
the females. This author says that he has seen hundreds of
stylopized male Halictus tumulorum, but has never seen a female
in this condition. Although Theobald's conclusions agree with
my own, his data do not furnish very strong support in favor of
his contention, since in A. lapponica the ratio of parasitized males
to females-is 1 : 1.5 and in A. nigroa^nea only i : 1.2. Skinner ('03)
counted 34 stylopized individuals among 140 Polistes texanus,
which he found at Pecos, Texas. He says that " most of theXenos
appeared to be females and only 4 males were secured."
The percentage in this case is very similar to that which I found
in P. metricus. Brues ('05) has published some statistics on two
co Ionics of the Texan P. annularis infested with Xenos nigrescens
Brues and X. pallidus Brues. In these cases the amount of para.
Effects of Castration in Insects 383
sitization was very great. In one nest there were 86 wasps, 44 or
51 percent of which contained X. nigrescens. There were from one
to seven in each wasp (an average of 2.6 per host), and of the total
number of Xenos (94)591 were males and only 3 females. In the
other nest there were 42 wasps, and 36 or more than 85 per cent
were stylopized. The total number of the parasites in this case
X. pallidus was 125 (81 males and 44 females); the highest num-
ber in a single wasp being 10, the average per host 3.6.
Fuller consideration must be given to the effects of the stylopids
on their hosts. This may properly begin with a resume of the
excellent work of Perez (1886) who examined stylopized speci-
mens of 47 species of Andrena. The effects produced by Stylops
in these bees is so considerable as to render their specific deter-
mination difficult. This is not surprising perhaps, when we con-
sider the vast number of closely related species in the genus. All
the known specimens of certain "species" (F. Smith's Andrena
insolita, separata and victima) have been found to be stylopized,
which gives force to Perez's opinion that these are not true species
but merely parasitized individuals of forms that are already known
under other specific names. Perez describes minutely the fol-
lowing modifications as characteristic of stylopized Andrenae:
(i) The abdomen is shortened and swollen and therefore more
globular, the shortening being due to an attenuation of the termi-
nal segments. (2) The head is usually smaller than that of nor-
mal specimens. (3) The villosity of the abdomen is more
abundant, longer and more silky, especially on the terminal seg-
ments, and its color is often greatly altered, becoming lighter
and more reddish or fulvous. The villosity of the thorax may
undergo similar bat less pronounced changes. (4) The puncta-
tion of the body becomes finer, .denser and more superficial in
correlation with the pilosity, which arises from the punctures.
These changes are common to both sexes and therefore affect
specific characters. They 'give the specimens a peculiar pseudo-
specific facies. Perez therefore rightly warns against basing
new species of Andrena on stylopized individuals.
The following changes affect the secondary sexual characters:
(l) The normal males of the genus Andrena, as in many other
384 William Morton Wheeler
genera of bees, have a greater amount of yellow or white on the
face or clypeus or on both than the cospecific females. Stylopiza-
tion tends to diminish this light color very perceptibly and hence
to make the face of the male resemble that of the female. In the
female the parasites produce the reverse effect, making the face
resemble that of the male. "It is difficult to find a stylopized
male of A. labialis, e.g., whose face is normally colored and, on
the other hand, it is quite as rare to find a stylopized female of
this species having the face entirely black." (2) The normal fe-
male Andrena differs from the normal male in the structure of its
hind legs, the tibias of which are modified for collecting pollen.
They are always robust and incrassated and have a brush of long,
curved hairs, especially on their internal surfaces. Similar hairs
are found also on the femora, COX;E and metapleura. The metatar-
sal joint of the hind legs is also kilated or enlarged and is furnished
with rows of stiff hairs on its lower surface. In the male the hind
tibiae and metatarsi are slender and bear only short, sparse, straight
hairs and this is true also of the coxae and metapleurae. The pres-
ence of Stylops in the abdomen of the female diminishes the de-
velopment of the pollen-collecting apparatus to such a degree
that the hind legs become like those of the male. The reverse
occurs in stylopized males, the organs under consideration be-
coming more enlarged and approximating to the female type in
their pilosity. The modifications in this sex, however, are rarer
than in the female and in both sexes they vary greatly in different
stylopized individuals. (3) The frontal furrow near the internal
orbit of the eyes, which s filled with velvety pubescence, is wellr
developed in the normal female, but feeble or absent in the normal
male. In stylopized Andrenae this furrow may'undergo diminu-
tion of development in the female and becomes accentuated in the
male. (4) Although the female Andrena has 12-jointed, the male
13-jointed antenna;, there is no modificationof the numberof joint
in parasitized individuals. The antennSe of the normal sexes may
differ in the length of the second funicular joint. In one species,
A. Trimmeriana, the second funicular of the normal female is as
long as the two succeeding joints taken together, whereas in the
normal male this joint is at most half as long as the succeeding
Effects of Castration in Insects 385
joint. In the stylopized male of this species Perez found the sec-
ond funicular attaining to two- thirds the length of the third joint
and to this extent approximating to the conditions in the female.
(5) The normal female Andrena bears a fringe o! long hairs, the
ana fimbria, on the edge o the fifth adbominal sternite, but th s
fringe is lacking in th normal male. Stylopization tends to sup-
press the development of the fimbria or causes it to disappear com-
pletely in the female and more rarely has the reverse effect on the
male. (6) The sting, which is peculiar to the female, is reduced in
size in the parasitized individuals, the copulatory organ of the
male is also reduced in length and becomes narrower and less
curved, while the paramera tend to become atrophied.
Perez concludes from these observations that, so far as the
secondary sexual characters of Andrena are concerned, the modifi-
cations induced by the Stylops are not merely attenuations, but
actual inversions of development. "The stylopized Andrena,
male or female, is not merely a diminished male or female; it is
a female which takes on male attributes; a male that takes on the
characters of the female."
The intimate correlation which exists between the structure and
instincts of all organisms, leads one to look for instinct peculiarities
corresponding with the morphological inversions described above.
Perez found only one stylopized female Andrena which had its
hind legs charged with pollen, and he therefore concludes that the
stylopized bees rarely or never forage or build nests like the normal
females. Normal and parasitized bees of both sexes, however,
visit flowers as this is not a unisexual instinct, and hence the
triungulins produced by the Stylops have an opportunity to
move off onto the plants, climb onto normal foraging bees and
thus get transferred to the brood in incipient nests. In this way
the perpetuation of the parasites is insured through a line of bees
capable of nourishing them.
The internal changes due to Stylopization have been studied by
Newport (48), Perez and Perkins (92). All of these authors
find that the testes and ovaries are not destroyed by the parasite
but are more or less reduced in size, in the male sometimes only on
the sideof the body bearing the Stylops. In the female the oocytes
386 William Morton Wheeler
or ova degenerate in their follicles and are evidently quite incap-
able of development, in the. male there may be ripe spermatozoa
in at least one of the testes. Perkins found motile spermatozoa in
all the stylopized males which he dissected, and Perez mentions a
male of Andrena decipiens taken in copula, so that this sex ma}
retain, at least occasionally, not only the normal mating instincts,
but the ability to fecundate normal females. The parasites before
maturity live on the fat-body and blood-tissue of their hosts and
do not attack the other organs directly. These undergo partial
atrophy through lack of nutrition. Observations similar to those
of Perez have been published by Saunders ('82) and Schmiede-
knecht ('83).*
Turning now toPolistes, we find that in this genus the secondary
sexual characters are in certain respects quite as clearly developed
as in the andrenme bees, but as wasps do not collect pollen, the
hind legs show no special modifications in the female. The fol-
lowing are the main external sexual differences observable in
Polistes metricus: The male has a slender thorax and long, nar-
row abdomen. The antennae are I3~jointed, with a long, slender
funiculus, not enlarging towards its tip; the second funicular
joint is little if any longer than the two succeeding joints taken
together. The face is long and narrow, with a pair of longitudinal
grooves running from the antennal insertions to the clypeus and
separated by a prominent longitudinal welt or elevation. The
clypeus is flat or even slightly concave and its surface is impunc-
tate. The whole face and clypeus, the anterior surface of the anten-
nae to within a few joints of the tip of the funiculus, the anterior
surface of the coxae, femora and tibiae, a series of transverse
bands or spots on the abdominal sternites behind as well as in-
cluding the first segment, are sulphur yellow. The two large ferru-
ginous spots on the first abdominal segment are usually well-
developed.
In the female the thorax is proportionally stouter and the
3 Though the publications of these authors antedate the article above reviewed, we are not to infer
that this implies priority of discovery. Perez says that he originally called the attention of these in-
vestigators to the facts and had himself published a preliminary account of his researches as early as
1880 in the Revue Internationale des Sciences, Tome I.
Effects of Castration in Insects 387
abdomen is decidedly shorter. The antennae are iz-jointed, with
a shorter funiculus slightly enlarging towards its tip; the second
funicular joint is nearly as long as the three succeeding joints taken
together. The face is decidedly shorter than that of the male,
the grooves and welt much less pronounced and the clypeus is
convex and coarsely punctate. The face is black, with the internal
orbit and sometimes portions of the clypeus, the anterior surface
of the scape and of the two first funicular joints, the anterior
surfaces of the tibiae and apical portions of the femora, ferruginous.
The sulphur yellow is restricted to the tarsi and the posterior
border of the first abdominal tergite, and the ferruginous spots
on the. first abdominal segment are obscure or wanting. The
wings are often somewhat more deeply infuscated than in the
male.
In stylopized Polistes metricus of either sex I fail to find any
modifications of a morphological character which could be
definitely attributed to the presence of the parasites. A few of
the more heavily stylopized females were abnormally small, but
with these exceptions, all the wasps were of normal stature. No
modifications of the antennas nor of the structure and proportions
of the face could be detected. A study of the coloration, however,
yielded more positive results, but even here, owing to the great
range of color variation to which P. metricus like all our other
species of the genus, is subject, the results are not capable of very
precise formulation. In the coloration of the face stylopized
males show no tendency to approach the female. In 14 out of 25
heavily stylopized females I find the clypeus of the usual black
or dark brown color; in the remaining II it is more or less ferru-
ginous or yellow. Some specimens have the free border of this
sclente sulphur yellow or its whole surface ferruginous, or only
its posterior border or sides of this color. One specimen has the
clypeus ferruginous with a small black spot in the center. It
would be possible to regard these cases as approximations to the
male type of coloration due to parasitism, were it not that per-
fectly normal, unstylopized females not infrequently exhibit the
same erythrism of the clypeus. I have not seen a sufficient num-
ber of P. metricus from different localities to be able to determine
JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 8, NO. 4
388 William Morton Wheeler
whether the percentage of this modification is so much greater
among stylopized than among unstylopized individuals as to
show that it must be attributable to the influence of the parasites.
I am inclined to believe, however, that it is part of a more general
erythnsm which affects also the abdomen of many parasitized in-
dividuals. This region, to a varying degree in such specimens, but
undoubtedly to a greater degree in those that are most heavily
stylopized, takes on in both sexes alike a distinct ferruginous tinge
which is usually most pronounced towards the posterior borders
of the tergites and sternites. Sometimes it may be very strongly
developed as in one rather small female taken August 29, and
bearing three male Xenos. In this case the second gastric segment
is entirely ferruginous, with the exception of a black anteromedian
triangle, and the posterior half of each of the remaining segments
and the whole clypeus, except its anterolateral corners, are rich
ferruginous. I have failed to notice m the legs, wings and antennae
of either sex in stylopized specimens any color modifications that
could not be regarded as falling within the wide limits of normal
specific variability. :
The color modification here described is not confined to styl-
opized specimens of P. metncus. It has also been observed by
Brues ('03) in two of the Texan species, P. rubiginosus and annu-
laris. "The stylopized Polistes," he says, ''can be recognized even
before the heads of the pupa cases begin to appear between the
sclerites of the abdomen, by their paler color. They seem never
to become as darkly colored as normal specimens. This lighter
color of parasitized specimens seems to apply only to the origi-
nally dark species, in P. rubiginosus there seems to be but slightly
if any lighter coloration. In the specimens of P. annularis from
which I raised Xenos, all of them females, the faded appearance is
especially noticeable upon the dorsuni of the abdomen. The
first abdominal which is normally piceous with a narrow apical
yellow band is in this case almost entirely bright ferruginous, or
is ferruginous with the border yellow. The remainder of the abdo-
men is normally piceous, but the posterior margins of the seg-
ments, especially the second and third tend to become more or less
broadly dull ferruginous in stylopized specimens."
Effects of Castration in Insects
389
There is also a modification of behavior in stylopized Polistes.
Several observers have noticed that such individuals are more
sluggish, that they fly about less actively, and Brues ('03) has
found that they are less inclined to use their sting, probably
because the voluminous parasites interfere with the exsertion of
this organ. A similar inability is observed in queen honey-bees
with ripe ovaries and in worker honey-bees with their crops full
of honey. The peculiarities of behavior in stylopized wasps are
such as would be expected in parasitized organisms for these al-
most invariably exhibit a general reduction of vitality due to
malnutrition.
Fig. 2. Abnormal abdomens of Polistes metrica; A and B, dorsal; C, ventral; O, lateral view.
Among the unstylopized female Polistes taken at Colebrook
there were three specimens with abnormal abdomens. Sketches
of these are shown in Fig. 2. The segments in some cases were
partially divided on either the right or left side, and in one case
there were several supernumerary sclerites. It might be inferred
that these abnormalities were the result of stylopization, for
although no Xenos were found in the specimens, these parasites
may have been present in the larvae from which the anomalous
individuals developed. I doubt this, however. At any rate, the
anomaly in question is not peculiar to wasps that are subject to
3QO ' William Morton Wheeler
stylopization or indeed to insects. Janet ('03) describes and
figures a very similar abnormality in Vespa rufa, an insect that
is not afflicted with Stylops or Xenos, and Con and Morgan ('92;
show that similar abnormalities are not uncommon in earthworms
and cestodes. In the case of Polistes the abnormality must be
produced either in the early embryonic stages while the metameres
are forming or at the time of the formation of the abdominal
sclentes in the pupa.
We may conclude, therefore, that Xenos produces no modifi-
cations of the secondary sexual characters of its Polistes host com-
parable to those produced by Stylops in the bees of the genus
Andrena, but merely a tendency to a reddish coloration of the
abdomen and face, a tendency which, s>o far as the abdomen is
concerned, io manifested equally by both sexes.
This general lightening of color in stylopized Polistes and its
reddish tinge remind one at once of the similar changes observed
by Perez in Andrenae, although in the latter insects it seems to
be confined to the pilosity. Pierce ('09, p. 32), cites the following
observations, .which show that a similar change of color was long
ago observed by Saunders in stylopized bees of the genera Pros-
opis and Hylaeus: "Prosopis gibba occasionally exhibits irregular
rufous patches on the abdomens of affected individuals (Saunders,
'50). Prosopis rubicola exhibits color changes regularly. The
nymphs of those Hy laei which are likely to produce the pale-colored
specimiens (H. versicolor), which prove, as anticipated, to be only
a variety of the H. rubicola consequent upon parasitic absorption,
may usually be identified within one or two days of their final
metamorphosis by assuming a yellow tinge, and may be set
apart as certain to produce male parasites. (Saunders '52.)" It
is not easy to account for this modification. Brues is inclined to
believe that "the reason that the reddish Polistes are not affected, is
that red is a more primitive color than piceous and that the color
simply becomes arrested at this stage and does not tend to become
so before the red stage." Thequestionof the development of varia-
tions of color in the species of Polistes is a very complicated one,
as Miss Enteman ('04) has shown, and a number of possible
explanations of the erythrism of stylopized individuals might be
Effects of Castration in Insects ' 391
suggested. The ontogenetic explanation suggested by Brues is
one of these, implying that a red stage precedes the brown or
black of the mature form of dark species like P. metricus. This
is borne out by the development of the color pattern in such species.
On this view st^Iopization inhibits color development in an
ontogenetically and presumably therefore in what corresponds to
a phylogenetically earlier stage. A second explanation is, how-
ever, suggested by Miss Enteman's studies. These tend to show
that the dark-colored races or species of Pohstes are due to cold
and moisture, the lighter yellow and red forms to heat and aridity.
This seems to be clearly indicated in the distribution of the spe-
cies, e.g., in such extreme forms as the yellow P. texanus and the
black canadensis. It is possible, therefore, that the erythnsm of
stvlopized P. metricus, which in normal coloration is closely
related to P. canadensis, is due to withdrawal of water from the
tissues by the developing parasites. This does not contradict the
ontogenetic and phylogenetic explanation but supplements them,
if we suppose that the primitive yellow or red color cannot pass
on to the piceous or black stage unless the tissues contain a suffi-
cient amount of water. Miss Enteman has shown that the piceous
or black color is in the form of pigment granules in the chitinous
cuticle of the wasp's integument, whereas the yellow is deposited
in the hypodermis. Erythnsm is probably due, therefore, to a
diminution in the cuticular pigment which permits the yellow hy-
podermal pigment to shine through. As both kinds of pigment
are the result of metabolism in the pupa, we can see how a disturb-
ance of metabolism either through withdrawal of water by the para-
sites or through other causes might lead to the deposition of a
smaller amount of the black pigment and hence to erythrism.
It is more difficult to account for the absence of all modifications
of the secondary sexual characters in stylopized Polistes, when
such modifications are so evident in Andrena. We may, perhaps,
account for this difference on one of the following hypotheses:
I. As will be shown in the sequel, complete extirpation of the
gonads in young larval insects, has produced in the few species
on which it has been performed, no appreciable effects on the de-
velopment of the secondary sexual characters. This indicates that
392 William Morton Wheeler
these characters may be so fixed and so nearly independent of the
gonads, except, perhaps, in the very earliest larval or late em-
bryonic stages, as to remain quite unaffected in their development
after the gonads have been completely removed. The degree of
this independence may be supposed to differ in different insects
and even in different individuals of the same species. It may be
slight or almost absent in Andrena and very well marked in Polis-
tes and this may account for the differences between the stylo-
pized specimens in the two genera.
. 2. The difference in the manifestation of changes in the second-
ary sexual characters may, however, be due to ethological differ-
ences between the two genera. Andrena has only male and female
forms and both under normal conditions are adequately fed in
their larval stages. In Polistes the larvae of the earlier broods in
the annual series, as Marchal has shown ('96, '97) are poorly fed
and as a result become sterile females, or workers. As imagines
they maintain themselves in a sterile condition by appropriating
very little of the food they collect to their own us.e, since they at
once lavish it in feeding the succeeding broods. Hence the females
of these earlier broods become sterile, in the first place through
alimentary castration of the larvae from which they develop, and
in the second place, maintain themselves in this condition as
adults through the nursing or nutncial function (nutncial castra-
tion). These peculiar phenomena will be more fully discussed in
the second part of this paper. Owing to these two forms of physio-
logical castration inhibition of the development of the reproduc-
tive organs is a common and normal occurrence in Polistes females,
and the parasitic castration induced by Xenos would not be ex-
pected to produce somatic changes of such magnitude or of
such a nature as Perez has observed in Andrena, all the females of
which are normally fertile mothers. In other words, the effects of
the Xenos on their hosts is of the same nature as the alimentary
castration to which all the earlier broods during the seasonal
development of the Polistes colony are normally subjected, and
this probably accounts for the absence of any specific effects on
stature and structure and the evident ease with which the volu-
minous parasites are borne and tolerated.
Effects of Castration in Insects 393
In the case of the male Polistes the matter is not so readily
explained, since this sex is not subjected to the two forms of nor-
mal physiological castration just mentioned. But it should be
noted that the effects of stylopization on the secondary sexual
charactersof the male even in Andrena are rarer than in the female
(vide, p. 384), owing to the fact that castration is much less
complete in this sex, as both Perez and Perkins have shown. This
is, no doubt, also the case in Polistes, for the development of the
testes requires much less food than does that of the ovaries, and
the presence of the Xenos probably, therefore, has much less
effect on this sex.
It has long been known that male pupana and adult femah
Xenos are found only in the late summer or fall brood of Polistes
in the brood, namely, which consists of fertile females and males
that are to mate and provide, after hibernation, of individuals of
the former sex, for the formation of new colonies during the en-
suing spring. Brues ('05) captured on May 22 a large over-win
tered female of P. rubiginosus containing a female Xenos ni
grescens that gave birth to a lot of triungulin larvae. Evidently,
therefore, the larvae of the wasp must be infested with tnun-
gulins in the spring, soon after the colony is founded. How come
it then, we are led to ask, that the adult Xenos appear only ir
wasps belonging to the last or autumn broods? If these wasp<
really belong to so late a brood they could not become infested;
unless we suppose that the triungulins hang about the wasps'
nest for a long period before entering the larvae. As this assump-
tion is very improbable, we seem to be forced to the conclusion
that the wasps that bear the Xenos in the late summer really
belong to early broods which have been greatly retarded in their
larval and pupal development. Dodd ('06) and Howard ('08)
have published some interesting observations which show that
the larvae of other insects (Lepidoptera,Formicid3e) parasitized by
chalcidids are greatly retarded in their growth and development.
If this occurs also in Polistes larvae infested with Xenos, as seems
probable, we may be able to account for the facts and understand
how the single generation ofXenos manages to survive till the
following spring to insure the perpetuation of the race in healthy,
394 William Morton Wheeler
incipient colonies of the wasps. The triiingulins are, in all
probability, carried to these colonies by healthy wasps from the
flowers onto which they crawl from their mothers after hibernat-
ing in their hosts.
Since the foregoing paragraphs were written Pierce's fine
monograph of the Strepsiptera has appeared ('09). This work
contains such a full summary of all that has been published on
this remarkable group of insects, together with so much new
matter, that I should have thought it .unnecessary to publish the
preceding pages, but for the fact that they were written for the
purpose of elucidating a problem which Pierce treats only
incidentally. Of the many interestingfacts contained in his paper
I shall cite only a few which have an immediate bearing on the
matters considered above.
The fullest statistics given by Pierce relate to two large colonies
of Polistes annulans infested with Xenos pallidus. These colonies,
which were collected at Rosser, Texas, September 23, together
contained 1553 wasps, 1311 males and 242 females. Of these 266,
or 17.1 per cent were stylopized, 259 being males and only 7
females. The highest number of Xenos observed in a single
wasp was 15, and this occurred in a male specimen! Pierce also
cites some statistics published by Austin (1882) on 50 Polistes
metricus collected at Readville, near Boston, Mass., August 20,
1879. Of these wasps, 14 of which were males and 36 females, 9
or 18 per cent were stylopized (2 males and 7 females).
Pierce figures the abdomen of a male wasp (Leionotus (Odynerus)
annulatus Say) which has the sclerites much distorted as in the
P. metricus shown in Fig. 2. Concerning his specimen, which
contained a female Leionotoxenus hookeri Pierce, he says: "It
seems that in pushing itself out between the segments the parasite
completely split the dorsal tergites of segments three, four and
five and split segment two halfway to the base. The parasite
was located behind segment three." He cites the observations of
Perez on the effects of stylopization in Andrena and adds the
following modifications observed by Crawford in specimens of
Andrena crawfordi infested with Stylops crawfordi:
"i. Puncturation of abdomen less strong, punctures finer and
sparser; especially noted on second segment.
Effects of Castration in Insects 395
"2. In females with male parasites the basal joint of the hind
tarsi is narrower, approaching the shape of the corresponding
joint of the male tarsi ; this joint not noticeably narrowed in female
with female parasites.
"3. Scopa of parasitized female thinner, plumosity shorter, not
so silky.
"4. Out of six males with male parasites two show the second
transverse cubital gone in both wings; one has stubs at each end,
however, in right wing; one has the transverse cubital slightly
interrupted in both wings. Out of about no nonparasitized males
none show any variation.
"5. Out of 38 females with male parasites one has the left
wing with three submarginals, the right wing with two submar-
ginals; one has two submarginals in both wings but right wing
with a stub of the nervure; one has first transverse cubital of the
left wing one-half gone; forty-five nonparasitized females show
no variation.
"None of the other salient alterations found by Perez could
be expected in this species because of the close resemblance of the
two sexes. Andrena crawfordi is a very generalized bee."
Pierce also calls attention to a single parasitized specimen of
A. advanans in his collection, with a spurious nervure in the third
discoidal cell, and believes that parasitism may affect the trachea-
tion of the wings, a modification not observed by Perez.
II GENERAL CONSIDERATIONS.
By employing the word "castration" in a broad sense to mean
any process that interferes with or inhibits the production of
ripe ova or ripe spermatozoa in the gonads of an organism, and
not merely in the concise original meaning as the sudden and
complete extirpation of the gonads, we are enabled to bring to-
gether a number of interesting but hitherto rather scattered facts
which have a bearing on the correlation of the primary and
secondary sexual characters. An adequate consideration of these
facts would go a long way, I believe, towards preparing us for a
profitable study of the recondite problem of sex determination.
396 William Morton Wheeler
Owing to the limits of this paper and to the fact that the depend-
ence of the secondary on the primary sexual characters in verte-
brates has been recently analyzed by several authors, notably by
Herbst ('o.i) and Cunningham ('08), I shall confine my remarks
very largely to the arthropods. Taking the word "castration"
in the broad sense suggested above, we may distinguish:
1. Surgical, or true castration, i. e., the sudden and complete
ablation of the male or female gonads, so that the organism is
deprived of its primary sexual characters, if we do not include
in this term also the gonad-ducts and copulatory organs. This
operation is of the greatest experimental significance, since, when
performed at the proper ontogenetic stage, it has been shown to
lead in many animals to interesting modifications of the secondary
characters of each sex.
2. Physiological castration. Under this head may be included
at least three different forms of inhibition in the development of
the gonads, leading to a failure of the individual to develop its
primary sexual characters, or, in other words, to an inability to
function as a male or a female. This inhibition is brought about
by an insufficient supply of nutriment and appears as the result
of a well-known law, according to which the organism provides
in the first instance for the growth and differentiation of its soma
and develops its gonads on the nourishment in excess of that
required for normal growth in stature and the complete differ-
entiation of the various tissues. The following three forms of"
physiological castration may be distinguished:
A. Alimentary castration. This term was originally given by
Emery ('96) to the suppression of gonadic development through
insufficient feeding of the organism during its larval life.
B. Nutricial castration. This term was first used by Marchal
('97) to designate the maintainance of the gonads in an undeveloped
condition in the adult, owing to the latter's devoting itself to
nursing the brood of other fertile individuals instead of itself
taking on the reproductive function.
C. Phasic castration. I use this term, for lack of a better, to
include all the cases in which the gonads are inhibited in their
development by seasonal or ontogenetic (growth) conditions. This
Effects of Castration in Insects 397
form of castration is not sharply marked off from the two preceding
but may be made to include them, since both alimentary and
nutricial castration can be suspended during the life time of the
individual and normal reproduction supervene.
3. Parasitic castration. This term was first introduced by
Giard ('87, etc.) in a series of studies on Crustacea. It refers to
the suppression or destruction of the gonads by parasites. By
enlarging the scope of Giard's definition we can distinguish two
forms of parasitic castration:
A. Individual parasitic castration, which is induced in certain
organisms when they contain parasites, and
B. Social parasitic castration, which occurs in ants when one
colony in becoming parasitic on a colony of a different species
eliminates the sexual individuals of its host.
A number of illustrations will bring out the fundamental
resemblances between these different methods of suppressing the
reproductive function and the resulting modifications of the
somatic characters of the individual or of their equivalents in
animal societies.
/. Surgical castration.
The pronounced modifications of the secondary sexual charac-
ters observed in vertebrates, especially in birds and mammals,
from which the gonads have been removed during early life, or
in which these organs have become diseased, have led some inves-
tigators to look for corresponding modifications in the secondary
sexual characters of insects subjected to a similar operation.
One observer, Hegner ('08), has succeeded in castrating the
embryos of a chrv'somehd beetle (Calhgrapha multipunctata) by
removing the very young sex-cells as soon as they are segregated
in the protoplasmic accumulation at the posterior pole of the egg
during the formation of the blastoderm. Although Hegner's
experiment, which consisted in pricking the chonon at the pos-
terior pole and allowing the sex-cells to flow out, was successful
to the extent of demonstrating that the embryo may continue its
development after the operation, nothing but a few young larva?
398 . William Morton Wheeler
were obtained. The experiment therefore, throws no light on the
question with which we are here concerned.
Much more important are the results of experiments per-
formed byOudemans, Kellogg, Meisenheimer and Regen in cas-
trating larvae.
Oudemans ('99) was the first to attempt surgical castration in
insects. , He removed one or both gonads from male and female
caterpillars of the gypsy moth (Ocnena dispar) before the last
and second last moults. About one-third of the caterpillars (30
out of 86) survived the operation and produced moths. From a
study of these, the Dutch investigator concluded that castration
has no influence, either on the external appearance, i.e., on the
secondary sexual characters, or on the behavior .of the moths,
since the castrated males copulated, though they had no sperma-
tozoa, and the females, though they had no eggs, nevertheless
stripped from their abdomens the mass of long hairs in which
they normally oviposit. Females castrated only on one side laid
eggs, and three normal females that copula ted with castrated males,
laid eggs which developed parthenogenetically.
Kellogg ('04) succeeded in castrating silk-worm caterpillars
(Bombyx mori) after the second, third and fourth moults by burn-
ing out the gonads with a hot needle. This" method was very
inferior to that employed byOudemans. Not only was the mortal-
ity of the caterpillars greater, to judge from Kellogg's remarks,
but the complete destruction of the gonads was obviously much
less certain. Like Oudemans, he failed to detect any modifica-
tions of the secondary characters of either sex in cases in which
dissection of. the adult moths proved that the gonads had been
completely destroyed.
More recently Meisenheimer ('07) has carried out much more
elaborate experiments than either of his predecessors, on about
600 Ocneria dispar caterpillars, of which 186 yielded imagines.
The smallest caterpillars castrated were between the second and
third moults, and about | cm. long, but he also used those be-
tween the third and fourth and between the fourth and fifth
moults. He was able to remove the gonads even before the second
moult but the larvae were too delicate to survive the operation.
Effects of Castration in Insects 399
Three series of operations were performed: first, the removal of
both gonads; second, the removal of the gonads together with
the gonad-ducts; and third, the transplantation of testes into
female and of ovaries into male caterpillars. The transplantation
of ovaries was more easily performed than that of the testes. In
these cases the transplanted organs not only developed to their
normal size, but the ovaries in some cases even united with the
male vasa deferentia. In one case a single transplanted ovary
united with one of the vasa deferentia, and as the testes of the
opposite side developed, an artificial hermaphrodite was produced.
Meisenheimer describes the results of his operations as follows:
"Oudemans' and Kellogg's experiments established the fact that
the removal of the gonads exerts no influence on the secondary
sexual characters. My results agree with these to the extent that
in my experiments the originally male caterpillar always produced
a male moth, the female caterpillar a female moth. The general
habitus of the respective sex was always perfectly preserved, both
in the form of the body, the structure of the antennae and the color-
ation of the wings, and this was true of the operations, both in
the case of the castrated moths and of the artificial hermaphrodites.
But on examining, in a comparative way, the material obtained,
a' certain effect of the operations seems, nevertheless, to be notice-
able. The moths subjected to the two kinds of operation may be
arranged in series, which in the males vary from dark to light
forms and pass over in the females from a whitish to a darker
color." But, as Meisenheimer observed, there is considerable
color variation in both sexes of normal gypsy moths, and this was
true also of his control series, though he believed the variations
to be greater in those developed from operated caterpillars. The
specimens with transplanted organs, however, showed no greater
modification than those of the castrated series. It is especially
noteworthy that in the cases of transplantation there was no
change in the copulatory or other organs, though these had not
yet developed at the time of operating. Hence, although Meisen-
heimer made artificial hermaphrodites, he did not succeed in
producing artificial gynandromorphs. 4
4 Unfortunately I was unable to secure a copy of the first part of Meisenheimer's final mono-
graph ('09) till after the manuscript of my paper had gone to press. The review here given of
his experiments is, therefore, inadequate.
400 William Morton ff heeler
It will be noticed that the preceding experiments were per-
formed only on holometabolic insects of the order Lepidoptera.
As such experiments on ametabolic insects might be expected to
yield different results, it is interesting to record that Regen ('09,
'10) has recently succeeded in castrating crickets (Grylluscampes-
tns L). In his first paper he gives us little more than an orienta-
tion experiment performed for the sake of determining whether
the insects would survive the operation, but his second contribu-
tion brings ampler and more satisfactory data. In order to perform
the operation he narcotized the crickets with CO 2 . The testes
were removed from 40 males (20 in the second last and 20 in the
last larval instar), and the ovaries were removed from 10 females
in the last instar. These 50 individuals were released in the open
field and each returned to the burrow which it is in the habit-of
occupying throughout its larval life. The operated individuals
were marked by cutting off portions of their wings, and near
their burrows stakes were placed with records of the necessary
data. After the crickets had reached maturity Regen recovered 9
males that had been castrated in the second last, 13 of those cas-
trated in the last larval instar, and 6 females. The insects were
left in their burrows. Ten days later he found that the crickets
had changed burrows and there was a tendency for them to as-
sociate i-n pairs, each consisting of a male and female occupying
a hole in common. Several individuals had migrated to other
parts of the meadow in which Regen experimented, but he suc-
ceeded incapturingand placing in a terrarium 10 males (4 castrated
in the second last and 6 in the last larval instar) and one female
On these specimens he made the following observations:
"l. Nine imagmal males, part of which had been castrated
during the last and part during the second last larval instar,
chirped throughout the remainder of their lives in as lively and
shrill a manner as normal males. Only one of the males, which
had been castrated in the last larval instar, chirped feebly and at
rare intervals.
''2. The behavior of the castrated males towards the females
was the same as that of normal individuals. They enticed the
females with their shrill stridulation and when a female approached,
Effects of Castration in Insects 401
emitted a soft, whirring sound, and tried to affix their sperma-
tophores to her, for
"3. The glands which secrete the spermatophore envelopes
produced these structures up to within a short time of the death
of the crickets and therefore performed their function independ-
ently of the testes.
J
"4. In external appearance the spermatophore envelopes of
castrated males were in all respects like those of normal males
(in some cases they were somewhat smaller), and contained a
white secretion, which was less abundant than in normal sperma-
tophores.
"5. The markings of the anterior wings, or tegmina and the
development of the stridulatory organ showed no modifications.
"6. The females were unable to distinguish between normal
and castrated males. They followed the call of the latter, mounted
their backs and permitted them, as if they were normal males, to
affix their spermatophore envelopes near the genital orifice.
"7. The castrated female behaved like one that had not been
castrated. She thrust her ovipositor into the earth and made
motions like a normal female, so that she -had every appearance
of desiring to oviposit. As time went on this "oviposition"
became abnormal, as the female kept on thrusting her ovipositor
into the earth but only to a slight depth."
Regen assured himself of the completeness of castration in
these crickets by dissection and by examination of the spermato-
phores, which were found to contain no spermatozoa. He also
kept a series of castrated individuals in captivity from the time
of operation, and when these reached maturity they were found
to behave exactly like the individuals that had been permitted to
mature in the field. His experiments, therefore, gave results in
complete harmony with those of Oudemans, Kellogg and Meisen-
heimer. It must be admitted that his insects were all castrated in
rather late stages. He informs us, however, that during the sum-
mer of 1909 he successfully castrated a number of much younger
larvae, measuring only 5 to 8 mm., and that these had grown to a
length of 20 mm., by December 1909 when he wrote his second
paper. At that time the females were readily distinguishable
402 William Morton Wheeler
from the males by their ovipositors. He intends to remove the
spermatophore glands from some of the males of this series and also
from some uncastrated males and to report on the results in a
further publication.
2. Ahmeniary Castration
The best examples of this form of castration are to be found
among the social Hymenoptera, i.e'., among the social wasps,
bees and ants. In these insects the majority of the female larva;
in each colony become what are called workers, because they are
fed on a limited diet, grow very slowly, pupate more or less
prematurely and hence as adults, or imagines are smaller in stat-
ure than the normal females of their respective species. These
workers are also distinguished by other morphological and etho-
logical peculiarities. Owing to their inadequate nourishment as
larvae, their ovaries are, as a rule, in a very rudimentary condition.
Very striking examples of this alimentary castration are seen in
the incipient colonies of ants, while the mother queen is bringing
up her first diminutive broodof workers, in the species of Carebara,
the queens of which are more than 1000 times as large as their
sterile offspring, and in Pheidologeton, in which there is nearly as
great a difference beween the stature of the queen and that of
the -smallest workers. In bumblebees, honey-bees, social wasps .
and most ants this difference is less pronounced, but it is never-
theless perceptible and clearly traceable to larval starvation.
Opinions differ as to whether the other characters peculiar to the
worker forms of these insects are the result of underfeeding, but
it is evident that none of these can be regarded as an approach to
the male type of structure. In other words, notwithstanding the
very decided inhibitory effect of larval starvation on the develop-
ment of the ovaries in the adult workers of the social Hymenop-
tera, the soma does not tend to become like that of the male, but
merely departs to a greater or less degree from that of the female
type. This departure is usually in the direction of greater simpli-
fication and is most pronounced in the ants, the workers of which
are wingless, have a smaller and much simpler thorax and smaller
eyes and ocelli.
Effects of Castration in Insects
403
The social Hymenoptera, however, are not the only insects
which practice alimentary castration. A very interesting case is
also seen in certain aphids of the genus Phylloxera, e.g., in the
Ph. caryae-fallax recently studied by Morgan ('09). The stem-
mother, or fundatrix of this insect makes and inhabits a hollow
gall on hickory leaves. She lays numerous eggs which may give
rise to two kinds of offspring. The eggs first deposited produce
individuals that grow up to form the wingless sexuparas, (Fig. 3 A),
Fig. 3. Large wingless female of Phylloxera caryae-fallax; B and C, dwarf females of same,
drawn to same scale as A. (After T. H. Morgan.)
while the eggs laid later give rise abruptly to very small females,
(Figs. 3$ and C), which Morgan calls "supernumerary or dwarf
females." These he describes as follows: "In the larger galls
as many as 46 eggs may produce the large individuals, and then
the smaller series abruptly begins; while in the smallest galls
only one to three or four or more large individuals are produced
when the small series begins. There seems to be here not a prede-
termined number of large and dwarf females, but the conditions
JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 8, NO. 4.
404 William Morton Wheeler
of life determine when the one kind ceases to be produced and the
other begins. The two types of individuals must, however, be
predetermined by alternative possibilities possessed by each egg.
The supernumerary or dwarf females differ from their large
wingless sister-forms, and from the young of the latter in a num-
ber of points. The shape of the body is entirely different and
resembles that of the sexual male; but it differs from the male
in two important respects; first, the dwarf individuals have a
very long proboscis which in this species is absent in the male;
second, there are no testes within the abdomen as in the males,
where they form a relatively enormous mass. Otherwise the
dwarfs are so similar in external form to the sexual males that
their true nature was uncertain until they were studied in serial
sections. These showed the absence of the testes and the presence
of rudimentary ovaries and ducts resembling those of immature
parthenogenetic females. There was nothing to indicate that
the dwarfs could become sexual females. In fact the latter con-
tain each an enormous egg when they hatch." Morgan believes
that the dwarf females "are destined to a -brief existence, and die
without progeny," and he gives good reasons for supposing that
they owe their origin to inadequate feeding of their parents. In
other words, w.e have here a case of alimentary castration differing
from that of the social Hymenoptera only to the extent that the
mother insect provides her egg with an inadequate amount of
yolk instead of feeding the larva from day to day on an insuffi-
cient amount of food. The resemblance of the dwarf females of
Phylloxera caryae-fallax to the workers of ants and other social
insects is very striking, although it seems not to have been
noticed by Morgan.
Perhaps the well-known "high" and "low" types of male in
many insects, notably of the Scarabaeidas and Lucanidae are to be
regarded as the results of larval feeding. If this is the case, the
low males may present examples of alimentary castration. This
peculiar male dimorphism certainly bears more than a superficial
resemblance to the female dimorphism of the social Hymenoptera.
These may, indeed, be said to have high and low females, which,
like the corresponding forms of the opposite sex in Scarabaeidae,
Effects of Castration in Insects 405
are sometimes connected by intermediates In ants the soldiers
and desmergates represent the intermediate forms. 6
But no one, to my knowledge, has studied the testes of beetles
with dimorphic males, with a view to ascertaining whether these
organs are more imperfectly developed in the low than in the high
individuals. The low males undoubtedly approach the female in
form, and might, therefore, be said to assume the secondary char-
acters of this sex, were it not for the consideration that in a large
number of scarabasid and lucanid species and, genera both sexes
have the same simple form. Thisindicates thatthelowmalesimply
fails to develop its secondary sexual characters and hence returns
to the ancestral type of the species in which these characters were
either very feebly manifested or were altogether absent. G.
Smith ('05*3) has shown that in the Scarabaeidse and Lucanidae,
as well as in certain Crustacea (Tanaidae), "the differentiation into
high and low males within the limits of a species has widely in-
fluenced the progressive differentiation among the different
closely related species of many groups." This is somewhat more
clearly expressed by saying that there are also high and low species
in certain groups, the larger species of certain genera having a
more pronounced male dimorphism than the smaller closely allied
species. This is also true of the sexual dimorphism of female ants,
as is seen in such genera as Solenopsis and Camponotus and among
the genera of the subfamilies Dolichoderinae, Camponotinae and
Mymicinae. It will be shown in the sequel that there is also another
way of accounting for the "high" and "low" forms of some insects.
In this connection, I may briefly consider two cases which, if
correctly reported, would appear to represent a complete loss of the
reproductive organs by alimentary castration carried back into the
early larval or embryonic period. Adlerz ('86) and Miss Bickford
('95) failed to find any traces of ovarian tubules in workers of the
common pavement ant, Tetramorium cespitum. If this negative
observation be correct, the workers of this ant must be regarded as
utterly sexless. In my opinion, however, renewed investigation
6 For a fuller account of the conditions in these insects the reader is referred to my paper on poly-
morphism ('07).
406 William Morton Wheeler
is required to establish the truth of this statement. The other
case is even more doubtful. Silvestri ('06) recently described
Copidosoma truncatellum, a chalcidid which is polyembryonic
and infests the eggs and caterpillars of moths belonging to the
genus Plusia, as possessing two very different larval forms. One
of these he designates as "asexual" and states that it lacks every
trace of the reproductive organs. It is very unlike the ordinary
sexual larva in having a large head, well-developed mandibles and
a very slender nematode-like body, and never lives beyond the
larval stage'. Silvestri believes that it has been developed for the
purpose of breaking down the tissues of the host caterpillar and
of thus rendering them more easily assimilable by the sexual larvae
which alone develop into imagines. The following considerations
seem to me to cast considerable doubt on this interpretation:
First, the asexual larva; figured and described by this investigator
are suspiciously like certain very young ichneumomd larvae, and
as their development is not satisfactorily traced to the same cell-
masses from which the sexual Copidosoma larvae arise, it is not
improbable that the two larval forms really belong to two very
different parasites. In other words, Silvestri's Plusia caterpillars
were probably infested with ichneurnonid in addition to Copido-
soma larvae. Second, I have been unable to find any larvae of the
asexual type in a number of American Plusia gamma caterpillars
which were heavily infested with Copidosoma truncatellum.
Third, as in many species of Chalcididae larvae of Silvestri's sexual
type are able by their own endeavors to break down and assimi-
late the tissues of their host, it seems improbable that a single spe-
cies should have developed a peculiar sexless and moribund larva
for this particular purpose.
J. Nutncial castration
The abortive or rudimental condition of the sexual organs
seen in the cases of alimentary castration may be normally pro-
longed and maintained throughout the adult life of the workers
among the social Hymenoptera, when these insects are compelled
to live on the slender remnant of food that remains to them after
Effects of Castration in Insects 407
they have satiated their queens and the young broods which are
continually hatching from her eggs. Marchal ('97) has called
attention to this condition in the wasps, and it has long been
known to obtain in ants and the social bees, though the causal
connection between the protracted immaturity of the ovaries
in adult workers and their primary function as nurses had not
been sufficiently emphasized. The form of castration which is
thus produced is, however, not necessarily permanent. If the
trophic status of a colony becomes highly favorable, or if the queen
dies, the ovaries of one or of a number of the workers may undergo
active growth and produce eggs capable of normal development.
In such cases the workers may be said to usurper to supplement the
function of the queen, but owing to the fact that the adult insect
cannot modify its external characters, there is no visible differ-
ence between the sterile and fertile workers, except in the size
of the abdomen, and even this may be so slight as to escape
observation. The primary cause of nutricial castration is to be
sought in the instincts of the individual itself, whereas alimentary
castration would seem to be attributable to the instincts of the
individual's living enviroment, i.e., to its nurses. This distinc-
tion, however, is probably more apparent than real, since as I
have suggested in a former paper ('07), it is possible that the
worker larva is from the beginning an organism predisposed to
assimilate only a portion of the nourishment with which it is
provided by its nurses. The growth and development of the
larva obviously does not depend on the amount of food admin-
istered to it but on the character and rate of operation of its
assimilating mechanism. A larva may be very voracious, but
its tissues may be constitutionally unable to appropriate more
than a limited portion of the food which enters its alimentary
tract. The administration of highly assimilable food, as in the
case of the ''royal jelly" which is fed to the larval queen bee, may
be, as I have maintained ('07), primarily for the purpose of
accelerating the development of her ovaries, and the secondary
characters of this insect, which are mostlyof an abortive charac-
ter (smaller sting, shorter wings, smaller hind legs) may be the
result of this development.
408 William Morton Wheeler
Nutricial castration is not confined to the social insects but
occurs also in mammals during the periods of gestation and lac-
tation and in birds during incubation, as the result of a very simi-
lar inability of the organism to expend in reproduction the ener-
gies demanded by the exigencies of the nursing function.
4. Phasic Castration.
The forms of sterility which I include under this term, though
temporary, cannot be sharply distinguished from the cases of
alimentary and nutricial castration, since both of these may be
abolished during ontogenetic development and yield to a fer-
tile phase, as, e.g., when worker ants become gynaecoid and nym-
phal termites become supplemental males and females. We may,
indeed, say that the great majority of animals exhibit alimentary
castration during their embryonic, larval and juvenile stages, but
that this is not universally true is shown by the many examples
of neotenia and paedogenesis scattered through the animal king-
dom. There are, however, several cases of temporary castra-
tion which, though intimately dependent on the trophic condition
of the individual nevertheless do not properly fall in the categories
previously considered. The following may serve as examples:
A. Many hermaphroditic animals are protandric, z. <?., develop
only their male reproductive organs at a very early stage and do
not mature their female reproductive organs till after the testes
are partly or wholly exhausted. Some of the most extreme cases
of this phenomenon are seen in the epicarid Crustacea and in
the singular parasitic worms of the genus Myzostoma. In the
crustacean Danaha the individual becomes a functional male
while it is still a minute and active larva. Later this form at-
taches itself to the abdomen of a crab, loses its limbs, and develops
a long proboscis which penetrates the tissues of its host. The
abundant nutriment thus acquired enables the parasite to grow
rapidly. Its ovaries then begin to enlarge, while the remains of
its testes degenerate and are devoured by phagocytes, and the
creature becomes a female. A very similar condition occurs,
as I showed several years ago ('96) in certain species of Myzo-
Effects of Castration in Insects
409
stoma (e.g., in M. pulvmar von Graff). In these striking exam-
ples the animal is only potentially hermaphroditic, since function-
ally it exhibits seasonal gonochorism through phasic castration
of the ovaries during its youth and of the testes during its adult
stages.
B. Geoffrey Smith ('05 a, '09) has called attention to a very
striking form of phasic castration in decapod Crustacea: "Dur-
ing the breeding season the males of Inachus mauritamcus fall
Fig. 4. Males of Inachus mauritanicus. A small breeding male with swollen chelge; B non-
breeding male, with slender chelae; C, large breeding male with swollen chelse. (After Geoffrey
Smith.)
into three chief categories: Small males with swollen chelae (Fig.
4.^), middle sized males with flattened chelae (B), and large males
with enormously swollen chelae (C). On dissecting specimens
of the first and third categories it is found that the testes occupy
a large part of the thoracic cavity and are full of spermatozoa,
while in the middle-sized males with female-like chelae the tes-
tes appear shrivelled and contain few spermatozoa. These nOi.-
breeding crabs are, in fact, undergoing a period of active growth
and sexual suppression before attaining the final stage of devel-
4io William Morton Wheeler
opment exhibited by the large breeding crabs." This same con-
dition was previously observed by Faxon ('85) in male crayfish
belonging to the American genus Cambarus. Of course, the three
stages distinguished by Smith are separated by moults. Ob-
viously we have here a condition like that observed in many male
fishes, amphibians and birds, which lose their secondary sexual
characters during the seasons when they are not breeding. Smith
regards the phenomenon as "obviously parallel to the 'high and
low dimorphism,' so common in lamelhcorn beetles," but this
is a mistake, as 'Cunningham ('08) has shown, for we are here
confronted with a case of seasonal sexual dimorphism. Nothing
comparable to the condition described above is seen in insects,
for the reason that these animals either do not mature their
gonads till after they have attained their fixed and final imaginal
instar, or if they become sexually mature as larvae or pupae (nep-
tenic and paedogenetic aphids, cecidomyids, chironomids, etc.)
they do not develop beyond this stage. It is not improbable,
however, that insects which live several years in the adult stage
and have seasons of sexual activity alternating with seasons of
infertility, may exhibit great periodical changes in the size and
development of the reproductive organs. I have been unable to
find any observations on this subject in the entomological litera-
ture.
5. Individual Parasitic Castration.
The first zoologist fully to appreciate the importance of para-
sites in suppressing the reproductive function and in incidentally
affecting the somatic characters of their hosts was Giard. He
published some twenty papers ('6g-O2') on a great variety of
cases which he observed not only among animals but also among
plants. The cases to which he devoted most attention were the
decapod Crustacea, especially species of Stenorhynchus, Portunus,
Carcinus, Cancer, Platydnychus, Eupagurus, Palasmon, Gebia and
Hippolyte, which are infested with extraordinary cirriped and
bopyrid parasites of the genera Sacculma, Portunion, Bopyrus,
Probopyrus, etc. Within more recent years these studies have
Effects of Castration in Insects
411
been continued and deepened by Geoffrey Smith ('06, '09) on the
spider crab Inachus mauritanicus infested with the cirriped Sac-
culina neglecta and by Potts ('06, '09) on hermit crabs (Eupa-
gurus meticulosus) infested with the cirriped Peltogaster curva-
tus. A summary of the work of these two authors will not be
out of place here, since they have reached rather definite con-
Fig. 5. Specimens of Inachus mauritanicus to show effects of parasitic Sacculina neglecta. A
normal male; B, normal female; C, male infested with Sacculina (final stage);/), abdomen of infested
female; E, infested male in an. early stage of its modification. (After Geoffrey Smith.)
elusions not without a bearing on the various cases of parasitic
castration in insects and other organisms to which I shall have
occasion to refer.
According to Geoffrey Smith ('09) the abdomen of the normal
male of Inachus mauritanicus "is small and bears a pair of copu-
la tory styles, while the chelipedes are long and swollen (Fig. 5^).
In the female (Fig. 5$) the abdomen is much larger and trough-
412 William Morton Wheeler
shaped, and carries four pairs of ovigerous appendages; the che-
lae are small and narrow.
"Now it is found that in about 70 per cent of males infected
with Sacculina the Body takes on to varying degrees the female
characters, the abdomen becoming broad as in the female, with
a tendency to develop the ovigerous appendages, while the che-
lae become reduced ( Fig. 5(7). This assumption of the female
characteristics by the' male under the influence of the parasite
may be so perfect that the abdomen and chelae become typically
female in dimensions, while the abdomen develops not only the
copulatory styles typical of the male, but also the four pairs of
ovigerous appendages typical^ of the female. The parasitized
females, on the other hand, though they may show a degenerate
condition of the ovigerous appendages ( Fig. 5-D), never develop
a single positively male characteristic. On dissecting crabs of
these varying categories it is found that the generative organs
are in varying conditions of degeneration and disintegration.
"The most remarkable fact in this history is the subsequent
behavior of males which have assumed perfect female external
characters, if the Sacculina drops off and the crabs recover- from
the disease. It is found that under these circumstances these
males may regenerate from the remains of their gonad a perfect
hermaphrodite gland, capable of producing mature ova and sper-
matozoa. The females appear quite incapable, on the other
hand, of producing the male primary elements of sex on recovery
any more than they can produce the secondary."
The following account is quoted from Pott's summary ('09)
of his own studies on the modifications induced in Eupagurus by
Peltogaster and of Smith's observations:
"The difference between the sexes of Eupagurus is shown only
in a couple of external characters, the position of the generative
apertures (as in all Decapods) and the character of the abdominal
appendages. The abdomen of the hermit crab is furnished on one
side only with a few appendages, insignificant, but with definite
functions. It is in the female that we see the full development of
the appendage as a swimmeret with two equal branches, the inner
one provided with long hairs affording a secure anchorage for
Effects of Castration in Insects 413
countless eggs while the outer one is of equal size in both sexes,
and in both by its paddle-movement maintains respiration cur-
rents in the shell. No use has been found for the outer branch in
the male and so has become quite rudimentary, but the effect of
the parasite Peltogaster is to stimulate the growth of this rudi-
ment. There isof course great variability of response to.this stimu-
lus but those individuals which experience the maximum amount
of change possess swimmerets exactly similar to those of a mature
female, even in the assumption of the curious branched or barbed
hairs which in this case can never bear eggs. As m the spider
crabs so here, the female appeared incapable of the reverse change,
and the large number of hermit crabs with typical female append-
ages and sealed genital apertures are undoubtedly to be regarded
in part as modified males.
"A protest will conceivably be uttered against the attribution
of a special sexual significance to the development of typical
swimmerets in the male in both spider crabs and hermit crabs. It
is of course well known that in the larval stages of these Crustacea
biramous abdominal appendages are found in both sexes to be
subsequently reduced or lost in the male. Lest this, then, be
deemed a happy opportunity for applying the term "reversion"
to this phenomenon I hasten once more to point out that when the
male develops biramous abdominal swimmerets they are of the
type associated with female maturity, and that the specialized
nature of their nursing-hairs cannot well be associated with ances-
tral conditions.
"Both Sacculina and Peltogaster inflict sterility upon their host
and apparently entire abortion of the gonad generally is the final
consequence. On the external appearance of the parasite the eggs
of the female shrink through absorption of their yolk and the
formation of spermatozoa is after a time suspended in the male.
The testis of the spider crab dwindles and disappears without
undergoing any particular histological change; but in the hermit
crab it is curious to note the presence of large cells with large
nucleus and abundant protoplasm in sections of the testis. These
instantly suggest ova in their appearance and call to mind the
instances of the occurrence of such cy tological elements as a nor-
414 William Morton Wheeler
mal experience in the testes of many animals. In sand-hoppers
(Orchestia) to quote a well-known case (and there are many others
in the Crustacea) spermatozoa are produced in the anterior part
of the young testis while posteriorly the whole space is occupied
by two or three large ova fytde Boulenger '08).
"The particular interest of the phenomenon in this case is its
association with a definite cause, that is, parasitism. We are also
able to come to some conclusion as to the degree in which such a
condition can be called true hermaphroditism. Some striking
evidence is offered by spider crabs which were once infected by
Sacculina but which have outlived their parasite and recovered
from its influence. Such crabs occur in nature in fair frequency
and the only reminder of their former condition is the chitinous
ring on the abdomen which surrounded the peduncle of the para-
site. After the death of the external part of the Sacculina the root
system may continue .to exist in the host and it is only when this
has disintegrated and been absorbed that regeneration of the
gonads becomes rapid, for the still living roots repress the devel-
opment of the sexual organs as effectually as the living parasite.
A few crabs however were found in which the gonads had again
attained full size and maturity. One was a female with a well-
developed ovary and four were males only slightly modified ex-
ternally, with glands producing large quantities of spermatozoa.
The remaining four cases were remarkable for the crabs showed
with a complete external hermaphroditism the corresponding
gonads. In all fo ( ur animals the reproductive gland consisted of a
male part with ripe spermatozoa, and a female division with large
pigmented ova. The ducts were usually absent, but one individ-
ual possesssed both vasa deferentia and oviducts. The sequel to
these observations is given by the experimental evidence which
Smith then obtained. It was attempted to destroy the parasite
by removing the external part and the crabs so freed were kept
under comfortable conditions for several months and the few
survivors then killed. Regeneration had obviously occurred to a
considerable extent, but the gonads were nearly always unisexual.
In one individual alone, which was externally a hermaphrodite
there was a gonad similar to those just described. In spite of the
Effects of Castration in Insects 415
comparatiyely small number of cases with fully formed herma-
phrodite glands we are not going too far in definitely asserting a
connection between their occurrence and parasitic influence, for
bisexual gonads have to my knowledge never been met with in
Decapod Crustacea under normal conditions. 6 But it thus ap-
pears that the curious condition in the hermit crab is an incipient
stage corresponding to the perfect hermaphroditism of the "re-
covered" spider crabs, and if the action of the parasite in absorb-
ing surplus nutrition were withdrawn the young ova in the testis
of the hermit crab would become large and pigmented like those in
the spider crab.
"These two cases have been described at some length as ex-
amples of extreme modification. In other Decapod Crustacea
which are infected by the same parasite an effect is observable
which is similar in kind but not in degree. The common shore
crab of England (Carcinus) is commonly afflicted (if affliction it
be) by Sacculina. Here again the male undergoes modification
while the reverse change never occurs in the female. The narrow
abdomen of the male is often exchanged at the moult after infec-
tion for one much broader but never attaining the full female
width. One may look in vain, however, for any reduction of the .
copulatory styles or for the appearance of the smallest rudiments
of swimmerets. The closure of the genital apertures nearly al-
ways follows parasitic attack in spider crab and hermit crab; but
they never become blocked up in shore crabs with Sacculina. Yet
the external change is apparently greater than that produced in the
reproductive glands. Dissection in every parasitized male showed
vasa deferentia of the characteristic milky white color due to
countless masses of spermatophores all packed with spermatozoa.
The testes though reduced, then, always remain in reproductive
activity. The parasites which infect sp'der crab and shore crab
are practically identical and presumably exert a very similar stim-
ulus yet the results are markedly different. It is obviously the
host which offers a different reaction in the two cases. In another
c In a footnote Potts states that "Caiman in the recently appeared volume Crustacea of Ray Lan-
kester's Treatise on Zoology refers to the unpublished observations of Wollebackon normal hermaphrod-
it'sm in certain deep-water Decapoda."
416 William Morton Wheeler
crab (Eriphia) examined by Smith there was infection both by
Sacculma and by a parasitic Isopod crustacean. Here the nature
of the parasite governs the result, and crabs with Sacculina alone
never showed the least trace of modification, while changes closely
similar to those described above occurred in those which har-
boured the Isopod."
Geoffrey Smith ('05 b} has also described parasitic castration
in Inachus dorsettensis by a sporozoon (Aggregata inachi) which
lives in the intestine of the crab and induces modifications not
unlike those induced by Sacculina. Smith says that of fifty males
of I. dorsettensis examined, "seven specimens were clearly dis-
tinguished by having the flat chelae characteristic of the females,
while the abdomen was much broader than is the case in normal
males of a corresponding size, thus converging on the female con-
dition. In one specimen there was present on the under side of the
abdomen a pair of swimmerets which are characteristic of the
female, these appendages being altogether absent in the normal
males." Dissection of these crabs showed the intestine "to be
covered with cysts of Aggregata inachi, the body cavity was also
full of liberated sporozoites, the haemolymph having a milky ap-
pearance due to the crowded presence of these bodies. The testes
were in all cases disintegrated, only the vesiculae seminales remain-
ing. Two modified males were also found to contain the cysts of
Aggregata inachi, but in none of these males were there larger
quantities of sporozoites in the haemolymph, so that it appears
that the hermaphrodite external characters are assumed by the in-
fected male at the moult which follows the liberation of a large
quantity of sporozoites." Smith made no observations on the
infected female Inachus, as this sex is much rarer than the male.
The foregoing examples of parasitic castration in Crustacea have
been reviewed at some length, because they show the phenomenon
in its most striking manifestation. Giard as early as 1888 (/>)
published a long list of other animals and plants known to becas>-
trated by what he calls "gonotomic" parasites. The most inter-
esting examples, apart from Andrena and the Crustacea just con-
sidered, are the castration of the nemertean Lineus obscurus by the
orthonectidlntoshialinea,of the planarian Leptoplana tremellaris
Effects of Castration in Insects 417
by Intoshia kefersteini, of the brittle, star Amphmra squamata by
theorthonectid Rhopalura giardiand by a copepod (Fewkes'88),
of thesnailsof the genera Paludina, Lymrieea and Planorbis by dis-
tome sporocysts (Distomum militare, retusum, etc.), of the crus-
tacean Cyclops tenuirostris by larval distomes (Herrick '83),
of the bumble bees (Bombus) by the extraordinary nematode
Sphaerularia bombi, and of the males of various North American
squirrels and chipmunks (Tamias lysteri, Sciurus hudsonius and
leucotis) by the bot-fly Cuterebra emasculator as described by
Fitch ('59), Riley and Howard ('89) and Osborn ('96). Among
plants Giard cites the castration of the rig by Blastophaga
gGossorum, of Melandryum album (Lychnis dioica) by Ustilago
antherarum and various grasses by smuts, ergots, rusts, etc.
The case of Melandryum and Ustilago which was repeatedly
studied by Giard('69, 'S/a, '88J, '890) bears a curious resemblance
to that of the male crab infested with Sacculina. The Melan-
dryum is "normally dioecious. The young flower is hermaphro-
dite but in certain individuals the ovaries abort, in others the
stamens remain rudimentary. When the parasitic fungus develops
on a male plant, it fructifies in the stamens, but when it falls on
a female plant, it seems at first as though it could not fructify
and that the infested plant must profit accordingly. But this is
not the case, for the plant develops its rudimentary stamens
completely in order to permit the fructification of the parasite,
just as the male Stenorhynchus enlarges its abdomen in order
to protect the Sacculina fraissei."
Castration frequently occurs in plants through petalody, or
petalomama, i. e. the conversion of stamens or carpels into
petals, producing the well-known "double" flowers. Molliard
('oi) has produced petalody experimentally in Scabiosa colum-
baria by artificially infecting the plant with the nematode
Heterodera radicicola. And this investigator, Meehan ('oo),
Giard ('02) and Cramer ('07) cite a number of observations
which indicate that petalody is often the result of infection of a
plant with root-fungi. Veuillemin ('07) has observed in Lonicera
infested with aphids a suppression of the carpels and a distinct
androgeny of a certain number of the flowers.
41 8 Willidm Morton Wheeler
Instead of stopping to review the various examples of parasitic
castration cited by Giard in his paper of 1888, and in many of
his later publications, it will be preferable to describe as briefly
as possible a number of selected examples, especially some that
have come to light more recently among insects. The stylopized
Polistes and Andrenas, having been adequately described in the
first part of this paper, will be omitted.
Grassi and Sandias ('93) describe a remarkable case of parasitic
castration in termites. They find that worker and soldier ter-
mites have the intestinal cascum, which occupies much of the ab-
dominal cavity, distended with enormous numbers of parasitic
Protozoa belonging both to the Ciliata (Dinonympha, Pyrsonym-
pha, Trichonympha) and to the Gregarinida. The Ciliata have
been studied by several authors, notably by Leidy ('77, '81),
Grassi ('85), Kent ('85), Porter ('97), and Dodd ('06). In ter-
mites infested with these parasites the reproductive organs, both
male and female, remain small and undeveloped, apparently as
the result of the pressure exerted on them by the distension of
the caecum. The parasites are absent in the very young termites
and in the sexual forms, which are fed on saliva. Grassi and Sand-
ias infer that the Protozoa must either be killed off or, at any
rate, prevented from living and growing in the alimentary tract
of saliva-fed individuals. These investigators are inclined, there-
fore, with some reservations, to regard the development of the
two sterile castes in termites as the result of infection with pro-
tozoan parasites. This infection is, of course, readily brought
about as the workers and soldiers are not fed on saliva like the
sexual forms but on dead wood and on the faeces of individuals
belonging to the same castes.
The researches of Grassi and Sandias have received a certain
amount of confirmation from Brunelli ('05), who finds that queens
of Calotermes flavicollis and Termes lucifugus sometimes become
infested with the parasitic Protozoa, and that when this happens
the young oocytes in their ovaries degenerate. Calotermes queens
are more susceptible to this form of castration then the queens of
Termes. Brunelli explains the winged soldier observed by Grassi
and Silvestri's ('03) 48 workers of Microcerotermes struncki with
Effects of Castration in Insects
419
well-developed reproductive organs (40 females and 8 males),
as being instances of fertility brought about by a disappearance
of the Protozoa through some unknown cause. Such fertile
soldiers and workers would be comparable to the "recovered"
spider crabs above described, except that there is no tendency
towards hermaphroditism.
It is not altogether improbable that the high and low males
among the Scarabaeidae, Lucanidas and Forficulidae are produced
Fig 6. A, normal worker of Pheidole commutata; B and C mermithergate of same in dorsal
and 1 teral view.
in some such manner as the workers and soldiers of termites. It
is certainly suggestive that all three of these families of insects
live on decomposing vegetable substances and in situations where
they become very readily infected with gregarines. Giard ('94^)
has given good reasons for supposing that the high and low males
of Forficula, which were made the basis of a statistical study by
Bateson (92), are produced by differences in the number of
gregarines they harbor in their alimentary tract. The French
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 8, No. 4.
42O William Morton Wheeler
observer says: "It is, indeed, possible to predict from the length
of its forceps whether or not a male Forficula possesses gregarines
and whether these are present in greater or lesser numbers. Since
these parasites produce a diminution of a secondary sexual char-
acter, that is, the length of the forceps, without bringing about
absolute sterility (complete castration being exceptional), it not
infrequently happens and this is the case both on the beaches
of Wimereux and on the Fame Islands that the individuals with
short forceps, namely, those containing parasites, are more nu-
merous than the individuals with long forceps." Giard is inclined
to believe that similar conditions may obtain in such beetles as
Xylotrypes gideon, Oryctes nasicornis and other Scarabaeidas
with high and low males. The low males of these beetles, however,
are not to be regarded as having acquired female characters, but
as having lost the male characters, so that, as Giard remarks, the
"infested individuals are generally paedomorphic as compared
with the normal form."
In two of my former papers ('oij '07) I described a peculiar
case of parasitism in a Texan ant, Pheidole commutata. The
larvae of this insect are occasionally infected with nematodes of
the genus Mermis and develop into peculiar forms, which I have
called mermithergates (Figs. 65 and 6C). These are much larger
than the normal workers (Fig. 6^), which they nevertheless resem-
ble in the structure and small sizeofthehead, although they possess
small ocelli and in this respect resemble the queens. In thoracic
structure they approach the soldier form while the gaster is enor-
mously distended with Mermis and retains scarcely any vestiges
of the fat-body, reproductive organs and other viscera. The
behavior of these parasitized individuals is also peculiar, since
they never excavate the soil, nor care for the brood like the nor-
mal workers, but run about in a state of chronic hunger, begging
food from their uninfested nest-mates. Emery ('90, '04) has re-
corded the occurrence of mermithergates in quite a series of neo-
tropical ants, including Pheidole absurda and several Ponerinae
of the genera Odontomachus, Neoponera, Ectatomma, Pachy-
condyla and Paraponera.
In the cases described by Emery and myself only the worker
Effects of Castration in Insects 421
formswere infested and modified by the Mermis, but Mrazek ('08)
has recently shown that the virgin queens of the European Lasius
alienus may become infested with this worm and that when this
occurs the insects develop abnormally small wings (Fig. jB).
These individuals, or mermithogynes, as Mrazek calls them, have
been seen by other investigators and described as brachi/pterous
to distinguish them from the normal macropterous individuals of
the species.
After seeing Mrazek's paper I examined a small collection of
seven brachypterous and as many macropterous females of La-
sius neoniger (a form closely related to alienus) which I had
taken from a single colony near Manitou, Colorado, August 9,
1903. Three of the short winged individuals were dissected and
each was found to contain a large coiled Mermis, 53 to 55 mm. long,
which filled out the whole abdomen, so that in the living indi-
viduals there could have been little left of the reproductive organs
and other viscera. There is nothing unusual in these females
except the small size of their wings, which measure only 6 to
6.5 mm. in length, whereas those of normal L. neoniger females
measure 10 to ir mm. These observations show that the queens
of our American Lasii may be affected by Mermis in exactly the
same manner as the queens of the related European species.
The species of Mermis are not, however, the only known gono-
tomic nematodes. A much more extraordinary form is Sphaeru-
laris bombi, which has been known ever since the days of Reaumur
(1742) to produce sterility in the hibernating queens of bumble-
bees. According to Leuckart ('87), who has written the best
and apparently also the most recent account of Sphserularia,
infested bees are sometimes found, "which have not a single ma-
ture egg in their ovaries. Structurally these organs are perfectly
developed and have ova in the blind ends of their ovarioles, but
ripe eggs are lacking. In other specimens one may find in addi-
tion to the young, also some ova of perfectly normal dimensions."
He says that he has "never seen an infested queen which had the
ovarioles as uniformly and richly provided with eggs as are the
ovaries of healthy bumble-bees at the same season. As a rule,
one finds only a few eggs, sometimes only a single egg." These
422
William Morton Wheeler
bees are therefore unable to found colonies, according to Schnei-
der and Leuckart. They keep flying about till late in June and
then die, whereas uninfested queens have started their colonies
and no longer fly at large after the beginning or middle of May.
Sphserularia occurs only in the queens, and has never been found
in those that have become mothers of colonies. It would be in-
teresting to know whether the colony-founding instincts of
Fig. 7. A, normal female of Lasius alien us; B, mermithogyne of same species (After Mrazek.)
infested queens show the same tendency to atrophy as the ovaries.
As the bees become infected in their imaginal instar, apparently
while seeking their winter quarters, the parasites can produce
no modifications in the external characters.
The Lasius mermithogynes described above recall some ob-
servations of Kiinckel d'Herculais ('94) on Algerian grass-hoppers
(Stauronotus maroccanus and other species) infested with flies
of the genus Sarcophaga. The maggots of the flies are entopara-
sitic, devouring the fat-body, and, according to Kiinckel d'Her-
culais, also absorbing the oxygen dissolved in the blood-plasma of
Effects of Castration in Insects 423
their hosts. The results are an atrophy of the reproductive organs
(parasitic castration) and a weakening of the wing-muscles, so
that the grasshoppers have a disinclination to fly. For this latter
condition, which is described as a "kind of rhachitis, " Kiinckel
d'Herculais suggests the name "aptenia." Like the brachyptery
of the Lasius mermithogynes, it points to an intimate correlation
between the development of the reproductive organs and the
wings, a correlation which is also clearly demonstrated in most
insects by the coincident maturation of the former and full devel-
opment of the latter organs, at the beginning of theimagmal ins tar.
The extensive literature on entoparasitic Diptera and Hymen-
optera, if carefully searched, would probably yield a number of
accounts of parasitic castration. Pantel ('09), in an important
paper, distinguishes both direct and indirect parasitic castration
as the result of the infestation of lepidopteran larvae with the
larvae of tachmid flies. In the former case the fly larvae live in
the testes of the lepidopteron and destroy the gonadic elements
directly. In the latter the gonads suffer atrophy through the
action of the parasites on the other viscera. The only cases I
have found in which the host shows a modification of its external
sexual characters as the result of such castration, are the homoptera
Typhlocyba hippocastani and douglasi, which are described by
Giard (-'896, '89^) as being infested with a dryinid hymenopteron,
Aphelopus melaleucus and a pipunculid dipteron, Chalarus
(Ateloneura) spuna. The females of both species of Typhlocyba,
when castrated by Aphelopus, have the ovipositor much reduced; the
Chalarus alone seems to have less effect on this organ. The penis
of the male T. douglasi islittle modified by either of the parasites,
but" in T. hippocastani infested with Chalarus, this organ shows a
decided reduction in size and simplification of structure so that
the specific characters become profoundly modified. None of
these modifications, however, indicates any tendency to take on
the characters of the opposite sex.
6. Social Parasitic Castration
This category is not sharply marked off from the preceding,
for if we define it as including those cases among social insects
424 William Morton Wheeler
in which the individuals that represent the reproductive organs
(i.e., the males and queens) of the colony considered as an organ-
ism of a higher order, are castrated by parasites, we should perhaps
include also the Lasius colonies containing mermithogynes and
the queens of Bombus infested with Sphamilaria described in the
foregoing paragraphs. But in these cases it is merely prospective
colonies, so to speak, which are castrated, since neither the mer-
mithogynes nor the parasitized Bombus queens have as yet be-
come mothers of colonies. For this reason I have treated them
as cases of individual parasitic castration. Here belongs also
the production of pseudogynes in Formica colonies infested with
the peculiar myrmecophilous beetle? of the staphylinid tribe
Lomechusini (Lomechusaand Xenodusa) which I have considered
at length in a former paper ('07). These beetles tend to sup-
press the development of the annual brood of virgin queens since
the worker ants of parasitized colonies either neglect the queen
larva; or endeavor to convert them into workers, after the period
during which this change can be successfully accomplished has
passed. The results of this behavior is the production of the non-
viable pseudogynes and the gradual degeneration of the colony.
In this case also the colony is not castrated, but the mothers of
prospective colonies may be said to suffer from misapplied alimen-
tary castration.
Leaving all these cases out of account we have left only those
in which a parasitic colony of insects prevents the development
of or destroys the fertile sexual individuals of the host colony in
which it lives. As parasites of this type I may mention the vari-
ous slave-making ants (Formica sanguinea andPolyergus rufescens
and their various varieties and subspecies), the temporary social
parasites (Formica rufa, ex^ecta, exsectoides, etc.) and the perma-
nent social parasites of the genera Anergates, Wheeleriella, Epi-
pheidole, Sympheidole and Eposcus. There are other social para-
sites that do not destroy the reproductive individuals of the host
colony, for example, the bees of the genus Psithyrus, which live
in the nests of bumble-bees, and among ants such species as Lep-
to thorax emersoni, Formicoxenus nitidulus and Harpagoxenug
sublevis. Stillother ants, such as the species of Strongylognathu S)
Effects of Castration in Insects 425
do not destroy the queen of their host colony (Tetramorium ces-
pitum), but since the workers, of this colony prefer to rear the small
sexual forms of the parasites instead of their own bulky males
and females, the development of future colonies of the host
species is rendered impo ssible and we have here again a case of
prospective social castration.
The conclusion which we reach after marshaling this long
series of illustrations of the various forms of castration is that
among insects the only case in which destruction or inhibition of
the reproductive function clearly results in any mpdifications of
the secondary sexual characters comparable to the modifications
observed in vertebrates under like conditions, is that of the sty-
lopized andrenine bees as described by Perez. In all the other
cases extirpation of or injury to the gonads may indeed result in
modifications of the somatic or secondary sexual characters, but
the latter do not take on the peculiarities of the opposite sex.
The most striking illustrations of the truth of this statement are
the insects that have been surgically castrated. These show that
the secondary sexual characters must be so independently and so
immovably predetermined and at so early a period in the onto-
geny that complete extirpation of the gonads during prepupal
life fails to produce the slightest curtailment or modification
either in the secondary sexual characters or in the sexual instincts
of the adult insect. This conclusion renders it imperative to rein-
vestigate the cases of stylopization in the andrenine bees for the
purpose of ascertaining whether Perez's interpretation is the only
one which they will yield, especially since it has been shown in
the first part of this paper that the study of stylopization in
Polistes leads to a very different view and one in complete harmony
with the other cases of castration in insects.
It is interesting to note that castrated Crustacea, to judge from
the observations of Giard, Geoffrey Smith, and Potts, show modi-
fications like those of castrated vertebrates and not like those of
the insects. This is in all probability due to the fact that the devel-
opment of the primary and secondary sexual characters is grad-
ual and continuous in the Crustacea and vertebrates, whereas
both these characters in insects are arrested in their develop-
426 William Morton Wheeler
ment and remain unaffected by the surrounding processes of
growth and differentiation till the imaginal stage is attained. In
holometabolic insects the secondary sexual characters are, of
course, segregated in the imaginal discs, or histoblasts, and even
in hemimetabolic and ametabolic insects there must be a similar
isolation of the cell-materials which will produce the somatic
sexual peculiarities of the adult.
The opinion 'here advocated, namely, that in insects the pri-
mary and secondary characters are .very loosely correlated dur-
ing ontogenetic development: or in a very different manner from
what they are in vertebrates or even in the Crustacea, receives
indirect support from two interesting classes of facts. One of these
classes comprises the anomalies known as gynandromorphs,which,
though always rare, are nevertheless much more frequently found
among insects than among any other animals. These anomalies
consist in combinations of male and female somatic characters
in the same individual, usually in such a manner that the two lat-
eral halves or the anterior and posterior portions of the body are
of different sexes. In the former combination the reproductive
organs may be hermaphroditic and correspond with the sex of the
halves of the body in which they lie, but this is not always the
case, and in anteroposterior, or frontal, or in mosaic,or decus-
sating gynandromorphs, which exhibit an irregular mingling of the
the sexual characters, the goriads may nevertheless be unisexual.
Herbst ('01) and Driesch (07) have emphasized the obvious
inference that these various arrangements of the male and female
characters cannot owe their origin to internal secretions, or
hormones, and indeed all those who have speculated on the ori-
gin of these anomalies are unanimous in holding that they must
arise either from peculiarities in the structure of the egg or from
irregularities in its fertilitation or early cleavage stages at the
very latest. Among recent speculations on the origin of gynan-
dromorphism those of Boveri ('02) and Morgan ('05, '09) may be
mentioned. Boveri believes that the gynandromorph arises from
an egg which has segmented prematurely, so that the male pro-
nucleus unites with one of the cleavage nuclei. Morgan is of the
opinion "that the results may be due to two (or more) spermaro-
Effects of Castration in Insects 427
zoa entering the same egg, one only futing with the egg nucleus
and the other not uniting, but developing without combining with
any parts of the egg nucleus." These hypotheses have no very
cogent facts to support them and I fail to see how they have any
advantage over the hypothesis which was advanced by Donhof
as long ago as 1860, to the effect that the gynandromorph arises
from the fusion of two eggs, only one of which, in the case of
the honeybee, is fertilized. In its original form Donhof's hypothe-
sis is incomplete, but I believe that its plausibility is increased by
addition of the following considerations. We may assume with
Beard ('02), von Lenhossek ('03), Reuter ('07), Morgan ('09)
and others that the gonochoristic Metazoa produce two kinds of
eggs, male and female, which may or may not differ in size but
differ in sex even as oocytes. Now we know from zur Strassen's
researches on Ascaris ('98) that two eggs may fuse and neverthe-
less give rise to a single embryo of perfectly normal structure
though of twice the normal size. In Ascaris the fusion occurs
after the oocytes have reached their full growth, but a fusion of
younger oocytes would be, in all probability, not only more readily
accomplished but lead to the formation of a single embryo of the
normal size. The structure of the ovanoles of insects indicates
that it would be a very easy matter for two young oocytes to be-
come enclosed in the same follicle, too easy, indeed, to accord, at
first glance, with the fact that gynandromorphs are such rare
anomalies. But if two female or two male oocytes fused no gy-
nan dromorph would result, and the chances of either of these fu-
sions of like oocytes occurring would be quite as great as that of
two oocytes of opposite sex. If this be the way in which gynandro-
morphs arise, we should have to explain the occurrence of the lateral
type of the anomaly by supposing that the plane of fusion of the
two eggs be omes the median sagittal plane of the future insect,
whereas in the frontal type this plane would be transverse to the
longitudinal axis. Finally, in the mixed and decussating types
we should have to suppose that the male and female egg-mate-
rials are mixed or interpenetrate one another toa variable degree.
The hypothesis here sketched has the advantage of permitting of
some slight cytological verification, for microscopic examination
428 William Morton Wheeler
of the ovarioles of a large number of Lepidoptera, which seem to
present the anomaly in question more frequently than other
insects, might re veal an occasional inclusion of two oocytes in the
same follicle or even various stages in their fusion. Or if hives
are ever again found like the famous Eugster hive, in which so
many gynandromorphous bees were produced, the cytologist
will have an opportunity to test the hypothesis here advocated by
a careful examination of the ovarioles of the queen.
But no matter what view we hold in regard to the origin of
gynandromorphs, we are compelled to admit that they demon-
strate the very early and rigid determination of the secondary
sexual characters, the possibility of their complete development
even when the gonads of the corresponding sex are lacking and
their independence of internal secretions. To this extent they con-
firm the results obtained by Oudemans, Kellogg, Meisenheimer
and Regen in their castration experiments. Indirectly they indi-
cate that the insect egg not only has its primary sexual
characters determined long before fertilization and independently
of the later nuclear or chromosomal phenomena, but that even
the secondary sexual characters are in some manner also prede-
termined at this early stage. Where great differences of stature
are secondary sexual characters, as in phylloxerans, some aphids
and rotifers, we find corresponding differences in the size of the
male and female oocytes. This is, of course, quite in harmony
with tbe remarkable predetermination of the embryonic regions
of the insect egg. Long ago Hallez ('86) and I ('89, '93) showed
that in many insect eggs the regions corresponding to the ventral
and dorsal, right and left, and cephalic and caudal portions of
the embryo are clearly established long before the maturation
divisions.
The second class of cases, which indicate that the primary
and secondary sexual characters of insects may develop indepen-
dently of one another, are found among certain species of ants,
the males of which, though developing gonads and external geni-
talia of the usual type, have nevertheless become decidedly femi-
nine in their secondary sexual characters. That this condition is
an expression of degeneration seems to be indicated by the fact
Effects of Castration in Insects
429
that it occurs only in parasitic species of the genera Anergates,
Formicoxenus and Symmyrmica or in species like those of the gen-
era Cardiocondyla,Technomyrmex and Ponera, which form small,
scattered colonies, often with a tendency to lead a secluded or
subterranean life. In the three parasitic genera the males are
always wingless and resemble the females and workers in the struc-
ture of their bodies. The resemblance to the worker is very great
Fig. 8. A, winged male of Ponera coarctata in profile; B, winged male of P. eduardi; C,
subergatomorphic male of the same species; D, ergatomorphic male of P. punctatissima (After
Emery.)
m the case of Formicoxenus. In Cardiocondyla and Ponera we
have a number of species whose males show a similar approxima-
tion to the worker and female type, and in one species of the latter
genus, P. punctatissima, shown in the accompanying figure (Fig.
%D) the male is indistinguishable from the worker except in the
structure of the genitalia. We have here, therefore, a true inver-
sion of the male, so far as its secondary sexual characters are con-
430 William Morton Wheeler
cerned, apparently as an adaptation to ethological requirements,
although the primary sexual characters have remained unaffected.
If it be true that the rudiments of the secondary sexual char-
acters are set aside so early in the development of insects and re-
main uninfluenced by the internal secretions, we can understand
why these characters exhibit no modification in cases of surgical
castration and why the modifications induced by alimentary,
nutricial and parasitic castration bear the aspect of inhibitions
or retardations of growth. Normal imaginal development in
insects, as is well known, depends on the amount of food accumu-
lated during larval life and stored up in the fat-body. In insects
surgically castrated during their younger stages there is nothing
to hinder the accumulation of this reserve material, and all the
imaginal characters, including the secondary sexual characters,
are thereby enabled to develop normally and completely. But
m insects that have been underfed or are infested with parasites
the reserve materials are either prevented from accumulating or
are consumed, so that the imago may have great difficulty in de-
veloping its imaginal characters. It is not surprising that under
such conditions the secondary characters are more or less reduced
or aborted, as we see in the forceps of parasitized Forficula males,
the thoracic and cephalic horns of male Scarabaeidae, the mandi-
bles of male Lucamdse, the wings of female Lasii, and many of
the other cases cited above. There is simply not enough nutri-
ment to permit of the full growth of the characters under consid-
eration. Their modification, therefore, is readily explained in
insects as due to malnutrition and we are not compelled to invoke
the internal secretions, or hormones, which play such an impor-
tant and interesting role in the sexual physiology of vertebrates.
Effects of Castration in Insects 431
BIBLIOGRAPHY
ADLERZ, G. '86 Myrmecologiska Studier. II Svenska Myror och deras Lef-
nadsforhallanden. Bih. till K. Svenska Vet.-Akad. Handl. xi, 18,
1886, pp. 1-329, pis. l-vii.
AUSTIN, E. P. '82 Collecting Stylopidae. Journ. Bost. Zool. Soc., vol. i, pp'
12-13, 1882.
BATESON, W. '92 On some cases of vanation in secondary sexual charac-
ters statistically examined. Proc. Zool,, Soc. London, 1892, p. 585.
BEARD, JOHN '02 The determination of sex in animal development. Zool.
Jahrb., Abth. f. Anat., xvi, 1902, pp. 703-764, I pi., 3 figs.
BICKFORD, E. E. '95 Ueber die Morphologic und Physiologie der Ovarien der
Ameisen-Arbeiterinnen. Zool. Jahrb., Abth. f. Syst., ix, 1895, pp.
1-26, pi. i, ii.
BOULENGER, C. '08 On the hermaphroditism of the amphipod Orchestia des-
hayesii Audouin. Proc. Zool. Soc. London, 1908, pp. 42-47.
BOVERI, TH. '02 Ueber mehrpolige Mitosen als Mittel zur Analysedes Zellkerns.
Verb. Phys. Med. Gesell. zu Wiirzburg, n. F., xxxv, 1902, pp. 67-
90.
BRUES, C. T. '03 A contribution to our knowledge of the Stylopidae. Zool.
Jahrb., Abth. f. Ont., xviii, 2, 1903, pp. 241-270, pis. 22, 23.
'05 Notes on the life history of the Stylopidae. Biol. Bull., viii,
no. 5, 1905, pp. 290-295, 2 figs.
BRUNELLI, G. "05 Sulla distruzione degli oociti nelle regine dei Termitidi in-
fette da Protozoi. Rend. Accad. Lincei (5), xiv, sem. 2, pp. 718-
721,1 fig.
CRAMER, P. J. S. '07 Kritische Uebersicht der bekannten Falle von Knos-
penvariation. Naturk. Verh. Holland. Maatschap. Wetens. (3),
vi , 3> 1 97> PP- I -474-
CUNNINGHAM, J. T. '08 The heredity of secondary sexual characters in re-
lation to hormones, a theory of the heredity of somatogenic char-
acters. Arch. f. Entw. Mech., xxvi, 1908, pp. 372-428.
DODD, F. P. '06 Notes upon some remarkable parasitic insects from North
Queensland. With an appendix containing descriptions of new
species by Col. Chas. F. Bingham and Dr. Benno Wandollek.
Trans. Ent. Soc. London, 1906, pp. 119-132, 2 figs.
432 William Morton Wheeler
DONHOF '60. Ein Bienenzwitter. Bienenzeitung, 1860, p. 174.
DRIESCH, H. '07 Die Entwickelungsphysiologie, 1905-1908. Ergeb. Anat. u.
Entwickelungsgeschichte, xvii, 1907, p. 157.
EMERY, C. '90 Studii sulle Formiche della Fauna Neotropica. Bull Soc. Ent.
Ital., ann. xxii, 1890, pp. 38-40, pis. v-ix.
'96 Le polymorphisme des fourmis et la castration alimentaire.
Compt. Rend, -f Congr. Intern. Zool. Leyde(Sept., 1895), 1896, pp.
395-47-
'94 Zur Kenntniss des Polymorphismus der Ameisen, Zool. Jahrb.,
Suppl. vii, 1904, pp. 587-610, 9 figs.
ENTEMAN, WILHELMINE M. '99 The unpaired ectodermal structures, of the
Antennata. Zool. Bull., vol ii, no. 6, 1899, pp. 275-282, 6 figs.
'04 Coloration in Polistes. Publ. Carnegie Inst. Wash., no. 19, Nov.,
1904, pp. 1-88, 6 pis.
FAXON, WALTER. '85 Revision of the Astacidae. Mem. Mus. Comp. Zool. Har-
vard Univ., x, 4, 1885.
FEWKES, W. '88 On the development of the calcareous plates of Amphiura
Bull. Mus. Comp. Zool. Harvard Univ. xiii, no. 4, 1888, p. 107,
Nature, no. 941, vol. 37, igth Jan., 1888, p. 274.
FITCH, A. '59 Emasculating Bot-fly (Cuterebra emasculator). 3rd Rep. on
the noxious, beneficial and other insects of the state of N. Y.,
1859, PP- 160-167.
('KKNZEL, J. '91 -Leidyonella cordubensis nov. gen., nov. spec. Eine neue Tri-
chonymphide. Arch. f. mikr. Anat., bd. xxxviii, 1891, pp. 301-316,
4 figs.
GIARD, ALFRED. '69 Sur I'hermaphroditisme de Melandryum album infeste
par Ustilago antherarum. Bull. Soc. Bot. France, xvi, 1869, Compt.
Rend, des Seances, 3, p. 213.
. '86 De 1'influence de certains parasites Rhizocephales sur les caracteres
sexuels exterieurs de leur holes. C. R. Acad. Sci. ciii, 1886,
p. 84.
"&"]a La castration parasitaire et son influence sur les caracteres exte-
rieurs du sexe male chez les crustaces decapodes. Bull. Sc. xviii,
pp. 1-28, 1887; Ann. Mag. Nat. Hist.,5 ser.,xix, 1887, p. 325;Natur-
wiss. Rundschau, ii, 1887.
Effects of Castration in Insects 433
'876 Sur la castration parasitaire chez Eupagurus Bernhardus L.
et chez Gebia stellata. C. R. Acad. Sci. civ, 1887, p. 1113.
'8/f Sur les parasites Bopyriens et la castration parasitaire. C. R. Soc.
Biolog., 5, ser. iv., 1887, p. 371.
'87^ Contributions a 1'etude des Bopyriens. Trav. du Lab. de Wime-
reux, .,.1887, p. 181.
'88a Sur la castration parasitaire des Eukyphotes des genres Palaemon
et Hippolyte. C. R. Acad. Sci. cvi, 1888, p. 502-505.
'SSk La castration parasitaire. Nouvelles recherches. Bull. Sci.
France et Belg., 3 ser., i, 1888, p. 12-45.
'88t Castration parasitaire probable chez Pterotrachea. Bull. Sci. xix,
1888, p. 309.
'88(/ La castration parasitaire de Lychnis dioicaL. par I'Ustilagoanthe-
rarum Fres. C. R. Acad. Sci. xvii, 1888, p. 757.
'Spa Note sur la castration parasitaire de Melandryum vespertinum
(Lychnis dioica). Bull. Sci. xx, 1889, P- I 5-
'896 Sur une galle produite chez le Typhlocyba rosae L. par une larve
d'Hymenoptere. C. R. Acad. Sci'. cix, 1889, p. 79.
'891: Sur la castration parasitaire de 1'Hypericum perforatum L. par Ceci-
domyia hyperici Bremi et par PErysiphe Martii Lev. C. R. Acad.
Sci. cix, 1889, p. 324.
'89^ Sur la castration parasitaire de Typhlocyba par une larve
d'Hymenoptere (Aphelopus melaleucus Dalm.) et par une larve de
Diptere(Atelenevra spuriaMeig.). C. R. Acad. Sci. cix., 1889^.708.
'89* Sur la transformation de Pulicaria dysenterica en une plante
dioique. Bull. Sci. xx, 1889, pp. 53-75, I pi.
'940 Evolution des etres organises. Sur certain cas de dedoublement
des courbes de Galton dus au parasidsme et sur le dimorphisrhe
d'origine parasitaire. C. R. Acad. Sci. cxviii, 1894, pp. 870-873.
'946 Convergence et pcecilogonie chez les insects. Ann. Soc. Ent.
France, 1894, pp. 128-135, Translated by Osborn in "Psyche,"
vii, pp. 171-175.
'02 Sur le passage de I'hermaphroditisme a la separation des sexes par
castration parasitaire unilaterale. C. R. Acad. Sc. cxxxiv, 1902,
p. 146.
434 William 'Morton Wheeler
GRASSI, B. '85 'Intorno ad alcuni protozoi parassiti delle Termiti. Nota
letta 15 Gen. 1885. Atti. Acad. Gioenia Sci. Nat. Catania (3), xviii,
pp. 235-240, 7 figs.
'93 Costituzione e sviluppo delle Societa del Termitidi. Accad. Gioenia
Sc. Nat. Catania, vi, vii, 1893, pp. 1-50, pis. 1-5.
HALLEZ, P. '86 Sur la loi de 1'orientation de 1'embryon chez les insectes.
Compt. Rend. Acad. Sci. Paris, ciii, 1886.
HEGNER, R. W. '08 Effects of removing the germ-cell determinants from the
eggs of some Chrysomelid beetles. Biol. Bull, xvi, no. I, 1908,
pp. 19-26, 4 figs.
HERBST, C. '01 Formative Reize in der tiertschen Ontogenese. Leipzig;, Arthur
' Georgi, 1901, -125 pp.
HERRICK, C. L. '83 Heterogenetic Development in Diaptomus. Amer. Natur-
xvii, 1883, pp. 381-389, 499-55> 794-795-
HORNE, CHARLES. '72 Notes on the hymenopterous insects from the north-
west provinces of India. Trans. Zool. Soc., London. 1872, pp-
161-196 pi. 19-22.
HOWARD, L. O. '68. A suggestion regarding development retarded by
parasitism. Canad. Entom. xl, no. I, 1908, pp. 34-35.
HUBBARD, H. G. '92 The life history of Xenos. Canad. Ent. xxiv,
257-261.
JANET, CHARLES. '03 Observations sur les guepes. Paris, C. Naud, Editeur.
: 93> 8 5 PP- 3 % s -
KELLOGG, VERNON L. '04 Influence of the primary reproductive organs on
the secondary sexual characters. Jour. Exper. Zool. i, no. 4, pp.
601-605.
KENT, W. S. '85. Notes on the infusorial parasites of the Tasmanian white
ant. Papers and Proc. Roy. Soc. Tasmania for 1884, pp. 270-273.
(Reprinted Ann. Mag. Nat. Hist. Ser. (5), c, xv, no. 90, June,
1885, pp. 450-453.)
KUNCKEL D'HERCULAIS, J. '94 Les dipteres parasites des acridiens: les mu-
scides vivipafes a larves sarcophages. Aptenie et castration parasi-
taire. C. R. Acad. Sci. cxviii, 1894, pp. 1106-1108. (Transl. in Ann.
Mag. Nat Hist. (6) vol. xiv, pp. 74-76.
Effects of Castration in Insects 435
LA BAUME, W. 'to Ueber den Zusammenhang primarer u. sekundarer Ge-
schlechstmerkmale bei den Schmetterlingen und den iibrigenGlieder-
tieren. Biol. Centralbl. xxv, 1910, pp. 72-81.
LEIDY, J. '77 On Intestinal parasites ofTermes flavipes. Proc. Acad. Nat. Sci.
Phila., 1877, pp. 146-149.
'81 The parasites of the Termites. Journ.Acad. Nat. Sci. Phila. (2),
viii, 4, (1881) pp. 425-447.
VON LENHOSSEK, M. '03 Das Problem der geschlechtsbestimmenden Ursachen.
Jena, 1903.
LEUCKART, Rud. '87 Neue Beitrage zur Kenntniss des Baus und der Lebens-
geschichte der Nematoden. Abhandl. Math. Phys. Cl. d. Kgl. Sachs.
Gesell. Wiss. xiii, no. 8, 1887, pp. 567-704, Taf. i-iii.
MARCHAL, PAUL. '96 La reproduction et 1'evolution des guepes sociales.
Arch. Zool. Exper. et Gen. 3 Ser. t. iv, 1896, pp. i-ioo, 8 figs.
'97- La castration nutriciale chez les hymenopteres sociaux. C. R.
Soc. Biol. (10) iv. 4, pp. 556-557.
MEEHAN, THOS. 'oo Contributions to the life-history of plants. No. xiv, i-
Fungi as agents in cross-fertilization. Proc. Acad. Nat. Sci.,
Philadelphia, 1900, pp. 341, 342.
MEISENHEIMER, JOHANNES. '07 Extirpation und Transplantation der Ge-
schlechtsdriisen bei Schmetterlingen. Zool. Anzeig. xxxii, 1907,
P- 393-
'09 -Experimentelle Studien zur Soma- und GeschlechtsdifFerenzierung.
Erster Beitrag. Ueber den Zusammenhang primarer und sekun-
darer Geschlechtsmerkmale bei den Schmetterlingen u. den iibrigen
Gliedertieren. Gustav Fischer, Jena, 1909. 149 pp., 2 pis., 55
text-figs.
MOLLIARD, M. 'pi Fleurs doubles et parasitisme. C. R. Acad. Sci., cxxxiii,
1901, pp. 548-551.
MORGAN, T. H. '92 Spiral modification of metamerism. Journ. Morph. vii,
2, 1892, pp. 245-251, 3 figs.
'05 An alternative interpretation of the origin of gynandromorphous
Insects. Sci. xxi, 538, 1905, pp. 631-634, I fig.
'09 A biological and cytological study of sex determination in Phyl-
loxerans and Aphids. Journ. Exper. Zool. vii, 2, 1909, pp. 239-352,
i pi., 23 text figs.
THE JOURNAL OF EXPERIMENTAL ZOOLOGY, VOL. 8, NO. 4.
436 William Morton Wheeler
MRAZEK, AL. '08 Myrmekologicke Poznamky. III. Acta Soc. Ent. Bohemiae,
v, 4, 1908, pp. 139-146, 4 figs-
NEWPORT. '48 -The natural history, anatomy and development of Meloe.
Second Memoir. Trans. Linn. Soc. London, xx, 2, 1848, p. 335.
OSBORN, HERBERT. '96 Insects Affecting Domestic Animals. Bull. no. 5, n.
ser. U. S. Dept. Agri. Eiv. Entom. Washington 1896, pp. 302, 170 figs.
OUDEMANS, J. Th. '99 Falter aus castrirten Raupen, wie sie aussehen und wie
sie sich benehmen. Zool. Jahrb., Abth. f. Syst. xii, pp. 71-88,
pis. iii-v.
PANTEL, J. '09 Recherches sur les Dipteres a Larves Entomobies, I Garacteres
parasitiques aux points de vue biologique, ethologique et histolo-
gique. La Cellule, xxvi, no. I, 1909, pp. 27-216, 5 pis.
PEREZ,, J. '86 Des effects du parasitisme des Stylops sur les apiaires du genre
Andrena. Soc. Linn. Bordeaux, xii, 1886, pp. 21-54, 2 pis.
PERKINS, R. C, L. '92 Stylopizecl bees. Ent. Month. Mag. (2) iii, 1892, pp. 1-4.
PIERCE, W. DWIGHT. '04 Some hypermetamorphic beetles and their hymenop-
terous hosts. Univ. Studies, Univ. Nebraska, iv, 2, Apr. 1904, pp.
1-38, 2 pis.
'08. A preliminary review of the classification of the order
Strepsiptera. Proc. Ent. Soc. Wash, ix, 1-4, 1908, pp. 75-85.
'09. A monographic revision of the twisted-winged insects compris-
ing the order Strepsiptera Kirby. Smiths. Inst. U. S. Nat. Mus.
Bull. 66, 1909, pp. xii, 232, 15 pis.
PORTER, JAMES F. '97 Trichonympha, and other parasites of Termes flavipes.
Bull. Mus. Comp. Zool. xxxi, no. 3, pp. 47-68, pis. 1-5.
POTTS, F. A. '06 The modification of the sexual characters of the hermit
crab caused by the parasite Peltogaster (Castration parasitaire of
Giard). Quart. Journ. Micr. Sci. S., N. L., 1906, pp. 599-621, pis.
33-34-
'09 Some phenomena associated with parasitism. Parasitology, vol li,
nos. I and 2, May 1909, pp. 42-56, 3 figs.
REAUMUR, M. de'4g Memoires pour servir a 1'Histoire des Insects. Tome vi,
- 1742, pp. 22, "23, pi. iv, figs. 10-12. [Sphaerularia bombi].
REGEN, J.'og Kastration und ihre Folgeerscheinungen bei Gryllus campestris L.
C? I- Mittheilung. Zool. Anzeig. xxxiv, no. 15, 29 Juni. 1909, pp.
4-77, 478.
Effects of Castration in Insects 437
'10 Kastration und ihre Folgeerscheinungen bei Gryllus campestris
L. II. Mittheilung. Zool. Anzeig xxxv, nos. 14-15, 15 Feb.
1910, pp. 427-432.
REUTER, ENZIO '07 Ueber die Eibildung bei der Milbe Pediculopsis graminum
(E.Reut). Festschr. f. Palmen, no. 7, Helsingfors 1907, pp. 1-39,
i text fig.
RILEY AND HOWARD '89 On the emasculating bot-fly. Insect Life. Vol. i.
p. 214-216, i fig.
SAUNDERS, EDWARD '82 Synopsis of British Hymenoptera. Pt. i. Trans. Ent. Soc.
London, 1882.
SCHMIEDEKNECHT, 83 Apidse EuropaEs, fasc. 6, 1883.
SILVESTRI, FILIPPO '03 Contribuzioni alia conoscenza dei Termitofili dell'
America Meridionaie. Redia, 1903, pp. 1-234, pi. i-vi.
'06 Contribuzioni alia conoscenza biologica degli Imenotter'
Parassiti. I. Biologia del Litomastix truncatellus (Dalm). Ann. R.
Scuol. Sup. d'Agri. di Portici vi, 1906, pp. 3-515 5 pis, 13 figs.
SMITH, GEOFFREY '050 High and low dimorphism, with an account of cer"
tain Tanaidae of the Bay of Naples. Mitth. a. d. Zool. Station
zu Neapel. xvii, 1905, pp. 312-340, pis. 20, 21, 13 figs.
'05^ -Note on a Gregarine which may cause the parasitic castra-
tion of its host. Mitth. Zool. Station zu Neapel. xvii, 1905, pp.
406-410.
'06 Rhizocephala. Fauna u. Flora des Golfes von Neapel.
29. Monographie. Berlin 1906.
'09 "Crustacea, "in Cambridge Natural History, vol. iv, 1939-
chapters iv and v.
THEOBALD, F. V. '92 Stylopized bees. Entom. Month. Mag. (2) vol.iii, 1892,
pp. 40-42.
THOMSON M. T. '03 Metamorphoses of the hermit crab. Proc. Bost. Soc.
Nat. Hist., xxxi, 1903, pp. 147-209, 7. pis.
VEUILLEMIN, P. '05 La Castration femelle et 1'androgenie parasitaires du
Lonicera Periclymenum. Bull. Mem. Seances Soc. Sci. Nancy,
1905, p. 19, 2 pis.
WHEELER, W. M. '89 The embryology of Blatta germanica and Doryphora
decemlineata. Journ. Morph. iii, no. 2, 1889, pp. 291-386, pis. xv-
xxi, 16 figs.
438 William Morton Wheeler
'93 A contribution to insect embryology. Journ. Morph. viii, no. I,
1893, pp. 1-160, pis. i-vi, 7 figs.
96 The sexual phases of Myzostoma. Mitth.,2ool. Stat. Neapel.
xii, 2 Heft. 1896, pp. 227- 302, pis. 10-12.
01 The parasitic origin of macroergates among ants. Amer.
Natur. xxv, 1901, pp. 877-886, I fig.
07 The polymorphism of ants, with an account of some singular
abnormalities due to parasitism. Bull. Amer. Mus. Nat. Hist,
vol. xxiii, 1907, pp. 1-93, pi. i-vi.
.SSEN, O. '98 Ueber die Riesenbildung bei Ascaris-Eiern. Arch. Ent-
wick. Mech. 'vii, 1898, pp. 642-676, 2 pis. 9 figs.