The social organization and breeding habits of the cotton-protecting kelep of Guatemala


Material Information

The social organization and breeding habits of the cotton-protecting kelep of Guatemala
Series Title:
Technical series / United States. Dept. of Agriculture. Bureau of Entomology ;
Physical Description:
55 p. : 23 cm.
Cook, O. F ( Orator Fuller ), 1867-1949
Place of Publication:
Washington, D.C.
Publication Date:


Subjects / Keywords:
Kelep -- Guatemala   ( lcsh )
Social hierarchy in animals   ( lcsh )
Insect societies   ( lcsh )
Kelep -- Sexual behavior   ( lcsh )
bibliography   ( marcgt )
federal government publication   ( marcgt )
non-fiction   ( marcgt )


Includes bibliographical references and index.
Statement of Responsibility:
by O.F. Cook.

Record Information

Source Institution:
University of Florida
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
aleph - 029625140
oclc - 27956863
lccn - agr06000611
System ID:

Full Text



L. 0. HOWARD, Entomologist.





0. F. COOK,
Bionomist in (Charge of Inre.tigatiown in A( rieulfural Economy of
Tropicrld and Si hbtrop iore Plants.





* ii



Washington, D. C., April 2.?, 1905.
SIR: I have the honor to transmit herewith a paper on the social
organization and breeding habits of the cotton-protecting kelep of
Guateniala (Ectttomma tubelc'.alati, 01.), by Mr. 0. F. Cook, Biono-
miist in Charge of Investigations in Agricultural Economy of Trop-
ical and Subtropical Plants, Bureau of Plant Industry, and tempo-
rarily on duty in this Bureau for the purpose of continuing the
investigations of this particular insect, whose cotton-protecting habits
hlie was the first to describe. This report contains evidence to show
that the breeding habits of the insect in question, especially its
methods of founding newv colonies, are e.seiitially different from those
of typical ants (family Formicida,) and resemble in important par-
ticulars those of the domestic honeybee. The possession of this type
of social organization will, in Mr. Cook's judgment, greatly facilitate
the establishment of the kelep in the cotton fields of the South if the
insect should be able to withstand the change of climate and other
natural conditions. I recommend that this paper be published as
Technical Series No. 10. Another bulletin dealing more fully with
the details of the life history of the species is being prepared and
will be fully illustrated.
Respect fully,
Entomologist and Chief of Bnreae.
Secretary of Agrie 'dt re.
0 (3)

Digitized by the Internet Archive
in 2013



Introduction --------. --------------------------------------------------- 7
Swarming and other time specializations-------------...------------------- 8
Annual mating concourses of termites and true ants ---------------------- 8
Breeding habits of the kelep ----------------.----------------------------- 10
Size of the kelep colonies-----.-. ..----------------------------------- 11
Population of kelep nest the same throughout the season---_----------- 12
Only one type of worker ------------------------.------------------- 12
Kelep colonies not hostile --------------------....----------.. -------- 14
Males continuously present ------------------.---------------------- 16
Queens inactive---------------------------------------------------- 17
Migration of colonies ---.-------.------------------------------------- 20
Queens carried by workers ----.----------.--------------------------- 20
Division of a colony -------------------------------------------------- 21
Colonies found without queens --------------------------------------- 22
Replacing a queen --------------------------------------------- 22
Types of social organization among insects ------------------------------ 23
The termite society -------------------------------------------------- 24
Explanations of worker caste------------------------------------- 24
Highly specialized worker castes --------------------------------- 25
The ant society ---.-------------------------------------------------- 27
The bumblebee society- ------------------------------------------28
The honeybee society -----------------------..... ...------ 1 ...-- 28
The kelep society ------------------------------I.------------------- 28
The driver society ------------------------------------------------ 29
Ambiguous use of the term "swarm"------------------------------- 32
Determinate and indeterminate colonies ---------------------------------- 33
Complete socialization of the kelep --------------- .----------------------- 34
Doctor Ashmead's new classification of the Ponerida ---------------------- 38
Synopsis of subfamilies of Poner:d e --------------------------------- 38
Synopsis of tribes of Ponerinx ------------------------------------ 39
Synopsis of tribes of Pachycondylin ------------------- ---------- 39
Habits of Ponerids --------4---------------------------------------- 41
Temporary brood cells of the kelep..------------------------------------ 43
Weevil-stinging wasps ----------------------------------------------- 44
Use of fibers in construction.-------------- ---------------------------46
Stinging habits of the kelep ------------------------------------------- 47*
Harmlessness of the kelep to man -------- ------------------------ 48
Adaptability of kelep organization to agricultural purposes --------------- 49
Summary.-------------------------------------------...-------------..... 50




In preceding reports treating of the kelep as an enemy of the cotton
boll weevil the distinctness of its behavior from that of the true ants
has been noted. To avoid in some measure the miisapprehension likely
to be caused by calling it an ant it seemed desirable to introduce with
the insect its distinctive Indian name, kelcp. In the minds of the
natives of Guatemala, the kelep is not a kind of ant, but an inde-
pendent animal not to be associated with ants. The more we learn
about it the more this aboriginal opinion appears justified, not alone
because the kelep is a beneficial insect, but because it has a different
mode of existence and a different place in the economy of nature.
The popular classification of the social Hymenoptera recognizes
three types-the ants, the bees, and the wasps, the ants being distin-
guished from the others by the absence of wings. The kelep falls,
however, into none of these groups. To call it a wasp or a bee would
not misrepresent the practical facts more than to call it an ant. In
reality the kelep represents a fourth category of social Hymenoptera,
as distinct from the other three as they are from each other. Authori-
ties on the classification of the Hymenoptera have admitted a rather
close affinity between the wasps and the ants, but the kelep differs
from both of these groups and approaches the b)ees in important
respects, and especially in those which affect the question of its
domestication and utilization in agriculture.
It was naturally supposed at first that the kelep would have the
same habits as the true ants which have been associated with it as
members of the same family or stfbfamily, but the differences were
greatly underestimated. If the Hymenoptera were classified by a
taxonomic system consistent with that applied to the higher animals,
the kelep would need to be recognized as the type of a new and
distinct family. It is, moreover, the first member of its family of
which the habits have become known. Under such circumstances
it was quite impossible, obviously, to determine in advance whether its
habits and instincts would permit its colonization in the United
States and its use in agriculture.
The fundamental difference between the ants and the kelep, and
that in which the latter resembles the honey bee, lies in the methods
of swarming. Among the bees and the keleps swarming results

directly in the formation of new colonies, but the swarming of the
ants is a distinct biological phenomenon having for its object cross-
fertilization. The kelep is completely socialized, like the honeybee,
while the ant is not. The keleps and the honeybees live only in com-
munities, while the ants at one stage of their life history leave the nest
and meet the vicissitudes of independent existence as solitary indi-
viduals, like the nonsocial insects. The social organization of the
kelep represents a line of development distinct from that of the ants,
and shows a relationship with the parasitic and predaceous wasps
rather than with the true ants.


The swarming of the ants is one of the many interesting phenomena
which might be grolupI)e(d under such an expression as biological
synchronism. Species are organisms, or at least organizations, and
in some of them there is manifested a simultaneity or time cobrdina-
tion of thle numerous members corresponding to the orderly develop-
ment of the cells of which the body of the individual is built. A
flock of birds or a school of fish, with the individuals separated at
equal distances an(d executing all their movements in exact unison,
is a striking example of such synchronism, but other no less myste-
rious adljustments'are necessary to enable animals and plants to keep
so exactly the annual appointments by which the interbreeding of
the members of the sl)ecies is maintained. The climatic vicissi-
tudes of temperate regions make complete simultaneity difficult, and
have led us to ascril)e the annutial recurrence of events to the change
of seasons rather than to recondite internal causes. The thirteen-
year and seventeen-year trysts kept by the periodical cicadas over
wide regions show, however, that more than sumn totals of heat, cold,
and food are involved, evenll in temperate climates.
Under the equable conditions of tropical existence, where the
seasonal explanation entirely fails, there are biological events which
might seen to show that plants and animals not only have drill-
masters but time locks. Some of the Asiatic bamboos grow for
thirty years or more by vegetative increase alone without producing
flowers or fruit, and then all the members of the stock blossom, bear
seeds, and die together, without reference to the age, place, or con-
dition of the individual plants which may have been propagated from
cuttings and carried to remote parts of the earth.

Nests of West African "white ants," or termites, are crowded for
weeks with winged individuals, but not one is to be found outside
until some moist afternoon or evening, when the young sexual in-

sects emerge from all the nests of the species in the same hour. The
air is filled as by snowflakes or a plague of locusts. Lights attract
the flying insects, and are smothered under heaps of toasted termites.
The insectivorous birds and reptiles gorge themselves to repletion.
By the next morning the detached wings have been blown together
in windows, and nothing more is to be seen of that particular species
for another year except by digging into their nests or galleries.
Probably not one pair of termites out of many thousands survive
to become the parents of a new colony, but the purpose of the sacri-
fice is accomplished if these have secured the interbreeding necessary
to maintain the incredible fecundity by which the termite queen
furnishes the population of a community to be enumerated in millions.
The mothers of such colonies have been seen to lay from 40 to 60
eggs per minute.
The true ants belong, of course, to an entirely different order of
insects, and their social organization and swarming habits have
been attailled(l quite independently. Nevertheless some of them, and
especially the families best known to entomologists, have a domestic
economy and a morphological diversification of the members of the
colony surprisingly paralleled to that of the termites, including the
habit of annual mating concourses of sexual adults. There are in
many species, both of ants and of termites, not only the two normal
sexes and the sterile workers, but some of the last are further special-
ized in structure, instinct, and social duties as soldiers, foremen,
nurses, etc. A family of American ants (Cryptocuderid) also re-
sembles a genus of African termites in maintaining an extensive
and highly specialized system of fungus gardens.a Other ants have
domesticated plant lice, mealy bugs. and leaf hoppers for the sake
of the honey dewv secreted by these animals, which are herded,
a With these fungus-cultivating ants and termites, at least, it would seem
that a new colony can scarcely be founded by a pair of sexual termites or by a
single fecun(dated female ant unless they carry their domesticated fungus with
them. It is possible, however, tlihat in both cases the newly mated insects are
adopted and set up in housekeeping and farming by workers of their own
species, who bring "'spawnu" of the fungi from the older colony with which
they are in communication. This might the more readily happen because long
subterranean galleries are a prominent feature of the architecture of the
fungus-growing insects, both ants and termites.
The keleps, indeed, may be said to have taken a step toward the domestica-
tion of the cotton plant. They have at least adopted it, and show an instinc-
tive interest and attraction for it in preference to other plants. That this also
extends to a special animosity for the boll weevil as an enemy of the cotton is
not, l)erhaps, to be claimed, but the habit of living on the cotton plant has
resulted, no doubt, in giving the keleps a special familiarity with the boll
weevils and a special skill in capturing and stinging them.

28929-No. 10-05 M--2


pastured, and cleaned with as much care as modern dairy cattle. i
Some ants hhve also a highly developed slave-making instinct, and
undertake regular raids on the nests of other species to capture their
young and carry them away to be brought uip in servitude.
It would seem, therefore, that in the highest members of both
groups of social insects the chief purpose served by this simultaneous
emergence, or swarming," as it is commonly called, is interbreeding,
or cross-fertilization, rather than the founding of new colonies in
more distant localities, as hitherto supposed. This may explain why
the species of these insects have achieved nothing very remarkable in
the way of geographical distribution, in spite of the immense fecun-
dity of their females. A single individual ant or a pair of termites
might be able to establish a colony in a new locality, but the lack of .
opportunities for cross-fertilization might prevent the perpetuation
and further extension of the species, which could increase its range
only by gradual, continuous expansion.


With such instances in mind it becomes easier to appreciate the fact
that the breeding habits of the kelep differ notably from those of the
termites and termite-like ants, and approach those of the domestic
honeybee. Although the actual migration has not been observed,
there are strong indications that. instead of emitting annual broods of
sexual individuals and founding colonies by means of solitary fecun-
dated females, the kelep communities simply subdivide after the
fashion of the bees, or in a still more practical and business-like man-
ner. the problem of cross-fertilization having been solved in another
ailThe extreme develoeI)net of the p;istoral instinct is to be tounl In an ant
which takes care of the e,,ggs of its plant-lice cattle through the winter.
"It is not merely that the ants milk them, defend them from attack, some-
times protect them by earthen iniclosure,; from too great summer heat, but over
and above all this they collect the eggs in autumn, keep them through the
winter, and plant then out on their proper plant in the spring. Some of the
root aphides may always be found in ants' nests, but I was much puzzled
years ago by finding in ants' nests some black eggs which obviously were not
those of ants. Eventually 1 ascertained that they belonged to a species of
aphis, which lives on the leaves and leaf stalks of plants.
"These eggs are laid early in October on the food plant of the insect. They
aire of no direct use to the ants, they are not left where they are laid, ex- :
posed to the sevecrity of the weather and to innumerable dangers, but are brought
into their nests by the ants and tended by them with the utmost carL through the
long winter months until the following March, when the young ones are brought
out and again placed on the young shoots of the daisy. This seems to be a most
remarkable case of prudence. Our ants may not, perhaps, lay up food for the J
winter, but they do more, for they keep during six months the eggs which will
enable them to procure food during the following summer, a case of prudence :1
unexampled in the animal kingdom."-Avebury, 1905, The Open Court, 19: 190. A


way.a The data which seem to support this conclusion are briefly
,-ummlarized in the following paragraphs.


The colonies, while very variable in size, range between 100 and
400 workers, and thus are of the same order of magnitude, instead of
forming a series from solitary queens to communities of thousands
or millions, as among the true ants and termites. Out of nearly 150
nests which have been explored only 3 or 4 had less than 100 workers.
This fact alone was sufficient to place the kelep in strong contrast
with the leaf-cutting ants, which are such conspicuous members of
the insect fauna of Central America. The colonies of the leaf cutters
contain enormous numbers of individuals, hundreds of thousands, or
millions, perhaps; and yet the young queens begin alone. The small
burrows of several such were found while kelep nests were being
opened, the annual mating flight of the leaf cutters having taken
place only a short time before.b
Most of the colonies brought from Guatemala in the first im-
portation were, as it now appears, mere fragments of normal com-
mlunities, containing from 20 to 5)0 workers. The jars obtainable in
Guatemala for use as cages were very small, and it was feared that
overcrowding would be detrimental. It was supposed, too, on the
analogy of the ants, that the colonies would readily replenish their
numbers if conditions should prove favorable in Texas.
In the Texas experiments, however, it became apparent at once
that in courage and general activity the behavior of the large colonies
was very differentt from that of the small ones, a fact which the char-
acter of the social organization permits us to appreciate more fully
than before. W1hlen permitted to do so the larger colonies generally
moved with promptness out of the cages and established themselves
a After the above was written Mr. McLachlan reported from Victoria, Tex.,
that there had been an increase in the number of colonies in two kelep settle-
ments in the cotton field near that place. One wire-netting cage which had
been supplied with four imported colonies was found to contain six colonies;
another had seven colonies, though only five had been left in it. Before leaving
Texas in October, I had noticed nkimerous young keleps in these cages, but had
observed no addition to the original number of colonies.
b It seems to be true, as the Indians say, that the leaf cutters are unable to
establish themselves in territory occupied( by the keleps. Presumably the keleps
kill the leaf-cutter workers as soon as they come out of the ground, and the
queen, being thus unable to raise a family to forage for her, soon starves to
(death. If there is a nest of leaf cutters near enougli to a cotton field to make
a raid upon it, the Indians protect it by a fence of tough leaves of the plant
called niosh (Calathwea). The same leaves are used by the black Caribs" of
Livingstou for lninig the waterproof baskets which they weave from the climb-
ing spiny palm (Desmoncus).


in new burrows of their own digging. Small colonies often refused
to leave the cages. Large colonies took boll weevils as soon as offered,
stung them, and carried them back into the nest to feed the young,
while some of the less populous communities showed none of the
hunting instinct and would tolerate the boll weevils for hours or
even for days.
It has been observed, too, by many students of bees and other social
insects, that colonies too much reduced in numbers may become
listless and discouraged and fail to manifest their normal activities
and instincts. To this rule the kelep is no exception; nevertheless,
even in very small numbers or as solitary individuals they do not lose
entirely their rational demeanor. This self-possession is probably a
consequence of tihe habit of the kcleps to spend considerable periods
outside their nests patrolling the cotton plants, or standing motion-
less, waiting for boll weevils or other insect prey. The tempera-
mental contrast with the honeybee in this respect is very striking.
The bee is above all. and even to a greater extent than the ant, a creature of
the crowd. She canll live onl3 in the 1mid(st of ai multitud(le * Isolate
her, and, however abulndant the food or favorable the temperature, she will
expire in a few days, not of hunger or cold. bitt of loneliness. From the crowd,
from the city, she derives an invisible aliment that is as necessary to her as

The numbers of the insects and young and other conditions inside
the nests of the keleps in Guatemala have been found to be the same
after an interval of over six months. The first exploration was made
ait the end of the dry season, in April, May, and June; tihe second at
the end of the rainy season, in November and December. Numerous
captive colonies also have been under continuous observation through-
out the same period. No indication has been detected of any seasonal
difference of habits, nor is it necessary to suppose that anything dif-
ferent takes l)lace in order to explain the domestic economy and
breeding habits of the species.'

The workers are all of the same form and of nearly the same size,
with no indications of the existence of a first brood of very small
individuals. In some colonies the workers average appreciably
larger than in others, but there is no such diversity as among the true
a Maeterlinck, M., 1901, The Life of the Bee, 30.
bWhile this report has been awaiting publication the period of observation
in Guatemala has been extended through the remainder of the winter and spring
months. The maximum of breeding activity appears to fall in the dry season,
at the end of the cotton-growing period, in March and April. Nests excavated
by Mr. G. P. Goll contained in some instances over twice as many cocoons as
adult insects. In other seasons this proportion is usually reversed.


ants. If colonies were established by solitary fecundated females, as
among the true ants, some of the nests would have shown examples of
the undersized workers, of which the first brood raised by the queen
ant is regularly composed.
It is now a well-established fact that every ant colony is founded by a single
fertilized female, or queen. The insect loses her wings and buries herself in a
small cavity in the soil or wood that is to formnt the future nest. After entering
the cavity she usually closes the opening so that she is completely shut off from
the outside world. She deposits, at the expiration of a certain time, a niumlber of
eggs, and when these hatch as larvae she does not go abroad in quest of food,
but feeds her offspring with substances regurgitated from her own body. These
substances are ultimately derived from the fat body, a store of nutrient aceu-
inulated during her life in the maternal nest, which she forsook to take the
nuptial flight. Of course, the insect must derive her own nourilshinuent from, the
same internal source, and as, in all ants, the development of the young extends
over a considerable period of time, it follows that the larv:m are of necessity
poorly fed. and after pupation hatched as dwarf workers (niicroergates).
The nunber, too, of these diminutive creatures is limited, so that the whole col-
ony in this incipient stage is a family consisting only of the huge mother and a
few dwarf offspring.a
Workers much smaller than any found in nature were raised in
some of the captive colonies of keleps, doubtless as a result of un-
favorable conditions or lack of the normal amount of animal food.
Nests with diminutive workers would not furnish proof positive,
therefore, that the kelep, any more than the honeybee, ever founds col-
onies by means of isolated queens. Disease, parasites, or starvation
might be expected to bring about in nature, as in captivity, a condi-
tion which gives the kelep colonrly a superficial resemblance to a
recently established community of true ants.
The finding of a colony of a few small workers of Odontomachius
clarus has been reported by Professor Wheeler as evidence that the
Poneridat agree with other ants in the method of founding colonies.
Other authorities on the classification of Hymenoptera recognize
Odontomachis as constituting a family distinct from the Poneride
by structural characters. Nevertheless, the similarity of habits be-
tween Odontomachus and Paclycondyla is so great that a difference
in social organization seems very improbable. The fact that the
Odontomacrhus nest found by Professor Wheeler was in a cavity in a
stone may be the explanation of its unprosperous condition
a Wheeler, W. M., 1902. Science, n. s., vol. 15, p. 768.
b Wheeler, W. M., Science, n. s., vol. 15, p. 7<69: In a former paper I main-
tained that the Ponerina' perhaps constitute an exception to the genera! method
of establishing colonies, but I have recently found in a small cavity in a stone a
fertile dealated queen of Odonomniucih u. clarus surrounded by five diminutive
workers. While it is certainly remarkable that one does not find similar
incipient colonies of other Ponerinve, this observation makes it prol)able nevpr-
theless that the ants of this family agree with the Componotinae, Myrmicinme,
and Dolichoderinvt in their methods of founding colonies."


There is supposed to exist among the worker bees a division of
labor, but nothing of the kind has been observed among the keleps,
except that young individuals, which are easily distinguishable by
their lighter color, remain in the nest for many days and perform
nurse duty, while the foraging devolves upon the older and darker-
colored insects.a
The demoralization of some of our imported colonies may have
been increased by the lack of a normal succession of young keleps to
serve as nurses. The predaceous instincts of the older workers may
incline them not only to neglect the larvae, but to yield more readily
to the cannibalistic tendency which some of the colonies have mani-

Kelep nests are frequently placed only a few inches apart, the
workers of the differentt colonies not being actively hostile. Members
of two colonies will forage on the same cotton plant or tree trunk
with no signs of animosity. Stranger ants introduced into captive
colonies for observation have not been attacked. They usually receive
little attention; if they enter the burrow they are likely to be brought
out and carried to the boundary of the inclosure, but are released
without injury. In natutire such stragglers, if any, would merely be
escorted to the border, as it were.
Under thle social ecoioniy of the true ants the species consists of
fewer and more scattered colonies of larger size. The workers from
different nests often have as nmiuch animosity for each other as for
members of distinct species. This hostility serves a practical pur-
pose, the close proximity of nests being, among the ants, a distinct
disadvantage. It is only in large and prosperous colonies that nuimer-
ouis sexual adults can be brought to maturity. Too many colonies
close together would mean a general scarcity of food and would keep
all the communities poor and unproductive.
The power of ants to distinguish at once between members of their
own and of other colonies has long been recognized as one of the most
remarkable refinements of instinct, and lhas been the subject of exten-
sive study and experiment. A recent and extremely careful investi-
gation has been made by Miss Adele M. Fielde, who finds that the
a Some species of ants and of termites have a special caste of very small,
slender workers which do not leave the nest, but are devoted to nurse dury.
These might be looked upon as representing a social specialization by which the
microergates are not confined to the first brood. On the other hand the very
great specialization of the so-called soldier castes of the termites would seem
to indicate that they are the oldest representatives of the worker series, and this
view seems to be supported by the fact that the "workers" of Calotermes,
which have the simplest social organization, are more like the soldiers of other
genera than they are like the workers.


olfactory sense, through which the instinct of animosity to strangers
is aroused, is so acute that ants will attack even their older sisters. if
they have been reared without contact with others of like age.a
It is much better policy, from the standpoint of the social economy
of the ants, for a given area to be occupied by one large colony than
by many small ones, and with many of the ants this instinct of hos-
tility is supplemented by that of retaining a part of the young females
in the nest. This enables the community to be expanded to the
greatest' extent possible and avoids the waste of unnecessary competi-
tion and conflict.
The more amiable dispos-ition of tle keleps has enabled them to
reach different and somewhat more democratic solutions of their
social problems. They take better cnre of their young females, and
do not find it necessary to make war upon their neighbors of the same
Like other carnivorous animals, the-kelep- become cannibals when
driven by hunger, but in normal conditions of Indian agriculture an
abundance of keleps in a locality would tend to increase the area
planted to cotton, so that their existence would not be subject to the
usual laws of competition in the struggle for existence.
Barring accidents which may compel a change of ivid(ence, an ant
colony remains indefinitely in the same place, but the kelep organi-
zation is decidedly imiore mobile. Colonies can change their locations
readily and may do so with p)romptne.s when a more desiralde spot
has been found. One of the ends gained by the kelep in moving
would be to leave behind parasites, with which their nest. sonietinies
become infested.b
It is a familiar fact in the study of plants that some grow in
clusters and others are solitary. If a -eed produces only a single
a Fielde. A. M., Biological Bulletin. 7: 227. In s4oie of tihe ants studied by
Miss Fielde the aversion to strangers is said to extend even to a maiinifest
preference" of the females for males fromin the same colony. It should not be
taken for granted, however, that the laying of pggs by queens captured before
the mating concourse proves a previous m.ating inside the nest, nor that all the
males which may be found in a nest origiJnated( in it.
SThe habit of frequent mnovinig might also explain the apparent absence
among the keleps of the guests or messmates which have attached themselves
to many of the true ants. In Guatemniala the kelep nests usually harlior a snail
or two, a small diplopod of the order Merocheta, a small creamny-white thysa-
nuran, and a worm which infests the bone yard or collection of dismembered
skeletons of their prey, which are stored in a special underground chamber.
A few specinimens of a hymenol)terous parasite identified by Doctor Aslimead
as Isoniaralia coronalUa Westwood have emerged from some of the captive colo-
nies during the journey to the United States, and even, in two instances, after
they had arrived fn Washington.
The worst enemy of the keleps, however, is a mite. which has become ex-
tremely numerous in some of the cages, especially those inhabited by small and
demoralized colonies.


stalk the plants are solitary; if there are numerous buds and branches
from roots or underground stems a cluster is formed and the plant
is said to have a cespitose habit, depending upon the different method
of reproduction. Similarly with the ants and the keleps the former 4
have their colonies remote and hostile, the latter adjacent and
friendly, or at least tolerant of each other. And though these dif-
ferences might at first seem slight and insignificant, they result, in
reality, from a fundamentally distinct system of social organization.
It is not only a regular and normal condition for ant colonies to
be solitary and remote from each other, but there is a provision of
mutual hostility which prevents the establishment of closely adja-
cent communities. The kele)ps, on the contrary, are social not only
to the extent of forming colonies like the ants, but they are social to
the second degree, as it were, in that the breeding habits of the in-
sects provide for the normal existence of closely adjacent communi-
ties. Solitary kelep colonies would be as abnormal as closely adja-
cent ant colonies and as clearly at a disadvantage. Adjacent ant
colonies might suffer for food, while a solitary kelep colony might
meet extinction through inbreeding if the visits of males from other
nests were excluded.

Winged males were found in the nests in about the same numbers
in November as in June. They appear to be regular inhabitants and
are received by all colonies with appl)l)arent indifference. They are
agile, active insects and( could readily pass from one colony to an-
other, thus providing cross-fertilization without annual swarming.
Males are not present inii all the nests, and the number varied from
1 to 7 or 8, a diversity itself strongly indicative of the absence of the
annual emergence, even of the males, at any stated time.a The num-
ber of males had also no apparent relation to the numl)er of females.
When the nests were opened in Guatemala thle males often
attempted to escape by running rapidly away, luit, did not take to
wving, although Mr. ILewton informs me that some of those brought
to Texas showed their ability to fly. They are, however, so active and
fleet of foot that they could pass readily from one colony to another
of the closely adjacent nests even without flying. Wingless males are
known in sonime of the species of Ponera, and also in certain parasitic
a In Guatemala in the spring the number of males raised by some of the col-
onies was found to be much greater. Mr. Goll counted about 40 in one nest.
Messrs. Kinsler and McLachlan saw numerous males at large in thle cotton fields
and observed, further, that they are often caught by the workers and carried
down into the nests.



The kelep queens, even when young, are distinctly less active than
the workers. Isolated queens have shown no ability or inclination
to excavate nests and very little interest in eggs or larva? which have
been intrusted to them.
It does not appear that the keleps have the art of regurgitating
food for their larva, or for each other, but they have, instead, the
curious habit of opening their mandibles wide and lapping up drops
of nectar, moistened sugar, or honey on their mouth parts. The
liquid is thus carried into the nest and dispensed(l to the other members
of the community, old and young. The queen is regularly fed in this
way, though in a few instances the queens of captive colonies came
to the surface to eat sugar with the workers.
In the true ants the young queen bites off her wings,.a excavates her
own burrow, and cares for her first brood of eggs after having
laid them, and appears not to be deficient in intelligence and activity
in comparison with the workers. The demeanor of the kelep queens
inside the nests differed notably from those of a species of Form ica
kept in the laboratory at Victoria, Tex., for purposes of comparison.
The kelep queen is the last to notice any (ditlurb)ance of conditions,
such as the admission of light, but the Formnica queens were even
more nervous and irritable than their workers and were the first to
run for shelter. The same was true of them when their nest was
1)eing dug out.
The only instances in which the kelep queens have shown any note-
worthy activity have occurred when they were condemned to solitary
confinement. As though realizing the uselessness of such an existence
these queens often make reckless attempts at escape and run rapidly
away, in complete contrast to their usual quiett and sluggish demeanor
when associated with workers and egcs or young. In nature, or rather
in the Guatemalan cotton fields, this activity on the part of a queen
which had for any reason lost her family might be of advantage in
enabling her to find a home with some other colony. Strange queens
seem always to be welcomed when introd(luced into a new community
with or without a queen of its own.b The removal of the queen, on
the other hand, does not seem to have any effect upon the actions of
the workers as long as there are eggs and larva- to- care for, but
workers alone become utterly listless and stand idly about with their
a Thiis was observed in Texas by Mr. Frederick L.. Lpwton in a young queen
of Crcmastogaster lwviumscila Mayr.
b In oue colony studied by Mrs. Cook two strange queens, while not actually
bitten, were dragged about in an unfriendly manner and did not long- survive.

28929-No. 10-4.035 M- 3

heads and antenna? in the air, as though waiting for something to
There is no indication that under normal conditions the kelep
queen ever leaves the nest voluntarily for mating purposes. There
seems to be no record of a queen of this family being captured
outside the nest. Mating inside the nest has been reported in another
member of the family Poneridae." Normally wingless females are
also known in some of the other genera, which proves that with them,
at least, the power of flight has been lost by the females as well as
by the workers. In some species of ants and of Ponerid.e there is
no very sharp line between the workers and the queen, all of the
intervening stages being occasionally found. Kelep queens and
workers are, however, quite distinct, so far as observed. The queens
are two or three times as large as the workers and are of a darker and
more reddish color. The workers occasionally lay eggs, but they are
of a distinctly smaller size than those of the queen. They remain
white instead of turning a deep gray or blackish, like those of the
queen, and are apparently recognized as worthless, for they are at
once fed to the larvae, a fact discovered by Mr. F. L. Lewton.b
When there are ino larva' the keleps seem n ot to know what to do
with these diminutive eggs, but continue to carry them around in
their inan(libles. In one small colony I have seen three worker
keleps with eggs at the same time. which would seem to indicate
either that they are laid in considerable numbers or that they are
carried about for considerable periods.
Thle wings of the female being useless, it might be expected that
they would b)e bitten off as promptly as in tile true ants, or even
sooner, but this is not thle case. It seems scarcely possible that the
queen could bite them off herself, even if so inclined, and of such an
instinct there is no indication. The wings seem to be worn for an
indefinite period. Queens with one or both wings frayed to half
their length or less are occasionally found. The wing stumps, also,
are of irregular sizes. It seems not unlikely that the workers gnaw
them away gradually as a part of the cleaning process.
In Pachycontdyla. as described by Professor Wheeler, and in Neo-
ponera,c which occurs in the vicinity of Victoria, Tex., the queens
a Wheeler, W. M.. 1901, Biological Bulletin, 2:49.
b The laying by workers of eggs of nearly normal size which also turn dark
and seem likely to develop normally was reported from Victoria, Tex., by Mr.
Argyle McLachlan. after the above was written.
c The species of Yeoponera, which commonly lives int the mesquite bushes
about Victoria, has been identified by Doctor Ashmead as N. villosa Fab. It
is not without interest with reference to the possibility of establishing the
kelep in Texas that the distribution of Neoponera rillosa extends from Texas
to Brazil.


and workers are very nearly of the same size, and, except for the
possession of wings by the former, look quite alike, while in other
genera, like Leptogcuy.s, the females, though wingless, are very
different from the workers. Such facts might have served as indi-
cations of considerable latitude of social specialization in the family
as a whole, as abundantly shown when the kelep is added to the series.
The Poneridae with wingless females seem to ally the group still
further with the Dorylidce and Mutillidae and increase the contrast
with the true ants, where the wings still remain as essential for the
females as for the males. It may be needless, perhaps, to add that
it is the loss of the wings by the workers and by the queens which
marks the course of recent evolution in the Hymenoptera as a whole,
as in all other iisectk. The most primitive insects were winged; all
the wingless insects are descendedd from winged ancestors.a The fact
that so small a proportion of the members of an ant colony have
wings, hnd these for so short a time, tends to make it appear that the
wings are a special provision, while in reality the specialization is
all in the direction of winglessness. The ants are in this respect
more primitive, since they have preserved a character which is being
lost in the Poneride and which has entirely disappeared in the
Dorylid (,T, Mutillidae, and Thynnidae.b
The winglessness of the queens has been attained, however, by an
evolutionary step distinct from that which resulted in wingless
workers, and has a very different significance from the standpoint of
the social organization of the species.
The winglessness of the workers may be looked upon as a part of
the general stunting and sterility of the individual for lack of ade-
quate food. In the early stages of the development of a new colony
it is undoubtedly an advantage not to have too many larvae, which
would increase the danger of starvation for the whole lot.- The loss
or disuse of wings by the sexual females of parasitic wasps, drivers,
and Poneridae, on the other hand, is a sequel of the abandonment by
a Cook, 0. F., 1902, The Earwig's Forceps and the Phylogeiiy of Insects; Proc.
Entom. Soc. Wash., 5: 84.
b Winglessness of one of the sexes also occurs in several anomalous ants which
live as parasites in the nests of other species and establish no independent colo-
nies of their own. In the genus Touiogtatliu.- the male is winged, but the
female is wingless and workerlike. Insects from the same nest have been found
not to pair. In the genera Antcrgates, Formnico.retmix, Syujmiyetnica, and Carvio-
conmdyla the conditions are reversed, the males being wingless and similar to the
workers, while the females have normal wings.
c The storing of honey by the bees had for its primary purposes, we may
believe, the avoidance of temporary striigency of food during the period of
brood rearing, not the laying up of supplies to enable the adult members of the
colony to pass the winter. The storing habit enables the bees to occupy tem-
perate regions, but tropical bees also accumulate supplies of honey.


that sex of the habit of flight, which means that the social organiza-
tion of the species is essentially different from that of the true ants,
in which both sexes emerge into the open air for a marriage flight.


In the transfer of a colony of keleps from one nest to another the
workers take the initiative. After an excavation has been made
eggs and larva' are carried over. First attention is given to the eggs,
the possession of which renders the new colony in a measure inde-
pendent of the presence of a queen. Even before the new burrow is
dug some of the workers load themselves with eggs and stand ready
to carry them in.
The ancient popular theory that the communities of the social
insects are organized on a monarchical basis finds little support in the
way of detailed facts. It has long been known that the nervous sys-
tem of worker bees and ants is more highly developed than that of
the sexual insects, and especially the parts which correspond to the
brain of higher animals. Even in the most highly organized society
of the termite.-, the workers are the intelligent and efficient part of the
community, the queii'en having no other function than the laying of
the eggs. from which the huge family is hatched. The termite
" mother," as the Africans more correctly call her, becomes enor-
moiisly disten(led and completely unable to move. She is kept with
her mate in a special chamber with very small entrances through
which only the workers can pass, and even these openings are
pronmpl)tly sealed up with earth when the colony is attacked. Among
the nomadic "driver ants" the mother of the colony does not lead
the procession and never goes abroad in daylight. The natives of
Liberia say, however, that in rare instances she is seen at night being
hurried along by her numerous family.
Among the honeybees, also, the workers, rather than the queen,
take the initiative in matters of the internal management of the
colony which lead up to swarming. Unlike the keleps, the bees do
not carry eggs or larvae with them, and are hence completely de-
pendent upon the presence of the queen to insure the perpetuation of
the new colony.

If the kelep queen does not follow at once to the new nest a worker
seizes her by the mandibles, raises her in the air, and carries her over
bodily. This has been observed repeatedly in connection with the
prompt transfers which many of the imported colonies made from
their cages into the ground. The queen submits to this treatment as
though it were a regular occurrence, and remains quiet and rigid
while being carried about. In one instance several workers also re-


mained behind, but were caught and carried by their sisters into the
new burrow. This simple expedient avoids many difficulties and
complications. The queen need not be possessed of any instinct of
leading or accompanying the swarm, as among the bees, and there is
no danger of her becoming lost as a result of her comparatively de-
ficient instincts.
The large straight-edged mandibles of the kelep are well adapted
for holding or carrying objects of considerable size. Ants with
smaller jaws could not catch and sting boll weevils, because they could
not. hold them in the right position. Numerous ants may be said to
attack boll weevils, but are able only to seize them by the legs or
snouts and drag them about for a time. As soon as released the
weevil escapes by its well-learned trick of feigning death. The
kelep can pick the weevil up bodily, and in normal positions on the
stemins of cotton plants can usually sting it while the weevil is still
"playing possum," before it begins to struggle. It is accordingly
not without interest that the large mandibles and long legs which
qualify the kelep as a weevil destroyer have other important func-
tions in the social economy of the species.
It may be mentioned also, as showing the special efficiency of the
keleps in moving their domicile, that they seldom carry one egg alone,
but will often go to considerable pains to pick up two or three at a
time. Sometimes the eggs adhere in considerable nuinmb)ers, though
not as in O idntomrhi .s, where they are cylindrical and are made up
into compact bundles. Eggs, pupa, and other objects are frequently
pushed between the mandibles by the end of the abdomen, brought
up from below as in stinging the boll weevil. Two cocoons are some-
times packed up in this way, so that they can be carried together. It
is also by virtue of this same flexibility of the abdomen that the
workers are able to assist themselves with their mandibles in the lay-
ing of the undersized eggs which they occasionally produce.

A colony with two queens spontaneously divided into two by the
departure of some of the workers and one of the queens. This inci-
dent was observed in Victoria, Tex., by Mr. Frederick L. Lewton, the
essential facts being stated as follows:
On July 22 the ants in colony No. 26 exhibited signs of restlessness, and as
several had been found dead a new nest was prepared in a larger jar. The
two jars were connected by means of a cotton-leaf bridge. The dead ants were
immediately carried over to the new jar and also the hard re iains of several
boll weevils. A considerable amount of earth was also carried from the old
nest to the new, but few ants remained in the new jar until its lower half
was protected from the light. The protection was renioved from the old jar,
which was also allowed to become dry, in order to hurry the keleps in removing
the colony.


At the beginning of the experiment a number of eggs and larv.e were on the
surface of the soil in the old nest. One larva was almost fully grown and
about ready to pupate. In twenty-four hours all the eggs and larvae, except
the large one, had been removed to the bottom of the new jar, while the one
remaining in the old nest had been taken below and covered with loose earth
preparatory to pupating.
In another twenty-four hours there seemed to be almost no communication
between the two jars and a close inspection wvas made. Then it was found that
the old colony had contained two queens. One of these had taken possession of
the new nest. accompanied by about half the workers and all of the eggs and
young hLI rv;e. only the full-grown larva and the workers remaining with the other
queen in the old nest. This she would not desert, in spite of the strong light
and the dry ness. The jir was therefore moistened and protected from the
light, and within five days thie queen began laying again. The one larva began
spinning its cocoon on July 23, and it was finished in about three hours.


Several kelep colonies containing eggs and( larva were very care-
fully explored( in Guatemala without finding queens. In view of the
simple structure (if the nests and the extremely favorable circum-
stances under which thle work was done, it is not considered probable
that the queens escaped notice in all these instances. It seems more
likely that some of these colonies had been established recently with-
out queens.

Captive workers alone, as already stated, become listless and con-
fused. but if supplied witlh eggs their I)ehavior soon becomes normal.
Several queen larva' have been raised in the laboratory at Victoria,
Tex., the first two in a colony in the hands of Mr. W. E. Hinds, the
others by Mr. Argyle McLachlan. One might argue from Mr. HKinds's
exp)erimellnt great intelligence oi thle part of the keleps. One of two
colonies had lot its queen by an accident, bult was supplied with eggs
from another onl the supposition that a queen would be raised to make
good the deficiency. The keleps appeared, however, to have no spe-
cial desire to have a queen to suppll)l)ort as long as they could secure a
regIlar supply of eggs, and( continued to raise workers only. The
other colony, however, from which the eggs were being removed, took
measures at once to raise a new queen, the deficiency of eggs having
given, perhal)s, the impression that the fecundity of their queen was
declining. Two larva? of unusual size were raised, one of which
emerged as a normal winged queen. Mr. Hinds found that the time
required for the development of a queen is about three months, the
larval and pupal stages being about one and one-half times as long as
those of the workers.
Two queens raised by Mr. McLachlan in January and February,

1905, spent 35-1 days in their cocoons, a period distinctly less than that
of two worker pupa? of the same colony at the same time, which took
40 and 39 days, respectively. These queens were raised in the same
queenless colony previously referred to, after the supply of fresh eggs
was cut off. This would seem to show that a scarcity of eggs rather
than the bodily absence of the queen may be neces-ary to set the work-
ers to the development of a new queen. Certain it is that there is no
such obvious reaction to the presence of a queen as among the bees.
If the accounts of Maeterlinck and other observers are to be credited,
there would seem to be among bees either a distinct system of royal
etiquette or some direct unexplained influence upon the actions of the
And one may mention here the curious fiact that the workers always avoid
turning their backs on the queeii. No si,'ner liis she alppro;iched a group thlin
they will invariably arrange themselves so as to face her with eyes and
antenllna, and to walk backwards before her. It is a token of respect, or of
solicitude that, unlikely as it miay seen, is nevertheless consta lt and general.a


Insects of several natural groups might be described as social. The
larvae of certain Lepidoptenira remain together and even make com-
munity webs or cocoons. Some of the bark lice (Psocida) also pas-
ture in drove-, and in their younger stages spin continuous silk can-
opies over their feeding place-. Thee communities, while obviously
held together by social instincts, are in no proper sense to be termed
organized, no division of labor or diversity of form or structure hav-
ing arisen as a result of the merely gregarious association of indi-
viduals. Among the Hymenoptera some of tlie of the family
Eumenide repre-ent this primitive social condition. Maeterlinck has
traced the social development of the honeybee.b
For an adequate outline of the diversity of breeding habits it seems
necessary to recognize at least six types of social organization among
insects. The contrast, between the different biological conditions
found in the several forms of insect society can best be shown by brief
formal descriptions.
a Maeterlinck, M., 1901, Tlie Life of the Bee, p. 214.
b We find even to-day, among the melliferous Hymenoptera, all the stages of
progressive civilization of our own domestic bee. At the bottom of the scale
we find her working alone, in wretchedness, often not seeing her offspring (the
Prosopis, the Colletes, etc.) ; sometimes living in the midst of the limited family
that she produces annually (as in the case of the humblebee). Then she forms
temporary associations (the Panurgi, the Dasypodae, the Halicti, etc.), and at
last we arrive, through successive stages, at the almost perfect but pitiless
society of our hives, where the individual is entirely merged in the repub-
lic * *."-Maeterlinck, M., 1901, The Life of the Bee, p. 31.



Interbreeding maintained by a simultaneous annual concourse of all the recently
matured insects of both sexes.
Colony founded by a pair of sexual insects which remain permanently mated.
Wings provided with transverse sutures which enable the insects to break them
off readily by an upward bending of the abdomen. Both sexes shed their
wings in this manner as soon as they have associated in pairs, and then
begin digging in the ground.
Workers of both sexes always wingless, numbered by thousands or even by
millions, generally of diverse form, or'of two or more distinct castes-sol-
diers, nurses, etc.
Several of these features are not shared by any of the types of
organization found among the bees, wasps, ants, or other Hymenop-
tera. It is only among the termites that the males are regularly
found with the females in the nests. Among the Hymenoptera the
males are relatively short-lived and*take no part in the work of the
colony. Sometimes they are not even tolerated in the nests. A single
fecundation suffices for the lifetime of the queen bee or ant, while
among the termites the process is probably repeated. Copulation
has never been observed among the termites; it does not take
place during the mating flight, as among the bees. The wings of
Ilyinenoptera are not provi(led with sutures to render them easily
detacliable. MIales are permanently winged in all the groups; in
some the workers are wingless, and in some the females also. Among
tlhe ants the queen is artificially wingless, as in the termites. She bites
off her own wings after the marriage flight.
Perhaps the most fundamental peculliarity of the termite organiza-
tion lies in the fact that the workers., soldiers, nurses, or whatever the
various castes may b)e called, are of both sexes, instead of being unde-
veloped females only, as among tlihe bees and ants. Whether a young
bee larva shall develop into a queen or become stunted into a worker
or so-called neuter depends upon the quantity, and perhaps also
the quality, of the food supplied to it. Among the honeybees special
brood cells are prepared in advance for the males, the workers, and
the queens which the colony proposes to raise.


It has been supposed that the differentiation of the members of
termite colonies is governed in the same manner, by nutrition merely,
but this is rendered very doubtful by the fact stated above, that the
workersof all the different castes represent stunted individuals of both
sexes. Another serious weakness in the nutrition theory lies in the
fact that the larvT of the termites are not mere helpless, inactive
grubs, as in all the social Hymenoptera, but are quite as capable of
locomotion as the adult workers, and are always traveling about to


all parts of the nests, so that the application of a consistent regimen
of feeding would become, humanly speaking, impossible.
In some of the relatives of the kelep native- in Texas, Professor
Wheeler has found that there is no clear distinction between the
fertile females and the workers, a condition which he ascribed to the
desultory feeding habits of the insects.a The deficiency may be con-
nected, however, with the more backward social organization of the
Texan species. The keleps, with the same predaceous habits, seem
always to be able to produce the two well-defined types; at least no
intermediate forms ha\'e been observed.
The question has not been put to an experimental test as yet, but
the suggestion has arisen in connection with the kelep that the de-
velopment of a larva into a worker may be brought about in the early
stages by withholding food, it having been noticed that the larvae
grow very slowly at first, but after they reach a certain size they often
coml)lete their development with great rapidity. A superabundance
of food or a special solicitude for the younger larvae might induce
the development of queens without the necessity of supposing that
the insects proceed by any deliberate intention.
But even among the keleps the worker caste is hardly to be
explained merely by inadequate nutrition, for the reason that it
possesses positive as well as negative characters and instincts which
the queens do not have, and which they probably never had in the
evolutionary history of the species. As previously stated, it is well
known that among the bees and ants the nervous system, and
especially the part which corresponds to the brain of the higher
animals, is much more highly developed in the workers than in the
sexually perfect insects, and this is accompanied by the accentuation
in the workers of the social instincts by which the colony is main-

The more recent evolutionary changes among social insects have
been largely in the direction of the workers, the males and females
a Now, while we can, perhaps. understand how these more specialized ants
may manage to control the quantity and quality of liquid food regurgitated
from their own crolis and salivary glands, it is not so easy to understand how
ants can exercise such control when they adopt a capricious nietliod of feeding
like that of the Ponerina. Such a method can hardly pro(luce clear-cut results,
I. e., either workers or fertile females. And a comparative study of the better-
known species of Ponerinw shows that in certain species, at least, there is no
such sharp distinction between the sterile and fertile female as we find in the
more specialized ants. Not only is the female sex in a state of miorphological
and physiological in stability-i. e.. di- or even tri-norphic-but the male sex
also is sometimes dimuorphic, at least in the same genus, if not in the same
species."-Wheeler, W. M., 1901, Biological Bulletin, 2:28.

28929-No. 10-05 iM---1


retaining more nearly the characters and habits of their nonsocial
ancestors. The development of the workers has not been confined,
however, merely to the more highly specialized nervous system and
instincts. In many of the ants and termites there are two, and some-
times three or four, distinct kinds of workers, very different from
each other and equipped with an equal diversity of instincts for
undertaking special parts of the work of the social organization.
One of the African termites has soldiers with immensely large heads
and biting mandibles, whose sole function is that of defending the
colony. When the wall of the nest is broken they sally out to en-
gage the intruder. They are wholly devoted to this purpose, and
lack even the instinct to return again, and soon perish from exposure
to the drier, outside atmosphere. A second form of worker, also
with large heads, but much smaller jaws, will not leave the nest to
fight, but appears to control the erection of new walls, though it does.
not itself carry earth, a work performed entirely by the third
caste, which might be called the worker proper. These build the
nests and bring in the dead wood, which, after a period of curing,
is made into fungus gardens. The fourth caste is small and slender,
and neither fights nor carries earth, but is occupied with the manage-
ment of the eggs and, presumably, also with the feeding of the young
We may be sure that this high degree of specialization of castes
hlias come about during the period of social organization. How the
queens are able to give birth to these distinct kinds of offspring is
unexplained as yet, either by difference of nutrition or by any of the
theories of hereditary transmission applied to the other groups of
Professor Weismann has appreciated the inadequacy of the nutri-
tion theory, even with reference to the honeybee, and has attempted
to explain the worker caste by means of his well-known theory of
a" This explanation [nutrition], however, even if correct as regards the degen-
erated parts of the workers, does not sufficiently account for the other differ-
ences between the two kinds of females. For the workers are not inferior to
the queen bee in all respects. On the contrary, the worker's sting is straighter.
longer, and stronger, and is provided with more teeth than the queen's. The
wings, moreover, are longer, the tarsal segment of the hind limb is provided
with the well-known brush, and the tibia has a depression known as the
'pocket' for carrying the masses of pollen which the insect collects. These two
characteristic parts are wanting in the queen. Important differences must also
exist as regards the minute structure of the brain, for the instincts of the
queen are very different from those of the workers. After fertilization has
taken place the queen lays eggs, but she neither gathers honey from flowers,
excretes wax, nor makes the honeycomb. It is therefore incredible that the
queen and workers should be formed by the agency of similar determinants.


It is apparent, however, from what has been said regarding the
termites of Liberia that Professor Weismann has greatly underesti-
mated the difficulties which his theory of heredity would encounter
in that group. As the various castes are not separated on sexual
lines, there are five forms of each sex, or possibly six if complemental
or substitution royalties are developed. Instead, therefore, of three
or four kinds of determinants there would need to be ten or twelve
kinds, and since it would be necessary to provide in the same way for
the preliminary stages of the various castes more than a score of
different determinants would need to be predicated-too many, it
would seem, to leave us much confidence in this method of accounting
for the existence of the worker caste.


Interbreeding maintained by a simultaneous amnual concourse or mating flight
of all the recently matured insects of both sexes.
Colony founded by a solitary fecundated female.
Wings present on young females, but bitten off after the mating flight.
Workers always wingless, often to be numbered by thousands, commonly of
diverse form or of two or more distinct castes. Workers of different col-
onies of the same species actively hostile.
Numerous deviations from this type of organization are to be ex-
pected among the various families of the tropical ants, the habits of
which are still very imperfectly known. The colonies of many ants,
too, are compound, a part of the young females remaining in the
parental nest. Miss Fielde describes such a colony of Stenamma

The germ plasm munist contain double determinants for certain parts of the body
of the queen and workers, respectively. But as we have already assumed the
existence of double determinants for the formation of male and female bees,
or at any rate for the development of those parts which differ in the two sexes,
we can only make the further assumption that the 'female' halves of the
double determinants miay themselves consist of two halves, corresponding to
the queen and worker, respectively, and that each of these halves must naturally
be looked upon as a complete determinant as regards size and structure. * *
In the case of bees the factor that determines which of the two halves of the
'female' determinant is to become active seems to be the quality of the food
supplied to the larva, so that the critical moment only arrives long after the
termination of embryogeny and before the chrysalis stage is reached. * *
We might, however, also assume the existence of three independent determinants
side by side, so arranged that they become active under other definite influences,
and this conception would better agree with the unavoidable assumption that
the three determinants which act vicariously are of a similar size. * *
The termites, in addition to the workers or stunted females, possess "soldiers,"
or males in which the sexual organs are stunted, which possess very strong
mandibles and differ in other important structural details from the ordinary
males. In this case, therefore, four determinants must be present, each capable
of being substituted for another, and only one of which can be active at a time."-
Weismann, A., 1893, The Germ Plasm: A Theory of Heredity, 377-379.


fulrvinnm scattered over an area 90 yards in diameter, and states that
she has been unable to find any other colony of Stenammas in its
vicinage."' a

Interbreeding maintained by free flight of males during the period of emergence
of young females.
Colony founded by a solitary fecundated female, the males and workers not
surviving the winter.
Workers permanently winged like thle sexual adults.
The organization of many of the social wasps is essentially the
same as that of the bumblebee. This close similarity of social habits
has been accepted by many entomologists as an evidence of relation-
ship between the two groups, but the present tendency is toward the
view that the social organization has been attained independently.


Interbreeding maintained by the free flight of males during the swarming sea-
son, when they are permitted to feed in any of the nests. The young
queens emerge from the nests for a nuptial flight, in which fecundation is
Colony established by subdivision of the workers of an older swarm, accom-
pained either by the old queen or by a young queen still unfertilized.
Males short lived, usually not tolerated in the nests except during the period
of emergence of newly adult females.
Females permanently winged, but the wings used, normally, only in the nuptial
flight and in that of swarming.
Workers permanently winged, of one caste, or of slightly different castes, nuim-
bered by thousands. The workers of different colonies are not actively
hostile except to marauding intruders.


Interbreeding probably maintained by the visits of the males to the various
adjacent nests; no concourse or nuptial flight of females.
Colony established by a body of workers from an older community. The work-
ers carry eggs, larvae, and queen with them for the complete equipment of
the new nest.
Males permanently winged, but apparently not inclined to fly; tolerated in all
nests, and probably present throughout the year.
Females winged at first, but not known to fly or emerge from the nest.
Workers wingless, all of one form or caste, numbered by hundreds; not actively

In other genera of the same family normally wingless queens are
frequently found. In some only wingless queens are known; in others
some are winged and others wingless, with no differences in other

a Fielde, A. M31., 1904, Biological Bulletin, 7: 245.


respects. Fecundation inside the nest has also been observed in a
member of the Poneridap, or kelep family, as already noted.
Though no doubt attained independently, the resemblance of the
kelep social organization to that of the bees is fully as striking as the
parallel between the ants and the termites. The most serious dif-
ference lies in the fact that the queen bee must go out of the nest to
be fertilized. Both the keleps and the bees are able to raise new
queens as long as eggs are obtainable, but the danger of failure is
much greater with the bees because of the vicissitudes of the marriage
flight to which the young queen is subjected. In the habit of an open-
air mating the bees resemble the butterflies, dragon flies, and other
nonsocial insects. Mating inside the nests represents a more advanced
and much more practical social system.
It is not difficult to assign a reason for a frequent subdivision of
colonies as a feature of the social economy of the kelep. A large col-
ony of predaceous insects would soon exhaust the insect fauna of its
immediate neighborhood. It could maintain itself only as the drivers
do, by foraging widely and changing quarters frequently. The keleps
do not go, apparently, more than a few feet from their burrows. They
seem ready to move their more favorable locations, but these
migrations are also, in all probability, for short distances only,
though the Guatemalan Indians say that the keleps will come in from
neighboring areas and occupy a cotton field after the crop has bean

Interbreeding maintained by annual emergence and free flight of large numbers
of males.
Colonies nomadic, probably increasing by the subdivision of migratory hordes.
Larvw carried about by the workers for long distances..
Females wingless, probably never emerging from the nest unless carried by the
Workers wingless, of two or more castes-soldiers, nurses, etc. Numbered by
hundreds of thousands, or millions.

The African "drivers and the army ants of the American tropics
represent another group of social insects popularly associated with
the true ants because the workers are wingless. In Liberia the males
sometimes fly to lights in the evening in large numbers, which may
indicate a definite breeding season for the sexual adults. Most Do-
rylidae have been described from workers and males, very few females
being known to science. The winglessness of the females may be said
to advertise the distinctness of their social economy from that of the
true ants, and to ally them with the Mutillidae. Doctor Ashmead has


also pointed out to me the strong similarity of the males of Dorylus
to those of the wasps of the East Indian Vespa doryloides Saussure.a
The scant evidence thus far obtained seems to show that a colony
of drivers has only one large, distended queen, like that of the Ter-
mites. Emery notes that the last joints of the legs are usually
broken off in the specimens known to him.b Such mutilation may
well arise while the huge creature is being dragged about, as stated
by the native Africans.
The interesting feature of the drivers, from the standpoint of the
study of the kelep, lies in the fact that the differences between them
and the true ants are in several respects the same, though much more
accentuated than between the ants and the kelep. Thus, interbreeding
is accomplished in both groups by the circulation of the males, the
females, apparently, never emerging from the nests. The drivers
carry their larvae, and presumably their queens also, for long dis-
tances, their close ranked, rapidly moving columns sometimes requir-
ing several hours to pass a given point. In food habits, too, they
are carnivorous, like the keleps. They not only capture insects and
other arthropods, but young or helpless vertebrates also. Active
creatures even, like rats and mice, are often secured, the pain inflicted
by the bites of the drivers causing them to roll about on the ground
and thus l)ermit more and more of their diminutive enemiies to
attack them. Any animal too large to be carried away by a single
worker is cut in pieces. The kelep follows the more social plan of
moving large insects and( centipedes by the combined strength of
several workers. The drivers are also like keleps in having a sting
which they do not use as a weapon of offense against man, but they
bite viciously and will not let go, even though their bodies be pulled
loose from their heads.
The association of the keleps with the drivers permits a better
appreciation of their habits in two respects. First, the apparently
high development of their instinct of place or ability to find their
way about, even in new situations. The drivers are fierce warriors,
a Doctor Ashmead informs me that the males of the tribe Amblyponinae, near
relatives of the kelep, are very closely similar to those of the Mutillida, even to
the structure of the petiole and the venation of the anterior wings.
The flying to lights of the males of the Dorylide, or drivers, is also shared by
the males of .lyrnmosa and related genera of parasitic wasps formerly placed in
the M1utillid.,. but which Doctor Ashmead now separates as a distinct family,
Myrmosidr. intermediate between the Mutillide and Vespidoe, but having wing-
less females like the Mutillidae and the Dorylidu?. Some of the Bethylid2e also
have wingless females, but these are not supposed to have any close affinities
with the other families. The Thynnidwe, which have been thought to connect
with the Mutillidai and the ants, likewise have wingless females. Finally, a
few of the Mutillida are wingless in both sexes.
b Emery, E., 1895, Ann. Soc. Entom., France, 64 :73.

which actively pursue and capture their prey. They are to be
described as predatory, while the keleps are merely predaceous, and
like many other predaceous animals their movements are slow and
deliberate except when they make a final spring to seize their game.
The slow movements of the kelep thus conduce to its efficiency as a
hunting insect, and may be looked upon as an evidence of specializa-
tion in this direction. Odontomachus and the Texan Poneridae, so
far as observed, are quicker and more excitable under ordinary cir-
cumstances, but they seem not to have developed any such skill in
capturing quick-moving insects. The similarity of the keleps to
the drivers is greater, however, when the colony moves and the eggs,
larvae, pupae, cocoons, and queens are carried over to the new burrow.
In one instance a few of the workers refused to leave the old nest.
but the majority would not permit the colony to be divided in this
way and finally captured and carried off the unwilling members in
the same way that the queen is removed. One of the Brazilian rela-
tives of the kelep is described by Bates as marching about in the
forests in a manner which must give them a considerable resemblance
to the drivers.a
It is true, of course, that the keleps and the drivers are very dis-
tinct, but careful comparison will show that the differences are in
degree rather than in kind. They represent possible developments
from the sante primitive system of organization, while to find a
common starting point with the ants it may be necessary to go back
to the nonsocial stage. Structurally, too, the keleps are very differ-
ent from the drivers, but the genus Cerapachys and its allies are
referred by some authors to the Poneridam and by others to the
Dorylidae. It should be no more surprising that some of the habits
of the true ants should be approximated by other groups of inde-
pendent origin than that the bees and wasps, or that the true ants
and the termites, should be so similar.
The idea that any existing group of organisms has given rise to
another group is generally, and probably always, erroneous. With
respect to any given character, one group may remain primitive while
another has advanced, but nothing organic is stationary, and the fact
that a plant or animal is primitive in one respect may well be taken
as a suggestion that good progress has been made in some other.
Looked upon as ants, the kelep and its relatives are undoubtedly
primitive, since they are more like wasps than like other ants, but the
a We were amazed at seeing ants an inch and a quarter in length and stout
in proportion lm.irching in single tile through the thickets. Tliese belong to the
species called Dinoponera grandis. Its colonies consist of a small number of
individuals, and are established about the roots of slender trees. It is a sting-
ing species, but the sting is not so severe as in many of the smaller kinds."-
Bates, H. W., 1875, Naturalist on the River Amazon, p. 9, ed. 4.

theory that their social organization must also represent the primitive
condition of the true ants seems not to find confirmation in fact. The
relatively small size of the colonies does not make them homologous
with incipient ant colonies. They represent a different system of
social organization. An increase in the size of the colonies of the
keleps would not carry them in the direction of the true ants, but
toward that of the drivers, unless they changed their habits and
became vegetarians.a
These various considerations do not exclude, of course, the possi-
bility that some of the families now included among the true ants
may have descended from Ponerid stock, but if so we may expect to
find that their breeding habits are not those now ascribed to ants as
a natural group, it being highly improbable that insects which had
become completely socialized, like the keleps, should abandon this
condition and return to the relatively crude and wasteful system fol-
lowed by the ants and termites.


Care has been taken in the preceding diagnoses of the types of
social organization to refer to the simultaneous annual emergence
of winged individuals of both sexes as a concourse, reserving the
term swarm for its original use with the bees, and for an entirely
analogous application to the kelep.
With the bee and the kelep the swarm is made up of workers
(undeveloped females), and it is obviously incorrect, or at least very
inexact, to apply the same word to a collection of winged adults
of both sexes, as among the ants and termites.
Swarming, as with the bee and the kelep, is a process which does
not take place at all in the other types of social organization. It
results in the foundation of a new colony by the spontaneous subdi-
vision of the workers of an older community, an occurrence not known
among the termites and true ants.
The simultaneous emergence of the newly matured males and
females from all the nests of the species has for its object cross-
fertilization. It is a time specialization not attained by the honey-
bees nor by the keleps, among which facilities for interbreeding are
maintained throughout the season. The swarming of the bee and
the kelep has, in fact, no connection with the cross-fertilization of the
a If the laying of few eggs and living in small colonies is to be looked upon as
a primitive character, the genus Ponera is the most backward of the American
series s of Ponerid;e and the kelep the most advanced. Such a criterion can not
be applied at random, however, since specialization may tend toward restricted
fecundity as well as in the opposite direction. Whether an insect would prosper
best in colonies of 10, 100, or 1.000 would depend on its habits and environment


sexual insects, which may take place either before or after the division
of the colony.

An insect colony in which all the eggs are furnished by a single
laying queen is a strictly determinate organization; that is, it reaches
a natural limit after the mother insect dies or ceases to produce. The
queens of ants and termites are sometimes enormously fertile and live
for many years, so that determinate colonies may sometimes attain
large size.
Colonies may be called indeterminate when the social economy of
the insect is such that a lost queen can be replaced. This is the case
with the honeybee, which may be said to have a simple indeterminate
colony because each hive has but a single laying queen. Colonies
with more than one egg-producing queen may be called compound
indeterminate. This condition has been reached in two different
ways. Some of the termites, for example, are able to replace a lost
queen by what are called complemental queens, individuals which are
brought to a precocious sexual maturity, but without becoming fully
winged. Other termites, such as the South European Termes b,1i-
fugus studied by Grassi, seem to have gone over entirely to the com-
plemental queen basis, thus reaching a completely indeterminate
organization. With the ants the saiiie result has been reached in a
slightly different manner. Some of the young queens are fertilized
and drop their wings before emerging from the parental nest, and,
being tolerated by their mother and sisters, they remain to contribute
to the numerical prosperity of the family.
A condition very similar, superficially, to that of the clustered col-
onies of the keeps is to be found in certain ants and termites which
have the habit of making what might be called multiple nests. In-
stead of constructing a single compact dwelling, a colony may spread
out into several distinct burrows or nests, connected by galleries,
sometimes very long. An instance of this among the ants has been
cited already on page 27. In one species of the termites, a single col-
ony often builds numerous nests of considerable size on widely sepa-
rated branches of the same or neighboring trees in the forests of
Liberia. The queen, however, does not live in these aerial summer-
houses, the eggs being carried up by the nurses through the galleries
which connect with subterranean burrows.
It is conceivable, however, that the compound indeterminate colony
might have originated in a more direct manner among insects which
have never passed through the determinate stage of social organiza-
tion. Social organization probably began, as already suggested, with


a condition of gregarious association of equal individuals. With
the ancestors of the termites and ants this became gradually more
centralized -and specialized until a completely monarchical system was
attained, which is still followed, at least in the early stages of new
With the kelep, however, there is no indication of a social history
of this kind. The primitive society of equal individuals became
specialized into two castes, the workers and the layers, so to speak,
but there is nothing in the habits of the insect to indicate that there
ever was a definite restriction to one fertile female for a colony.

With the ants the colonies are founded by solitary queens and held
apart byl definite instincts of hostility, but the keleps separate into
small colonies merely for economic reasons, as it were, to find new
feeding grounds. The colony itself is a different social phenomenon
in the two groups. Viewed from what might be called the geo-
graphical standpoint a colony in a new place is a new colony, how-
ever planted. but viewed from the standpoint of the insects and their
-oci-1l orgaiiizationl there is a fundamental distinction. A new colony
of auits is really a new social group, a distinct family, but there are
nIo such things, apparently. as lnew" groups of keleps; there are only
sIIl)(ldi 'visions of older groups. The female or queen ant leaves the
old grulp) and exists for a considerable period as a solitary inde-
pelldenlt insect. bItlt there is no provision inii nature for a solitary kelep.
The keep i'- conipletely socialized, if such an expression may be per-
nmitted, while thle ant is not. Indeed, it may lbe claimed that the
social organization of tlhe kelep i, more perfect than that of any
insect whose life history lhas become known thus far. It is less than tlihat of the honeybee, l)erhaps, but is more complete,
for even with the bees the young queens are obliged to leave the hive
for the nuptial flight.
Complete socialization ilnvolve.s, apparently, the transfer of the
chief responsibility, the social center of gravity, as'it were, from the
queens to the workers. The colonies of the bumblebees, wasps, and
termites are governed, or at least founded, by the queen or mother;
they represent the social l)rinciple of matriarchy, while colonies of
the bees, keleps. and probal)ly the divers also, are founded and man-
aged by the workers. In the matriarchy of the bumblebees, according
to Maeterlinck, the queen-mother has to protect her eggs to prevent
their being destroyed by her older daughters, but in the ergatarchy
of the honeybees the queen is restrained by the workers from destroy-
ing her own fertile progeny and possible successors, against whom
she entertains sentiments of deadly jealousy. It is not to be sup-
posed, of course, that the queens of the ants and termites issue orders


or actively control the colony when once established, but even then
there is an important difference in social economy, as pointed Out in
another place. The success or vital efficiency of a colony of ants or
termites depends upon the number of sexual insects or royalties it
can produce, but among the bees and keleps prosperity is measured,
primarily, by the number of workers, which have become quite as
necessary as the queens in the establishment of new communities.
The founding of a new colony among the ants and termites is an
incident or direct sequel of the proce-s of sexual reproduction. The
sexual insects emerge simultaneously from the different nest, of the
species for their mating flight, but do not return, the few survivors
remaining to found new colonies where chance has carried them.
The workers do not figure in the srarlrming process at all; they func-
tion merely as nurses to assist the queen in raising their sexually
mature brothers and sisters.
The honeybees have modified this programme to the extent that
the young queen does not attempt, by herself, to establish a new com-
munity, but returns to her family after the mating flight.a The
keleps, however, have achieved a much more complete separation of
their political system from strictly reproductive complica t ions. The
foundation of a new colony not only does not depend upon the initia-
tive of the queen, but may be undertaken quite without her presence.
Cross-fertilization or interbreeding among the members of the
species is undoubtedly a necessity for the keleps, as for all other highly
developed orgai'ismi., but it may not be nearly as.acute a requirement
as among the termites, honey1bees and true ants. whose social svytems
make far greater demands for fecumndity on the part of the queens.
If, on the analogy of the bees., the e .gg, of an unfertilized queen will
produce only males, the deficiency of the latter sex would be auto-
matically made good for that particiilar neighborhood at leist.
The predaceous habits, the winghle-,iss'of the workers, and the
lack of any instinct to forage a long distance from the nest forbid
the formation by the kelep of large communities like those of the bees,
but this need ilot keel) us from appreciating the advantage. of the
kelep organization, which by its greater mobility enables the insects
to keep nearer the source of supplies without being obliged to waste
so much energy as do the ants in carrying in their food, and without,
like the drivers, giving up a settled existence altogether.
It becomes, also, even more obvious than before that the keleps may
prove to have only the most superficial similarity to the ants, and
that the social systems of the two groups may be of entirely independ-
ent development. The striking parallel between the bees and the
a The imnpregnation of the young queen bee is said to take place occasionally
during the swarming flight.

keleps does not prove, of course, a common descent from a social
ancestor any more than does the agreement of the ants with the ter-
mites. or that of the bumblebees with the wasps. There are still
many solitary bees among which a colony organization might grow
up; also solitary wasps, some of which associate gregariously; and,
finally, there are solitary wasps with wingless females which have
undoubted relationships with the drivers and the keleps. The ant
organization can hardly be looked upon as a further specialization
from this series. The breeding habits prove conclusively that the
history of the ants involves an independent specialization, probably
from some nonsocial ancestry, but certainly not from one which had
attained any such organization as the keleps or the drivers, or any
such tendency to winglessness in the female sex as shown in the
Mutillide and other related groups of parasitic wasps.
The feeding of the larva with captured insects by ants of the fam-
ily Myrmicid&e has been cited by Professor Wheeler as evidence in
support of a suggestion of Professor Emery that this family of ants
was derived from the PoneridT.a But even if this correspondence be
accepted as a true indication of relationship it can scarcely mean more
than that the ancestors of the ants were predaceous insects like the
keleps and the drivers. The latter groups have kept the predaceous
food habits, but among the ants there is still to be found the even
more primitive instincts of an annual emergence of both sexes from
the nest and the founding of new colonies by solitary females.
On the other hand, it must be admitted that the habits of the genus
5ti&0ma tominia, as described by Professor Wheeler, offer many more
points of resemblance to those of the kelep than do those of Ponera
and the Texan genera Parhycondyla and Lepto.qenys studied by
Professor Wheeler. The larva of Stigmatomma, which does not
coiform to the Ponerine type," is much more similar to that of the
kelep. as are also the eggs.
,"There still exists an ecological (or biological, in the German sense) con-
nection between the IPonerinme an(i the Mymiciniie, as I have lately ascertained.
Since describing the peculiar method employed by the Texas Ponerine in feed-
ing their larvie, I have found, that one of our New England Myrmicine ants,
SICaumin, I .phiu'ogastir) ful' ion Rog., subsp. aquia Buckl., var. piceum
Eiery-ani ant very common under stones and in rotten logs along the edges
of wood,-has essentially the same method of feeding its young. My attention
was fir-4t called to the f:ilct ini an artificial nest belonging to Miss Adele M.
Fielde, at Woods Holl, Mass. One afternoon Miss Fielde left a lot of queen
ipupl and larvaie of Crenastogaster lincolata within reach of the Stenamma
colony. By the following morning the Stenammas had carried these into their
nest, cut off their "heads and abdomens, and had distributed the pieces freely
4,mong the larvn, which could be seen singly and in groups of from two to five
eagerly feeding on the juices in the same manner as Ponerine larva,."-Wheeler,
W. M., 1901, Biological Bulletin, 2: 05.

It becomes apparent that the species of so-called Ponerida or Po-
nerinae of which the habits are known belong to five different natural
groups which are likely to be recognized at some future time as dis-
tinct families; at least the differences of habit and social organiza-
tion are as great or greater than those of many families or orders
among the higher plants and animals. It is becoming more and more
appreciated that such characteristics afford as good or better indica-
tions of evolutionary history than do mere differences of structure
and proportions. Indeed, the evolutionary significance of a structural
character can not be determined until the biology of the plant or ani-
mal is known-that is, its mode of existence and place in the economy
of nature.
There may be said to be two methods of classification, correspond-
ing to the two states of knowledge regarding a group of animals or
plants. Under what may be called the analytical system, classifica-
tion is based primarily on a logical analysis of formal characters, the
various groups, such as orders, families, genera, and species being
reached from the top downward by repeated subdivision. The con-
crete or synlthetic method reverses this procedure. Coherent groups
of normally interbreeding individuals are recognized as constituting
species; closely allied species are grouped into genera, and genera into
Owing to their vast numbers and diversity, the classification of
insects is still largely in the analytical stage, so that recognition of
the kelep as representing a distinct family would not be consistent
with the cuiirrent classification. Nevertheless, it may be fully expected
that in due time the insect groups will be treated in better accord with
the methods by which the mammall, birds, and other higher animals
are classified.
The five groups referred to above may be typified by the genera
Sti(matomm,. O idntomwnachus, Ponera, Pachycondyla, and Ecta-
tonmmna. The first and second groups have already been recognized
as families or subfamilies becaui.,e of structural peculiarities of the
adult insects, the Amblyponin a (including Stigmatomma) having
no deep constriction behind the node or first abdominal segment.
The Odontomniachidwl are characterized by their remarkable mandi-
bles, which are inserted close together near the middle of the head
instead of at the anterior corners, as in all the related groups. The
true Poneridae, as represented by the typical genus Ponera, show the
most backward stage of social organization and have very small
colonies. The larve are provided, according to Professor Wheeler,
with four pairs of peculiar adhesive dorsal tuibercles, supposed to
have the function of holding them in position against the sides of the
nest chamber. In Pachycondyla and Leptogenys the larve are, like

those of Odontamachws provided on each segment with a transverse
row of conical tubercles, each surrounded with a circle of hairs, and
except for these the surface is naked. In Ectatomma, on the con-
trary, the larvae, like those of Stigmiiatomzma, have a uniform coat of
fine hairs. In Odontomachida, and Pachycondylidw the eggs are
cylindrical-oblong and adhere in bundles; in the Amblyponidte and
Ectatommidaw they are elliptical and adhere only in loose and irregu-
lar masses. The Amblyponidxe, Ponerida, and Ectatommide exca-
vate subterranean chambers, while the Odontomachida? and Pachy-
condylidf, make, as far as known, only irregular anastomosing


On consulting Dr. William H. Ashmead, of the United States
National Museum. the eminent specialist in the study of the Hymen-
optera, it was learned that in a still unpublished revision of the clas-
s.ification of the ants and allied insects he has recognized all these
five groups as representing distinct families, subfamilies, or tribes
on the basis of structural characters of the adult insects.
As this new arrangement lund(loubl)tedly represents an important
advance toward a natural and adequate classification of the kelep and
its relatives, Doctor Ashmead was requested to l)ermit the publication
in this l)lace of a brief extract from his manuscript, and he has kindly
consented. The following analytical keys show the characters by
which the kelepl) and its immediate relatives may be distinguished
from other subfamilies and tribes.


1. Hind til)i;Pe with ir'o apical spiurs------------------------------------- 2
Hinid til)ial usually with Mui al)ical spur. rarely without an apical spur;
claws simple, or p)ectinate----------------- Subfamily I. PONERINE
2. Abdomen with only one joint to the petiole, the constriction between the
second and third segments very shallow, the second segment therefore
not nodiform ; claws simple, cleft, or with a tooth beneath.
Abdomen with apparently tiro joints to the petiole, the constriction between
the second and third segments very deep and strong, the second seg-
anent therefore nodiform, nearly as in the Myrmicidte; claws cleft or
with a tooth beneath ------ ---- Subfamily III. MYRMICI[NE. a
a This subfamily represents a peculiar genus of ants found in Australia re-
sembling the genuine Myrmicidm, and formerly classified with them, on account
of the two-jointed abdominal petcole; but they are undoubtedly Ponerids, and
Forel and Emery are right in removing them to this family. All of the species
fall in a single genus, Myrmicia Fabricius.


1. Claws simple------------------------------------------------ 2
Claws pectinate----------------------------------------------- 5
2. Hind tibiae with one apical spur --------------------------------------- 3
Hind tibie irithout an apical spur.-------------- Tribe I. ONYCHOMYRMICINI
3. Promesonotal suture usually present; apex of the abdomen from the third
segments normal, not curved downward, but porrect--------------- 4
Promesonotal suture wanting, or obsolete; apex of the abdomenii froni the
third segmeiit abno'ral, curved downward in such a way that it is
directed forward beneath; antenna ending in a distinct club.
4. Head oblong, the eyes not especially small. played at or a little blonml the
lateral middle; thorax not constricted at the middle.
IHead oblong or oval. the eyes usually small, oval, placed before the lateral
middle sometimes ivnear the base of the mandibles (in a single case very
minute and placed at the middle)----------_-----__Tribe IV. PONERINI
5. Head oblong oval, the eyes inot iiismall, pllac(ed near the lateral middle, thorax
long, slightly .onstricted before the middle, the metanotum lollnger!
the mesonotumn, convex above; mandibles subtriaaigular, curved down-
ward with an oblique but broad masticatory edge.


1. Claws cleft or with a tooth beneath_----------------------------------- 6
Claws simple without a tooth lbeneath- -------------------------------- 2
2. Head usually oblong, the ey.s, when present, never large; thorax usually
with one or more sutures; if without sutures, the head., thorax, and first
two segments of the abdomen are longitudinally grooved--- ___ 3
Head quadrate or subqtuadrate, very little longer than broad, the eyes very
large, placed at the sidles close to the base of the ini;idiblcs; the thorax
oblong, subquadriangular, without sutures above or with only the meta-
notal suture present; scale of abdominal petiole sulquadrate.
3. Head variable, although usually oblong, but not longitudinally grooved, the
eyes rather small oval or oblong oval, placed anteriorly before the lateral
middle, wimetimes near the base of the mandibles; mandibles usually
more or less triangular, the masticatory margin dentate; if slender, the

a Tribe I. Onychomyrmicini: Represeiited by a single genus, Onychomiyrmc.r
Emery, found in India.
Tribe II. Cerapachyini: Well represented in America. The genera are:
Sph iinctom irmes Mayr, Ph yracacc.s Emery, Cerapaci is Smith, Para.syscia Emery,
Syseia Roger, and C!.stias Emery.
Tribe III. Proceratiini: Genera : D-i:cothyrea Roger, Sy.sp.iiucta Roger, Pro-
cerati' m Roger, and Prionopelta .Mayr.
Tribe IV. Ponerini: A group of world-wide distribution. Genera: Centromiyr-
me.r Mayr, Trapeziopclta Mayr, Myopias Roger, Cryptoponi Emiery, Rhopaloponiie
Emery, and Ponera Latreille.
Tribe V. Lep)togenyini: A tribe readily recognized by its imenibers always
having the claws pectinate. Genera: Prioniogeniys Emery, Leptogeny.s Roger,
Lobopelta Mayr, and Simoponwc Forel.


mandibles have a large tooth within neart thle middle; thorax narrowed
posteriorly, with all the sutures, or at least the promesonotal suture,
present; first abdominal segment well separated from the second by a
strong, constriction and a movable joint ---------------------------5
Head oblong, sometimes longitudinally grooved, the eyes, when present, oval
or oblong, placed on the sides near or a little beyond the middle, never
far anteriorly; mandibles most frequently rather long and narrow, with
teeth along the inner margin; if triangular, the masticatory margin is
edentate; thorax rarely much narrowed posteriorly, the sutures some-
timies present, sometimes wholly absent; first abdominal segment not
well separated from the second, usually broadly sessile with it, and
without a movable joint between-------------------------------- 4
4. Head. thorax, and first two segments of the abdomen not longitudinally
grooved, the sutures of the thorax distinct; mandibles rather long and
narrow, acute at apex, and armed with teeth along the inner margin;
eyes sometimes absent; if present, oval, placed at or near the middle of
thile sides of the head------------------------- Tribe II. AMBLYOPONINI
Head, thorax, anid first two segments of abdomen longitudinally grooved, the
sutures of tihe thorax entirely absent; mandlibles obliquely truncate at
apex. wiithiit teeth; eyes oblong. placed on the sides of the head a little
behind the middle; scape of antenna short, not half the length of the
head ---------------------------------_-Tribe III. CYLINDROMYRMICINI
5. Thorax always na rrowed losteriorly, with the pro-, meso-, and metanotal
sutullres usually present, or at least with the pro- and inesonotal sutures
present, scape of antenna long. usually reaching to the apex of the head
or nearly to it: scale of al dominal pietiole usually higher than long; eyes
placed oni the sides of the head. usually between thie middle and the base
of the mandibles, rather close to thIe mandibles.
6. Thorax rather long, narrowed posteriorly inucli as in the Pachycondylini
andi with usually one or more of tlie notal sutures present, the sides
never paraIllel, usually more or less constricted mniedially, but never
strongly constricted. 11iand(ibles never very long. and not curved upward,
usually with a slight curve downward; head oval-quadrate or oblong,
tlie eyes not ],lrge. placed on tie sides, usually a little before the middle,
rarely behind tim middle -----------------------ribe V. ECTATOMMINI
Thor:x very long and cylindrical, the sides pirall,?l, or nearly, the pronotal
sutimure alone distinct, tie ilieso- liigl inetalnotun closely united without
a trace of a dividing suture; ii.naddibles alinornmai, very long, slender, and
curved upward, the inner margin from the basal angle armed with two
rows of minute teeth: head oblong, the eyes large and placed close to
the base of the mandibles----------------Tribe VI. DREPANOGNATHEINI.a

a Tribe I. Lioponerini : Represented by a single genus, Lioponcra Mayr.
Tribe II. Amblyoponini. Genera: .lyopone Roger, Mystrinm Roger, Emeryella
IForel. Stigmatommu Roger, Mesorena Smnith, and Amiblyopone Erichson.
Tribe III. Cylindromyrmicini: A tribe found only in the Americas and rep-
resented at present by two genera, Cylindromnyrmex Mayr, and Thaumatomyr-
1c.r Mayr.
Tribe IV. Pachycondylini: This tribe is well represented in America, Asia,
and Africa, probably a dozen genera being already characterized. Those known
to me are: Odontoponera Mayr, Diacammnia Mayr, Plectroctena Smith, Both-
rioponera Mayr, Belonopelta Mayr, Pseudoponera Emery, Brachyponera Emery.


It is worthy of notice that Doctor Ashmead finds structural char-
acters which require the separation of Leptogenys from Pachycon-
dyla, Neoponera, and other immediately related genera. The fact
that Professor Wheeler found the habits, eggs, and larval characters
of Leptoyenys and Odontomachus so closely similar to those of Pacley-
condyla only increases the significance which may well be attached
to the differences in these respects between Ectatomma and the
Pachycondyla series.


The habits of the insects of the family Poneridw, with which the
kelep would be associated in current zoological classification, were
almost completely unknown until the recent investigations by Pro-
fessor Wheeler. It is evident, however, from his interestir papers,.
that the Texas Poneridae studied by him must belong to series quite
remote from the kelep. The eggs of the kelep, for example, are
elliptical in shape and not cylindrical, as in Odontomach/s and in the
Ponerida, investigated by Professor Wheeler.
The breeding habits and the characteristics of the eggs and larvae of the
Ponerina' exhibit striking deviations. frnim tltose of other ants. I have not seen
the eggs of Onhftomachitus, but throughout the month of May I have often hap-
pened on the eggs of Pacliifonvyla and Lc'pfo!cniy.s. These are white and of a
slender, oblong shape, somewhat smaller in the latter thaln in the former genus.
They differ in shape from the eggs in species of E'ifoi, C(i iouiotus, Forminiva,
Poeon.oni yrm e.r, Soleunop.-is, an(d Tapinoumai; for the ants of these genera, repre-
senting several sulbfamilies, agree in having elliipti'al and In,1li ]e<-s slendier
eggs than the Ponerin.e. The Ponerinab also keep tlieir in more regular
packets, the long axes of the different eg's being placed parallel with one
It would probably be an advantage for the kelep to have eggs like
those of the Odontomaclus, which readily adhere in bundles and can
be the more easily carried, but the diffic.'-ty is overcome by the greater
intelligence of the kelep, which adroitly uses the posterior end of its

Mesoponcra Emnery, P.acliyc-o dyla Smith, Euponera Ford, .Yoponci(ra Emery,
anid Eclfvoini t/ruio.r Ma.yr. The genera P.silomiiic, yrie. Andr(. and Hetcropoinc'r
Mayr, which I have not seen, appear to belong to this tribe.
Tribe V. Eetatommini: This tribe is abundantly represented in Central and
South America and in Africa. Only a few forms occur in North America. and
these in the Southern States. The genera placed here at present are: Plat~f-
thiyrea Mayr, A.foria Emery, l'tictoponcra Mayr, Ectatomnma Smith. Gnamp-
togeniy. Rogers. Acanthi iiopoicra Mayr, Prviomonrnim Smnith, Holcopoimra Mayr,
Rh/ytodoponera Mayr, Stri'blo!l alhu.i fs Mayr, Dinoponcera RP,,ger, Paltothyreus
Mayr. and Magaponera Mayr. .
Tribe VI. Drep)anognathini: This is a peculiar little group known only frtomn
Asia and Africa. and but few species have been described, all falling into a
single genus, Drepmiognfathus Smith.
a Wheeler, W. M., 1901, Biological Bulletin, 2:14.


abdomen to load several eggs between its mandibles. Sometimes also
the eggs of thie kelep adhere somewhat in irregular clusters, so that
several can be lifted and carried at once.
The superior social organization of the kelep is also rendered evi-
dent by important differences in feeding habits.
The l'onerina' do not seem to feed one another, like the specialized ants. In
captivity P. harpa, would eat the yolk of an egg or even sugar, but it would not
e;at termites. L. cloafata devoured termites and small caterpillars with avidity,
but would not eat flies. 0. hla'matodcs is more omnivorous; besides caterpil-
lars, house flies, beetles, and small Hemiptera, it will eat sugar, bread, cake,
The kelep does not appear to have the art, of regurgitating food as do
the true ants, but it is the regular custom of the workers to gather liup
on their mniouth parts large drops of nectar, sirup, or honey, which
are carried into the nest and freely dispensed to the remaining mel-
1ers of the community, as well as to the queen and larvfe. The ulse of
nectar an(l other sweet sul)stances by other Poneride seems not to
have l)een reported. The adult keleps seem to be able to live for an
indefinite period oil sugar alone, though animal food is probabl)ly
necessa;iry for the normal growth of the larvae. Termites and insects
of all the principal groups are eaten readily.
The extent to which the keleps normally depend upon nectar has
not been ade(qiately learned as yet. It may be that they use it largely,
if not exclusively, for feeding the very young larvar, since these do
not seem to ly1 regularly' fed with animal food, captured insects
iIlwavys b)(ing given, as fir as observed, to the large larve. Colonies
fed exclusively on sugar or honey have raised larvr to nearly the full
size, )11t these seldom, if ever, pupate normally,; and in some of the
cal)tiye (olo lies very few l)pupMv have survived to emerge as adults.
Sometimes the cocoons are opened and the l)pupe taken out and eaten;
in other instances they are throws on the bone yard or refuse heap
which each colony maintains. Messrs. Lewton and McLachlan are
inclined to I)believe, after a series of very careful observations, that
this mortality is sometimes due to the attacks of the mites, especially
whlien colonies become weak and discouraged. Prosperous colonies,
on the other hand. may receive no apparent injury from the presence
of large numbers of mites.
The most prosperous of the colonies which have been kept under
lal)oratoryv observation is one at Victoria, Tex., which was used as
the basis of a feeding experiment to see how many boll weevils it
would destroy. From 10 to 30 weevils was the regular ration,
a Wheeler, W. M., 1901, Biological Bulletin, 2:11. Mr. MeLachlan reports that
a colony (if Pachycondyla hliarpax, captured near Victoria, devoured termites

but in one instance over 50 weevils were killed inside of twenty-four
hours, and in the course of the experiment the colony disposed( of
between 1,200 and 1,500 weevils.
The nests of the keleps, though small and simpl)le. ill conmparisoll
with those of the termites and true anilts, consist of chambers several
square inches in extent, with level floors an(d arched roof-!, and denote
a state of architectural advancement much greater than that rel)orted
by Professor Wheeler in the Texan genera studied iby him.
The nests of the three I'onerinae agree in being of a very primitive structure.
They consist of ;a few simple lnd irreguilir burrows, rI' galleries, some of which
run along the surface of the soil imumediiately Ibelle.ath the ,stotne or log. while
others extend down into the soil obliquely or vertically to i dlepthli of S or 10
inches. These burrows ma;y anastomlose, but they are not widelled at cert;aiinl
points to form (chambers, ;as in the nests of the inmore specialized antis Alttu.
Pohlonollmy1rmic.r. (' unipot a1x. etc.). Even illn artificiall nests of thle Lubbock
pattern the I'onerin-r dig only anaist'oniositg galleries sc(:arcely mllore tha;il a centi-
meter in d(liamneter.a


The alliance of the kelep system of organization with that of the
wasps and bees receives support also from the instinct of inclosing
the pupating larva, in what mnay be termle(d temporary brool cells.
It has been reported of some ants and of some general of 1Poieridau
that they bury their larva' before pupation. The keep. instead,
constructs over them a ('cell of earth. if no other materials 'are at hand,
but prefers pieces of old cocoons if these are obtainable."
The bees and wasps avoid the labor of inclosilig etachl rigglilg
larva separately by a system of permanent l)brood cells inl which the
larva are reared., and(l which have merely to be sealed over at jpupa-
tion. Many of thle true aints (andd also the dri'velrs) have abanldonled
the use of cocoons and of the b)roo(l cells in which they can be spliin.
With them the digging of extensive subterranean nests. an(1 tlhe care-
ful handling of the larvv have made external protection inneces-
sary. Nevertheless, the failure to make b)rood cells deprived tihe
a Wheeler. W. M., 19( 1. Biological Bulletin. 2: :
b The honeybee also makes use of silk from old cocoons ill thle contruction (or
brood cells.
"Into the material used in constructing brood combs bees often incorporate
bits of wax and fiber-like gnaw'ngs of cocoons from old combs in which brood
has been reared, and if given ca)ppIings or trimmings of conibs they will work
them all over and utilize most of the material. Also when the bees have abun-
dant suppI)lies of pollen much of this is incorporated into the material of broodl
combs, thus saving the costlier substance-wax."-Benton. Frank, 1899. The
Honey Bee, Bul. 1, n. s., Division of Entomology, U. S. Department of Agri-
culture, p. 27.


ants of the possible evolution of the art of storing liquid food like
the bees.a
The position of the brood cells has doubtless had an important
influence on the evolution possible to the different groups of social
insects. The brood cells of the kelep are built over the larva as it
lies onil the floor of the chamber. Those of the honeybee are also
horizontal, though piled, as it were, in double tiers. The cup-like
cells of the bumblebees stand erect.
The instinct of making honeycomb was evidently attained as a
further step in the practice of constructing permanent brood cells
in advance of the egg laying. Among insects, as with man himself,
progress toward civilization has largely come through the accentua-
tion of parental instincts.
The underground activities of the keleps are much less extensive
than those of the ants. The galleries of their nests are simple and
nearly straight, and the chambers relatively small. There are none
of the winding passages to which the termites and true ants are so
partial, and which make the complete investigation of their domestic
habits so difficult.


The incredulity with which the report of the weevil-stinging ant
was received by entomologists shows how completely incongruous
such a habit appeared when ascribed to an ant. The biological
anomaly largely disappears, however, with the recognition of the
fact that thei natural relationships of the kelep lie with the para-
nitic wasps rather than with the ants. The stinging of other insects
by lpredaIceols wasps is an old and familiar fact. It is true that
most wasps prey upon spiders, caterpillars, flies, cicadas, crickets,
cockroaches, or even uipon bees or ants; but there is one family, the
Philanthida, which regularly attacks beetles. One of these wasps
even makes a specialty of beetles of the weevil family.
The species of the genus ('cerccris are numerous in Europe, and several of
them are known to make burrows in tlhe ground and store them with beetles
fir the benefit of the future larv:p. The beetles chosen differ in family accord-
ing to the species of Ccrccris.: but it appears from the observations of Fabre
.n1d Dufour that one kind of Ccrcris never in its selection goes out of the
limits of a particular family of beetles, but, curiously enough, will take insects
most dissimilar in form and color provided they belong to the proper family.
This choice, so wide in one direction and so limited in another, seems to point
to the existence of some sense, of the nature of which we are unaware, that
determines the selection made by the insect. In the case of our British species
a Though also making special 'honey tubs.' some of the bumblebees commonly
use old brood cells for the storage of honey."-Sharp, David, 1901, Cambridge
Nat. Hist., Insects, 2: 56.


of Cerecris, Smith observed C. arcnuria carrying to its nest Cureulionidl; of
very diverse fornus, while C. labiala used ;i beetle-Iallaica abidla-of tlhe
family Chrysomelidte.
The beetles, after being caught, are stung in the chief articulation of the ioly,
that. namely, between the pro- and nmeso-thorax. ('crccri, biipricstwidia C'ion-
fines itself exclusively to beetles of the filamily Buprestida,. It was by observa-
tions on this insect that Dufour first discovered the fict that the insects stored
up do not decay; he thought, however, that this wns due to thle liquid injected
by the wasp exercising some antiseptic power: ibut tlhe o1lb-er'vatiois of Falre
have shown that the preservation in a fresh state is due to life not being extii-
guished: the stillness, almost as if of deth, Ibeing due to tlie destruction of tlie
functional activity ( f the nerve centers thnt govern tlie m1114ements of tlie limls."
Between the habits of such wasps and those of tlie keleps tlhe dif-
ferences are certainly less titan between tlie keeps and thle true ants,
and another wasp of tlhe same family shows, how tlie tradition from
the parasitic to the prledaceos stage could be passed very graldually.
Instead of concluding its maternal duties iNy stocking its nlest or bur-
row with paralyzed insects in advance, IJiull thf/, u,1pit.i'u.S, which
preys upon the domestic honeybee, provides only one individual at
first and returns later to bring others, as required by thle growing
larvaw. Moreover, it has thle habit of crushling its prey in its nialidi-
bles, thus giving the larva easier access.' The kelep system repl)re-
sents a further improvement, the captured, weevils being pulled to
pieces and distributed to numerous larva,. In one group of Hymen-
optera, the sawflies, the larva? have wvell-developed legs and are able
to crawl about like caterpillars, whichli they very nuch resemlble. Il
all the other families the larva' are legless, helpless grubl)s. Most of
them are, in the larval stage, parasites, either of plants or of animals.
Those which are not parasitic require, obviously, to be fed and cared
for by their mothers, a condition most conducive, obviously, to the
development of social habits.
The distinction between l)redaceolus and parasitic habits is not
easy to draw. Mrs. Cook saw three young kelep larva' at one time
attached to a termite which the workers were still carrying about to
feed to the larger larva', which seem to secure nearly all of thle direct
attention of the workers. If the prey remained alive such attached
larvae would be looked lupl)Onl as parasites.
The manner in which the keleps feed their prey to their larva, is
thus to be looked upon as a derived rather than as a primitive trait,
compared with the habits of the other carnivorous Hymenoptera.
a Sharp), David, 1901, Cambridge Nat. Hist., Insects. 2 :125.
b There are two other significant approximations of habits between the kelep
and these wasps of the family Philanthidw. They seize their Iprey in their
mandibles and sting it by bending the body around underneath, to reach the
vulnerable point. Phiuanthus also makes long burrows in the ground, with
chambers at the end in which the eggs are laid and the young reared.


The parasitic groups avoid all further labor in behalf of their young
by simply laying their eggs inside the victim, upon which the larvae
feed at pleasure. The solitary wasps stock their brood cells in ad-
vance with whole spiders or insects. Some of the social species com-
minute or mangle their prey, and others, including many of the true
art,;. regurgitate the partly digested food material. The drivers
probably feed dismembered pieces of their prey like the keleps. The
kelep larvT are not so completely helpless as those of bees and true
ants. being provided with mouth parts adapted for eating out the
soft interior tissues of insects, and long, flexible necks to enable them
to reach inside and clean out the sections of boll weevils laid by the
workers carefully on the fat stomachs of their baby sisters. Two
such. lying side by side, each provided with a weevil's front leg to
nibble, was the ludicrous sight observed in the nest of one of the
captive colonies in Texas. Mrs. Cook has noted another instance
of feeding which well illustrates the extent to which the social
organization has developed in this respect.
A worker seized a termite as soon ais it was dropped into the nest and held
it in its jaws for fully five minutes, the termite vigorously protesting with its
antenmit. After it was lead. or at least motionless, the kelep took it below
w1wre itlier workers assisted in feeding it to a large larva. It was very
hard to get tlie termiite lioperly j)laced; time and again it fell from where it
liad libeen put, and was turned over and twisted in all sorts of ways in the
effort to briltng it inlto a position so that the larva might take hold of its head.
Tihe larva imeanlwhile Imoved its own Ihead back and forth, evidently trying to
get hold oni its own1 account, and a little larva near by did secure a hold on the
other end of thie termite, so that the kelep)s had to move both larva and termite
inll their furtllther efforts to giv(, time latter to thie large larva. The little larva
was almost as large as the termite. Finally the matter was arranged, the ter-
minte lyilg across tihe two larva', which remained l)ea'efully side by side, the big
one eating at thie lead, the little one at thie tail. A worker had to take the
hliead of thie bIig larva between his jaws and fore legs and put it in contact with
the termite, amid tlien stood over it as though to see that the larva did its duty.
Thie weight of thie small larva kept pulling thie termite off the body of the large
larva before it had become firmly attached, so a worker stayed by and kept pull-
ing thlie termite back in position. Finally time large larva got to work in earnest,
and time faithful nurse left to hlelpl) another kielep with another termite.


A possible reminiscence of an upper air existence by social ances-
tors of the kelep is to be found in a curious tube or gallery
which is often constructed at the entrance of the nest to extend the
underground passage upward on the stem of the cotton plant or
other object to which the opening of the nest is adjacent. The
workers always prefer to dig against something rather than in the
open ground. The tubular structure in question may be only an
inch or less in length, but it sometimes extends upward fori 5 or 6


inches. It is built of shreds of woody material, bark, fibl)er-, o01 even
of cotton lint, sometimes with pellets of earth intermingled, but
usually thin and with something of the papery texture and appear-
ance of a wasp's nest. The object of this curious structure has not
been determined with certainty. In some situations it serves the
purpose of keeping the loose earth from falling into the burrow and
may afford protection against enemies of some kind as yet unknown.
After the tube has been built the insects seem reluctant to crawl over
it to the ground outside. They even carry the pellets of earth
brought out of the burrow up the stem of the plant and far out
on the branches, and then drop them off.
The most obviously wasp-like habit of the kelep is, of course, the
adroit stinging of its prey to produce paralysis and consequent help-
lessness. It has long been known that the mud daubers and the
digger wasps stock their brood cells in advance with paralyzed insects
or spiders, though very few of them are known to attack beetles.
Tlhe insects are permanently paralyzed, but not killed outright, and
afford the young wasps an adequate supply of fresh food on which
to grow to maturity before emerging from the cell in which the egg
was deposited by the mother insect. It has been claimed by some
observers that paralysis results because the insects captured by the
wasps are always stung in a nerve ganglion. This extreme refine-
ment of instinct is doubted by some, but seems to have been estab-
lished by careful observers.
In dealing with the boll weevil there can be no doubt that the
kelep shows great instinctive skill, and often persistence as well, for
the armor of the weevil permits the insertion of the sting at only two
points-on the middle line of the ventral surface, at the two joints
of the body, one between head and thorax, the other between the two
parts of the thorax. The difference in the use of the sting bet,.een
the keleps and the true ants is most effectively shown when repre-
sentatives of the two groups are brought together and permitted to
fight. The ant tries to bite its antagonist, the kelep to sting and
paralyze, the latter strategy being much more effective.
This habit of stinging other insects allies the kelep not only with
the predaceous wasps, but also with the Mutillidpe and other para-
sitic groups, which, instead of preying on other insects or storing
them up for their young, lay their eggs in the living insect direct or
in the nests of the social Hymenoptera. The stings of the worker
bees and ants are, as is well known, merely modified ovipositors.
The males do not have stings, and the tendency to use them is gen-
erally less in the queens, in which the egg-laying function remains


Contrasts between the kelep and the true ants also appear in con-
nection with the sting. The kelep, which has an effective sting for
use on other insects, does not make it a weapon for general defense
against intrusion. The ants, on the other hand, fight each other with
their mandibles, but many of them sting viciously at any foreign
object with which they may come in contact. Curiously enough, too,
many ants in which the sting has become a mere harmless rudiment
still go through the motions of stinging with as much promptness
and apparent gusto as their more effectively armed relatives. The
decline of the sting among some of the ants may be associated per-
haps with the fact that they are vegetable feeders. At least, it would
seen to )e an indication of their remoteness from the parasitic groups
of 1Lymenoptera.

Lest the present recognition of the similarity of the kelep to the
bees and wasps should lead to another misapprehension, it may be
well to repeat lere the fact that the insect is entirely harmless to
man. Its sting is used with instinctive promptness to paralyze
boll weevils and other insects which it undertakes to capture, but
there seems to be a complete lack of any tendency to defend itself
by stinging, except when actually caught and held.
This has been shown to the entire satisfaction of all who have had
the interest to watch the kelep colonies which were brought to the
United States in July (1904). The insects have been handled on many
occasions, and by many differentt persons, without any threat or
symptom of stinging, except in the case of two or three young men
nt Victoria, Tex., who had thle curiosity to make a test of the injury
which thle insect's venom could inflict. The result was quite the same
ns we had exl)erienced in Guatemala, a slight and temporary irritation.
Mlessrs. Goll and Collins, who recently excavated the nests of about
40 colonies in Guatemala, were not stung at, all, though taking no
measures to protect themselves.
That an insect which is so ready and skillful in stinging its prey
should be so l)eaceable and harmless in other respects may well appear
almost incredible, but a similar specialization of instinct is known
to exist in the domestic honeybee, where the queen has no inclina-
tion to use her sting except for the purpose of dispatching her rivals.a
a" But though this sting is always ready to strike, though they make con-
stant use of it in their fights among themselves, they will never draw it against
a queen; nor will a queen ever draw hers on a man, an animal, or an ordinary
bee. She will never unsheath her royal weapon-curved, in scimiter fashion,
instead of being straight, like that of the ordinary bee-save only in the case
of her doing battle with an equal-in other words, with a sister queen."-Maeter-
liuck, M., 1901, The Life of the Bee, 100.



It is plain that the size attained by the colonies and the fecundity
of the individual queens by no means determine the rate of increase
possible for such an insect as the kelep. A social economy which
provides for an indefinite increase of the numbers of colonies and of
laying queens may more than compensate for the absence of the ex-
ceptional productiveness of the queens of the termites and the true
ants. The marvelous fecundity of the females in these groups is,
indeed, to be looked upon as an adaptive specialization )by which
the species is able to maintain an existence in spite of adherence t'o
a most wasteful social policy, in which so much is staked on the
vicissitudes of a simultaneous annual emnergence of all the young
and unprotected individuals of both sexes. This precarious period
in the life history of the ants has been well described by Professor
The sexual individuals, when finally liberated from the nests, are thrown on
their own resources, and for a time tlhe struggle for existence sets in with great
severity. One has an opportunity of actually witnessing both catastrophic and
personal eliminnation, often on a magnificent scale. The struggle among the
males for the possession of the females is intense. Thle lives even of the for-
tunate among the former are rapidly extinguished. The surviving fecundated
queens set to work to establish their colonies, an arduous and complicated
undertaking, which ruthlessly eliminates all tihe poorly equipped. Even before
they can dig their nests hundreds of these insects are devoured by 1 irds, lizardls,
spiders, etc. And many more of them d(lie from exhaustion while iiigginig their
nests, or from hunger while raising tlieir first litter of young, or from the at-
tacks of subterranean predatory in'sects., mparasitic fungi, etc. This struggle,
however, terminates on the appear:llce of tie first workers, and thle successful
queens thenceforth again lapse into a condition of domestication till the close
of their often very long lives.a
The contrast can best be expressed, perhaps, by saying that tlhe
social system followed by the termites and true ants involves the loss
of nearly all of the females, while in that of the keeps all females are,
or may be, saved and utilized in the increase of the species. There is
nothing to show that the very large colonies of the ants and ter-
mnites are an arrangement advantageous to tihe species as a whole.
They are rather the result of the failure of these insects to adopt the
habit of swarming, as practiced by the honeybee and the kelep.
From the agricultural standpoint, too. the superiority of the kelep
organization is obvious. The division of the species into small bands
enables the insects to spread uniformly over the fields, while large
isolated ant-hills inhabited by hostile colonies would exclude the pos-
sibility of an efficient protection of all the cotton. The hills and pit-
a Wheeler, W. M., 1902, Science, n. s., 15: 769.


falls made by many ants would be a serious inconvenience in the cul-
tivation of the crop. To avoid them would be difficult, and to drive
over or plow through them would mean the wholesale destruction
of the insects, so that the utilization in agriculture of a true ant-hill
ant might well be deemed improbable. The social organization of
the kelep avoids all of these difficulties. The colonies are small,
but not hostile, so that all the plants of the field can be visited. In-
stead of an ant-hill and a maze of underground passages, they have
a simple burrow, deep enough to give them protection against injuries
incidental to ordinary cultivation. If the entrance is accidentally
closed the inhabitants can readily reopen it; and if the situation
proves to be inconvenient the whole community generally has the
intelligence to move over close to the cotton plants, on which it regu-
larly forages for nectar and game.

In summarizing the former report, of July, 1904, the investigation
of the kelep was divided into five phases or questions, upon three of
which evidence was submitted, while two others remained uncon-
sidlered. The three facts which seeined to be established were:
(1) The kelep protects the cotton plant against the boll weevil,
which it regularly kills and eats. In the presence of sufficient num-
bers of keleps the protection is entirely adequate, as shown by com-
parative field tests in Guatemala. The keleps have made it possible
for the Indians to maintain a field culture of cotton in the presence
of the weevil, and have thus enabled the Indian variety of cotton to
develop weevil-resisting characters which give partial protection,
even in places where the keleps are few or wanting.
(2) The kelep does not attack plants, or have any other habits
which would make its introduction into the United States injurious
or undesirable.
(3) It is feasible to bring colonies of the insects to Texas and estab-
lish then in the cotton fields.
The continued study of the insect has furnished, of course, many
additional data bearing upon the above statements, but all have been
of a confirmatory nature. The present paper deals with the further
question, whether the habits of the insect will enable it to breed and
multiply in captivity or as a domesticated inhabitant of the cotton
fields of our Southern States. From the analogy of the habits of the
ants, it appeared to specialists in the study of that group of insects
that the kelep could not be applied to agricultural uses. The facts
detailed in the present paper seem to warrant a different answer.
(4) The kelep is not a true ant, but has a distinct type of social
organization, much more like that of the domestic honeybee.


The utility of the kelep will depend, therefore, upon the answer of
the remaining question, whether the kelep can survive and multiply
under conditions of climate, soil, and food supply to be found in the
United States. In structure, habits, and instincts it is wonderfully
a(lal)ted to the work of destroying the cotton boll weevil.
It is, in short, a new and efficient insectivorous animal, in all prob-
ability capable of being utilized for the protection of cotton and other
crops in many tropical and subtropical regions, whatever may be
the results of the present effort to naturalize it in Texas.
It is still impossible to predict the fate, in a new country, of an
insect which has so recently become an object of scientific study, but
whatever the experiments may ultimately show regarding its ability
to become established and thrive in the United States, it seems certain
that the social organization of the species does not disqualify it for a
future of agricultural utility. To prejudge its prospects by refer-
ence to the habits of the true ants would be the same kind of error as
to discredit the honeybee on the ground that the bumblebees and
wasps are worthless and undesirable insects.
Recent advices from Texas, received just as this paper is being com-
pleted, seem to indicate that the colonies which have been left out in
the open ground of the cotton fields, without care or food, will not
survive the winter, though they have lived long enough to show that
low teml)erature alone is not fatal, thus confirming the result of a
cold-storage experiment made in Washington last August (1904). As
already explained in the present paper, the colonies of the first
importation were far too small to make the experiment a fair ohe,
and they were planted in Texas after the middle of July (1904),
much too late in the season. A knowledge of the social organization
of the insect will permit these and other obstacles to be avoided in the
importations to be made in the coming spring (1905), and it is hoped
that a satisfactory and conclusive test will result.



Abdomen, flexibility of------------ 21
Agricultural uses of kelep--------- 49
A M BLYPON IDE -------------------- 38
AMBLYPONINE ------------------_ 30, 37
Ancrgates, males wingless--------- 19
Annual emergence-------------- 8,32,49
Ant society---------------------- 27
Ants, breeding habits-------------- 36
cocoons abandoned----------- 43
colonies compared with kelep_ 27,50
colony, sequel of sexual repro-
duction _------------------ 35
domesticators of other insects- 9
indeterminate organization -___- 33
mating concourse_------------ 8
not adapted to agricultural re-
quirements --------------- 49
number of individuals-------- 27
polymorphism of------------ 26
similarity to kelep superficial- 35
social system wasteful ------- 32
unable to destroy boll weevils- 21
wings, removal of----------- 27
workers hostile------------- 16,27
ASHIMEAD, W. II., references to.. ------ 29,30
classification of the
PoN ER I D.-E --- 38
AVEBURY, reference to------------ 10
Bamboos, periodical flowering of--- 8
BATES, 1I. W., reference to-------- 31
Bee, fecundation of queen --------- 24
sting of worker --------------- 26
worker caste, cause of ------- 24
Bees, dependent on presence of
queen ------------------ 20
honey, brood cells--------- 24
purpose of honey production- 19
resemblance to nonsocial in-
sects------------------- 29
royal etiquette among------ 23
similarity of keleps to--- 29,43,50
solitary --------------------- 36
young queens of----------- 34,35
BENTON, FRANK, on brood cells of
bees -------------------------- 43
BETHYLIDE_ ---------------------- 30
Biological synchronism----------- 8
Boll weevil, vulnerable points------ 47
Boll weevils, feigning death ------- 21
killed by keleps.------ 42
Breeding habits, diversity of------ 23
of kelep--------- 10
Brood cells, horizontal------------ 44
of kelep------------ 43
Bumblebee, foundation of colony-__ 28
interbreeding of------- 28

Bumblebee society --------------- 28
workers winged__ -------- 28
Bumblebees, governed by queen ---- 34
honey tubs of -------- 44
Calathwa------------------------ 11
Calotermes ---------------------- 14
Cannibalism --------------------__ 14, 15
Cardliocondytla. males wingless ..- 19
Castes, specialization of---------- -26
Cerapa]iys ---------------------- 31
Cvrccris. a beetle-eating wasp------ 44
Cicadas, periodical--------------- 8
Classification, two methods of----- 37
Clustered colonies------------- --
Cocoons, opening of-------------- 42
transportation of -------- 21
Colletes, social organization of----- 23
Colonies, determinate and indeter-
minate-------------- 33
formation of new-------- 8,
21, 28, 29, 32
increase of ------------- 11
migration of--------- 11, 15, 20
not hostile-------------- 14
population of----------- 11,12
small, inactivity of ------ 12, 42
without queens---------- 22
Colony, compound indeterminate ori-
gin of------------------ 33
new, foundation of 5------ 3
new, from geographical
standpoint ------------- 34
Communities, large, not practicable- 35
Community, expansion of-------- 15
Concourse, use of term ----------- 32
COOK, MRS., observations of------- 45,46
Cotton field, occupancy of-------- 29
Cremastogastcr lwa'riusula --------- 17
lineolata eaten by
SIcna im -------- 36
Cross-fertilization_ ------ 8, 10, 12, 32
necessity of------ 35
CIYPTOCERhIDE --- ------------- 9
Determinants, theory of----------- 26
Dinoponcra grandis -------------- 31
Division of labor----------------- 14
DORYLIDE ---- --- ----- ---------- 29, 30
Drivers ------------------- -9.30, 31, 46
Ectatomma and Pachliyco ndyla. di ffer-
ences between_ ------- 41
as type of family ------ 37
larva, description of.---- 38
tubcreula tin ----- ---- 5
ECTATOMMID.E ------------------- 38
Eggs, adherence of --------------- 21
carried by workers---------- 18,21


Eggs, description of--------------
distinct from those of Odon-
tomnachius ------ ---
in bundles-------------
laid by workers --- --
of an unfertilized queen ..
EMERY, E.. references to----------
Entrance to nest_----____
Erga tarchy -- -- --
Evolutionary history as shown by
social organization _
significance of struc-
tural characters ___
Evolution of honeycombs----------
of social habit----------
of storing liquid food----
Experiments planned--------
Fecundation inside nest-----------
Feeding experiments--------------
grounds, object of separate ion
habits -----------
Females and workers, distinction
bet ween _-------
winged at first----------
Fibers, use of, in construction-__.
FIELDE, A. M., references to-------
Food, animal, necessary------
liquid, storing of--------
Formicoxcinus, males wingless------
Fungus gardens _- _- ___
Galleries, construction of----------
Geographical distribution of social
insects -- ----------
Germ plasm __- _- __ __
GoLL. G. P., observations by.-----
Guests in kelep nests----------
Habits, not injurious -- ----------
Haltica tttbida, stung by wasp .....
IHarmlessness of kelep------------
Heredity, theory of---------------
HINDS, W. E., experiments of------
Honeybee - - - - -
brood cells of----------
colony indeterminate .....
less completely social -.-.
Honeybees, ergatarchy of ----
social development of --
sting of queen---------
(Itqueen, jealousy of----
Honeycomb, instinct of making ----
Honey tubs of bumblebees---------
HYIMENOPTERA, classification ------
primitive social con-
dition of------
Indians. Guatemalan-------------
Insects, classification of-----------
social organization of------
Instinct of place-----------------
Interbreeding ------------------ 8,
a necessity-----------
Isomaralia coronata-------------
Kelep, different from ants---------
Indian name--------------
method of killing weevils---
relationship with parasitic
and predaceous wasps ----




17, 42


12, 10


Kelep society, formal description of- 28
strength of----_------------- 21
KINSLER, J. H., observations by .--- 16
Larva, feeding habits ------------ 46
food of------------------ 42
method of feeding not prim-
itive trait---__----------- 45
of unusual size ------------ 22
Leaf-cutting ants, size of colony--- 11
LEPIDOPTERA, social habits of------ 23
Leptogenys and Pachycondyla, dif-
ferences between--- 41
eggs of -------------- 41
elongata, feeding habits- 42
lack of resemblance, to
kelep -------------- 36
larva, description of--- 37
queens--------------- 19
LEWTON, F. L., observations by--.... 17, 42
I.iberia. drivers of -------------- 29
termites of-------------- 33
MAETERLINCK, A. M., matriarchy of
bumblebees 34
references to 12,23
sting of bee-_ 48
Males, dimorphism of-----------__ 25
from unfertilized eggs------ 35
number of---------------- 16
present throughout the year- 16, 28
visits of ------------------ 28
winged ------------------- 16, 28
Mandibles, adaptation of--------_-- 21
Mating concourse----------------- 8
flight of ants and termites-_ 35
of honeybees-------- 35
Inside nest --------------- 18,29
Matriarchy -------------------- 34
MCLACHLAN, A., observations by--- 22,42
Migration, for short distance only- 29
of colonies------------ 20
Mites, parasitic------------------ 15,42
Mobility of kelep organization----- 35
Mouth parts -------------------- 46
Movements, slow----------------- 30
Multiple nests------------------- 33
MI'TILLID.E ------------------- 29,30.47
MYRMICIDE, derived from Ponerida?- 36
feeding of larvae ------ 36
Myrmosa ----------------------- 30
Nectar, as food of kelep---------- 42
Neoponera ---------------------- 18
and Leptogenys, differ-
ences between--------- 41
Nests, construction of------------ 42,43
of Philanthus, similar to kelep 45
ODONTOMACHIDE ------_---_------ 38
Odontomachus, as type of family--- 37
clarus ------------- 13
disposition of------- 31
eggs of------------ 21
hwmatodes, feeding
habits----------- 42
larva, description of- 38
Pachycondyla-------------------- 13,18
and Ectatomma, dif-
ferences between -- 41


Pachycondyla and Leptogenys, dif-
ferences between __ 41
as type of family.---- 37
eggs of------------- 41
lack of resemblance
to kelep---------- 36
larve, description of- 37
PACHYCONDYLID.E ---------------- 38
Parasites of kelep---------------- 15
P'HILANTHID.E, attack beetles ----- 44
habits similar to
kelep------------ 45
Plowing over colonies------------- 50
Political system------------------ 35
Ponera ------------------------ 32
as type of family---------- 37
hiarpa.r, feeding habits------ 42
lack of resemblance to
kelep------------------- 3-0
larva, description of-------- 37
PONERID.E----------------------- 19,38
classification of--------- 38
families of------------- 37
habits of -------------- 41
Population of colonies------------ 12
Predaceous and parasitic, distinc-
tion between-------- 45
habits --------------- 30
Primitive characters-------------- 31
Pyosopis, social organization of ---- 23
Protection of cotton-------------- 50,51
PSOCID.-E, social habits of-------- 23
Pupation------------------------ 42
Queen -------------- 17, 18. 20, 22, 23, 25
carried by workers ---------- 17, 2o
complemental -------------- 33
development of------------ 25
fecundity of, in termites,
honeybees, and ants------- 35
inactive ------------------- 17
removal of----------------- 17
replacing of---------------- 22
time required for develop-
ment of----------------- 22, 29
Rate of increase of kelep---------- 12,49
Regurgitation of food--------- 17,42,46
Sex, forms of-------------------- 27
SHARP, DAVID, on stinging of in-
sects----------- 45
on storage of honey- 44
Slave-making instincts of ants----- 10
Socialization of kelep. complete---- 34
Social organization of insects------
origin of------- 33
wasps, habits of----------- 28
Stenainimma fulrviumn. feeding habits of- 27,36
Stigmatomma, as type of family ___ 37
larva, description of 38
resemblance to kelep- 36
Sting, effects of ------------------ 48
loss of, among ants--------- 48
modified ovipositor--------- 47
use of, distin.-t from ants--- 47
Stinging habits------------------- 47,48

Stinging insects ----------
of insects for preservation-
Summary --------- --
Swarm, ambiguous use of term----
Swarming ------------------ 7,8,
of ants and termites-...
.S'ym n yrm ica, males wingless------
Synthetic method of classification-
Tf riiCes ltcifiiiis ----- ---
Termite queen, habits of----------
society ------------
Termites----------------- 8, 9, 20,
colony, new, sequel of
sexu,.1 reproduction----
colony, vital efficiency of-
complemental queens ----
larve not helpless----
males regiiTlarly found in
nest -------- -----
of Liberia ---------
similar to true ants------
social system wasteful ----
workers, polymorphismn of -
THYNNID.E ----- -------
Time specialization -- ---
T7ibnfignithiux, male winged___
rTube to nest, construction of----
V'tility of kelep ----------
Wasps and keleps, similarity be-
tween .
transition be-
tween -----
governed by queen----
parasitic, relationship with
keleli -----------------
relationship with drivers
and keleps -------
solitary --- -------
method of feeding
larvae -------
Weevils, number destroyed -------
Weevil-stinging wasps ------
WVEISMANvN, A., reference--------
WHEELER, W. M., references to ---

45, 47
10, 32
8, 35







2 13,
25. 37

on habits of PO-
NERID.E ------- 41
on habits of Stig-
m atomma ----- 36
on annual emerg-
ence of ants __ 49
on construction of
Ponerine nests- 43
White ants. See Termites.
Wings ---------------------------18, 19
Worker caste, character of 20, 25, 28
Workers as nurses --------------- 35
diminutive -------------- 13
diversity among---------- 12
number of-------------- 28
one type of-------------- 12
removal of-------------- 31
time of development------ 23
wingless----------------- 28


.. . :E : .:" ": '"

. .. ",......

... .. ... k... .!

11111111111111!11111 11111 I 111111 I I
3 1262 09229 6150