Myrmecocystus

Honeypot ants found in arid areas of the western United States and Mexico. The habitat of these ants for storing sweet liquids in replete workers was notable, and well known, to indigenous inhabitants living within the range of this genus. H. C. McCook subsequently published "The Honey Ants of the Garden of the Gods and the Occident Ants of the American Plains" in 1882, which brought greater attention to these ants and their habits.

Identification
Myrmecocystus can be distinguished from all Nearctic genera of Formicinae by the elongated maxillary palpi, of which the fourth segment is as long as, or longer than, the combined lengths of the two following segments.

Myrmecocystus subgenus/species-groups.

Myrmecocystus species keys.

Biology
Snelling (1976) wrote an extensive and well informed account of the biology of the species in this genus as part of his masterful revision of the group. He begins with the following paragraph, then devotes some attention to the taxonomic history of the group and its constituent species. He then proceeds to discuss their biology, much of which is provided below.

Writing at the time of the Spanish conquest of Mexico, Bernardino de Sahagun described some of the flora and fauna of Mexico. He mentioned the honey ants and stated that the Aztecs called them "Nequazcatl" (Curran, 1937). According to T. Smith (1807) children in Michoacan, Mexico, were very fond of the repletes and the Mexicans supposed that they contained " ... a kind of honey collected by the insect ... " Simmonds (1885) observed that repletes, fastened to squares of paper, were sold in the market places of Mexico City. He also noted that the "honey" was fermented into an alcoholic drink.

The production of repletes by some members of a colony is an outstanding feature of these ants, one which attracted literary mention as early as Llave's observations in 1832. Llave's account was followed by that of Wesmael (1838). Several other mentions of these ants appeared during the period between 1840 and 1880 including the translation of Llave by Lucas (1860), but these add little of use and much of confusion and are not worth mentioning. The earliest commentary, that of Llave, is fairly straightforward, being nothing more, in essence, than observation of the fact that peasants in the area of Mexico City excavated nests in order to gather repletes. Error crept into the commentary made by Wesmael, although it must be admitted that the fault was not his. Rather, he was the victim of misinformation. His informant assured him that the ants manufacture the honey and that this is, in turn, stored in special cells similar to those of honeybees.

The first detailed, and reasonably objective investigations of a honey ant are those of H. C. McCook, as published in his oft-quoted "The Honey Ants of the Garden of the Gods and the Occident Ants of the American Plains" (1882). Since this account is so well known I think it sufficient to point out the barest essentials. The ant which he observed was Myrmecocystus mexicanus (reported as melliger var. hortus-deorum) in the Garden of the Gods, near Manitou, Colorado. He found the foraging activity to be nocturnal, and that foraging workers gathered the sweet exudates from cynipid galls on shin oaks, Quercus undulata. This honeydew is regurgitated to the developing repletes upon return to the nests. McCook asserted that this sugary exudate seemed to furnish the primary food source for these ants.

The next observer of importance was Wheeler (1908) who presented an excellent resume of all previous accounts. The reader interested in these is referred to Wheeler's paper. Following this resume Wheeler reported on the results of his own observations on many colonies in the southwestern United States. In general these observations confirmed much of what McCook had discovered, but Wheeler was able to add some embellishments of his own. He, too, made observations at the Garden of the Gods, but found that site "... so overrun by tourists that careful and continuous observation of ant-nests is out of the question!" I can affirm that sixty years later the situation was not improved. Elsewhere in Colorado, however, he found conditions more suitable and repeated much of McCook's work on mexicanus. He discovered, as McCook did not, that the ants obtain much of their nectar from coccids and aphids, even in areas where the parasitized shin oaks are abundant. He also found that repletes develop from callow workers.

Finally, with respect to mexicanus, Wheeler presented some thoughts on nest construction which he felt were ". . . explainable only as adaptations to the development of repletes." The great hardness of the soil was commented upon as was the propensity of this ant to construct its nests on the summits of ridges. The size of the nest openings, too, seemed related to replete production. Hard soil was speculated to be best suited to the construction of firm ceilings for replete chambers, ridge tops provided for rapid shedding of rain, and large tunnels for complete aeration of the nests to inhibit the development of fungi. These speculations conclude with the supposition that there is " ... obviously a reciprocal relation between the replete habit and an arid environment; the ants store honey because they are living in an arid region where moisture and food are precious and the storing of honey in replete workers, in tum, is possible only in very dry soil."

Wheeler also published observations made on many other forms. Those made on other species which he found to produce repletes did not differ greatly from those made on mexicanus. He did, however, become aware that certain forms never seemed to develop repletes. This led to the speculation that such is a physiological phenomenon, characteristic of these forms. These physiological forms were, in his opinion, incapable of developing repletes and derived the bulk of their sustenance from predatory activities against other arthropods. This speculation had profound and unfortunate results when applied to the systematics of the genus. In 1908 Wheeler described melliger mimicus, an ant in which he noted the lack of repletes even though he excavated many nests. However, a few years later he received samples from San Diego County, California, of Myrmecocystus mimicus which included repletes. Rather than abandon or re-evaluate his theory Wheeler (1912) described these ants as a new subspecies (lomaensis) closely allied to mimicus. It is true that he made note of some differences to separate the California specimens, but it is equally true, as Creighton (1950) has shown, that these differences will not hold up. Indeed, they do not even exist, for it is quite possible to find specimens in the type series of melliger mimicus which differ from melliger mimicus in precisely the same manner as does melliger lomaensis.

Creighton (1950) ventured the opinion that replete production was a matter of opportunity, rather than an innate, physiological phenomenon. He further offered the idea that replete production is not a response to arid desert conditions but rather that it is the result of a previous xerophile having invaded less rigorous climes. In other words, replete production is the result of the availability of an excessive amount of nectar available in mesophytic environments.

Over a period of several years I have had opportunities to conduct observations on most of our species of honey ants. These are far from complete but they do permit some generalizations. The first of these is that replete production is largely a matter of opportunity and has absolutely no bearing on the systematics of the group.

The first non-replete forming subspecies to fall was melliger orbiceps Wheeler. Parks (1929) noted the presence of repletes in nests of this form in the vicinity of San Antonio, Texas, an observation subsequently overlooked in the years which followed. Creighton and Crandall (1954) published the results of excavations by the latter of a colony near Tucson, Arizona. Within this colony were found large workers with orbiculate heads (the structural peculiarity which identified this subspecies) and large numbers of repletes (the physiological peculiarity which identified the nominate subspecies). Obviously, the only sensible recourse was to synonymize the subspecies melliger orbiceps.

Another highly insectivorous form which did not, in Wheeler's opinion, develop repletes was Myrmecocystus kennedyi (= melliger semirufa, sensu Wheeler). This species is common in exceptionally arid regions of the western deserts, and often constructs its nests in very fine sand. Cole (1938b) reported finding a single replete in a colony 10 mi S of Cameron, Arizona. In 1967 I found many repletes in a colony excavated in Malheur Co., Oregon. Since then they have been recovered from colonies in other areas.

Wheeler, in describing melliger comatus (1908) remarked that he was unable to find repletes in the colonies which he studied in the Davis Mountains of Texas. This form is an outright synonym of melliger, the lectotype of which is a replete. Furthermore, I managed to find a few repletes in a colony which I excavated in the Davis Mountains on 20 August 1967. Another of Wheeler's highly insectivorous forms was melliger mendax Wheeler. Repletes of this ant, which I consider a distinct species, are included in a sample collected by R. H. Crandall in the Santa Rita Mountains, Arizona, and I have taken them from colonies in the Huachuca Mountains, Arizona.

The above cases have one thing in common: all are forms which Wheeler believed to be strictly insectivorous and therefore physiologically unable to produce repletes. These cases are not proof, of course, that all species form repletes, but they do demonstrate that several forms for which this was claimed as a character do, under suitable conditions, develop them. The production of repletes can be demonstrated for most of our species, and I see no reason why we should doubt that all are capable of doing so. One more interesting case might be cited at this point: repletes were found in a nest of Myrmecocystus flaviceps Wheeler, excavated near Shorty's Well, Death Valley National Monument -100 feet elevation. Death Valley is, of course, considered to be one of the more arid, inhospitable areas of our deserts.

Haskins (1939) addressed himself to the question of why certain individuals within the colony become repletes while others do not. He postulated that these individuals "quite independently take to storing nourishment," that such action is a "mental differentiation" or mental mutation. While this is an intriguing speculation, I do not consider that there exists any evidence to support it.

Repletes develop from callow individuals, usually the largest available. That small individuals within the nest may become repletes was shown by Parks (1929), so size is obviously not the sole factor. My own belief is that all callows will accept any quantity of food offered them and that the actual number of repletes developed, and, therefore, which individuals become repletes, is determined solely by the amount of food and the number of callow workers available. Observations made on developing colonies of Myrmecocystus creightoni and Myrmecocystus testaceus are pertinent. When food supply was limited to a small excess, only a few largest individuals were developed as repletes. When, however, the excess was large enough, all available callows, irrespective of size, were filled to repletion. Because of these observations, I cannot accept the notion that some ants become repletes because of "mental mutation" while other callows are endowed with normal instincts which do not allow such development.

Misconceptions exist on the nature of the fluids which repletes store. The sweet fluid commonly stored is not true honey. It consists rather of simple sugars, unmodified from their original state, i.e., nectar from flowering plants, exudates from galls and the secretions of aphids and other homopterans.

Nectar and honeydew are not the only fluids which these ants store. The first to point this out to me was W. S. Creighton (personal communication) who commented on water-filled repletes of mexican us, for which he coined the term "aquapleres." These specimens were secured by M. A. Cazier in Arizona. I have since recovered these "aquapletes" from colonies of this species in Arizona, and from testaceus Emery in California. A third type of replete has been found, also in nests of mexicanus. While I have not identified the fluid, it is my guess that it consists, in large part, of the body fluids of insect prey. Following this discovery, I found similar appearing repletes in preserved samples of testaceus and Myrmecocystus navajo Wheeler. Since these are now preserved, I doubt that the usual procedure (i.e., tasting) would be of much avail in attempting to determine their nature. They are, however, similar to the repletes of mexicanus in that the stored fluid is opaque, whitish to gray, with a considerable amount of sedimentary material. Obviously, too, a detailed chemical analysis of the fluid should be made to determine its composition. If the stored fluid proves to consist largely of proteins, i.e., insect fluids, then it is important to note that such repletes are thus far known only within a single complex of species, the nominate subgenus.

While Wheeler was incorrect in his assumption that some species are almost exclusively insectivorous and others derive their sustenance entirely from nectar, it is correct to state that some apparently are less strongly insectivorous than others. The Myrmecocystus, s. str., species are a case in point. If one observes returning foragers of Myrmecocystus placodops, a member of the subgenus Endiodioctes, the quantity of insects and insect fragments carried into the nest is impressive. Excavation of the nests of this species reveals chambers packed with fragments. Not so in the case of mexicanus or others of this group. Most returning workers seem to be carrying nothing, although a somewhat swollen gaster belies this. On several occasions I have observed mexicanus and testaceus workers feeding on a freshly dead insect. However, the insect was not carried back to the nest; its severely masticated corpse remained where the ant dropped it when through feeding.

It seems timely to diverge to the subject of larval feeding. The larvae of Myrmecocystus have a very characteristic appearance for formicines. Instead of being rather short and uniformly stout for their entire length (as is usually the case of formicine larvae in this region), Myrmecocystus larvae have small heads and the body becomes progressively broadened caudad, so that the larva appears to possess a long, slender neck. Normally, formicine larvae are fed directly by the adult workers (trophallaxis), but this does not appear to be true of Myrmecocystus. In excavating colonies of the supposedly wholly insectivorous species such as kennedyi I have been impressed by the large number of insect fragments present in the chambers containing larvae. Observations on captive colonies confirmed what seemed a logical assumption. The insect fragments are placed among the larvae and these must fend for themselves. I have seen no indication that larvae of these species are fed by trophallaxis. Colonies of members of the subgenera Myrmecocystus, s. str., and Eremnocystus differ significantly from the above. Insect particles are seldom found in the larval chambers and there is some evidence, though by no means proof, that larvae may be fed by trophallaxis, at least in part. If, as I suppose, the bulk of the insect protein utilized by members of these groups is carried to the nests in liquid form and stored within the bodies of repletes, then certainly trophallaxis must take place. More observations and experiments on captive colonies may provide answers.

Some further observations on captive colonies, even though fragmentary, are pertinent to the point at hand. I wish to stress the incomplete nature of these observations which have been to date on but three species: testaceus, creightoni and kennedyi. Observation colonies of all of these had been maintained on three different diets. The first consisted of a strictly honey-water diet (A), the second of insect fragments and water (B), the third of a combination of the two (C). In all cases, colonies were supplied with all the food material they would accept, and some callow workers in all instances developed the replete conditions. Some repletes of both testaceus and creightoni fed on diet B were filled with an opaque, whitish fluid, but none were seen in the single colony of kennedyi maintained on diet B. The colonies fed on diet A did not thrive. Larvae developed very slowly and often died prior to pupation. Most of those which pupated failed to mature. A few pupae did mature and these appeared to be normal in all respects. It is obvious, in these cases, that the colonies were not thriving. The larvae in the colonies fed on diets B and C developed rapidly and normally. The colonies previously restricted to diet A were ultimately converted to diet C. Of these, one colony of testaceus immediately began to improve, but all the other colonies continued to decline and finally expired. These data, albeit fragmentary, seem to indicate that protein, such as that derived from the tissues of other insects, is essential for larval development. It should be pointed out that the adults of the colonies fed on diet A did not seem adversely affected by their restricted diet.

Earlier I noted the limited quantity of insect fragments carried back to the nest by species of Myrmecocystus, s. str. This appears to be true of those of Eremnocystus as well. Further the nests of these species conspicuously lack chambers filled with insect fragments. This, of course, may also indicate that useless bits of integument, not suitable for food purposes, are removed from the nest to the surface of the ground. This latter seems more reasonable to me, but should be verified by further observation.

The nectar sources of these ants are varied. The observations by McCook (1882) on mexicanus in Colorado indicated that this species utilizes the exudates of galls on Quercus undulata, while Wheeler (1908) recorded this species soliciting aphids and "coccids" in the same state. No doubt the latter belong to the so-called mealy-bugs, now placed in the family Pseudococcidae. Recently, McKenzie (1968) indicated several species of mealy-bugs with which were associated species of Myrmecocystus. I have observed most of our species, at one time or another, soliciting aphids for honeydew. Wheeler (1908) expressed doubt that pseudococcids (and aphids) could be considered an important adjunctive food source for the Myrmecocystus of arid regions. I am inclined to disagree, although I cannot offer proof to the contrary. At the time Wheeler wrote pseudococcids were a poorly known group with a very sparse representation in desert areas. As Mckenzie's study of the California pseudococcids has shown, however, the family is surprisingly well represented in arid regions. Furthermore, many of the genera found in the deserts are those with which ants have been found. And, I doubt that the apparent sporadic occurrence of mealy-bugs in desert areas adequately reflects the true situation, but that it does reflect on the inadequacy of the collecting techniques. Mealy-bugs simply are not the sort of thing that casual collectors turn up. Many species feed on roots while others are found in tightly compacted basal clusters of blades of grasses and such plants as Yucca, or on the underside of cactus pads.

Shields (1973) reported finding kennedyi tending larvae of Philotes rita pallescens (Lepidoptera: Lycaenidae) on flowers of Erigonum kearneyi var. kearneyi in Nye Co., Nevada. He also noted that the ants "... fed in the flowers." At this location, Philotes larvae were also tended by Conomyrma insana (Buckley) [reported as Dorymyrmex pyramicus (Roger)..and now known as Dorymyrmex insanus!] and Camponotus nearcticus Emery. Whether all three ant species competed on the same plants was not indicated, nor did Shields mention ants transporting the lycaenid larvae back to their nests.

Another source of nectar, and one which I believe to be very important, was completely overlooked by Wheeler. During the early spring months much of our desert areas are covered with an incredible profusion of perennial and annual flowering plants. Other areas exhibit similar profusions of blossoms during the summer and autumnal months. Myrmecocystus workers may often be found in large numbers exploiting this abundant, though short-lived, food source. In many areas, especially on mountain slopes and in small canyons, there are, in addition, some species of perennials (such as Eriogonum and Euphorbia spp.) which persistently maintain blossoms through much of the year. I believe it safe to say that there is an abundance of plant-source food, either directly or indirectly (mealy-bugs, aphids, etc.), in even the more arid parts of the range of Myrmecocystus.

My own field work has indicated still another factor involved in the food problems which these ants face. During the driest parts of the season, rather few larvae are present in the colonies. However, shortly after plant growth begins a revival and flowering period, larvae become more plentiful in the nests. Growing plants naturally result in an upsurge in insect populations directly dependent upon plants as a food source. Hence, more insects are available as prey to be fed to growing larvae. The maintenance of small larval populations during the dry periods is advantageous since it imposes a small drain on the food procurement capabilities of the foragers at a period when insects, as a food source, are scarce. But, it also provides potential callows to function as repletes shortly after the plant growth cycle enters the active phase.

Shortly following the initiation of this phase of desert existence, larvae are to be found abundantly in Myrmecocystus nests and the number of larvae remains high until the cycle begins to decline. This is a normal and necessary adjustment to accommodate the rigors the arid environment imposes on all organisms. And, honey ants are equaled only by the harvester ants of the genus Pogonomyrmex in their successful adaptation to survival in this harsh habitat. It is interesting that the North American component of Pogonomyrmex, in number of species, is about the same as Myrmecocystus: twenty-two species for the former, twenty-seven for the latter. Both groups are basically inhabitants of arid and semi-arid habitats and in the arid regions are about equally common. Each group has proven successful in desert regions largely through the ability to utilize and store large amounts of food. In Myrmecocystus, however, this stored food seems to be largely for consumption by the adults and colony reproduction is curtailed during periods of limited prey availability.

Nomenclature

 *  MYRMECOCYSTUS [Formicinae: Lasiini]
 * Myrmecocystus Wesmael, 1838: 770. Type-species: Myrmecocystus mexicanus, by monotypy.
 * Myrmecocystus junior synonym of Cataglyphis: Roger, 1863b: 12; Mayr, 1863: 431.
 * Myrmecocystus revived status as genus: Forel, 1878: 372.
 * Myrmecocystus senior synonym of Endiodioctes, Eremnocystus: Snelling, R.R. 1981: 403.
 * ENDIODIOCTES [junior synonym of Myrmecocystus]
 * Endiodioctes Snelling, R.R. 1976: 25 [as subgenus of Myrmecocystus]. Type-species: Myrmecocystus melliger, by original designation.
 * Endiodioctes junior synonym of Myrmecocystus: Snelling, R.R. 1981: 403.
 * EREMNOCYSTUS [junior synonym of Myrmecocystus]
 * Eremnocystus Snelling, R.R. 1976: 92 [as subgenus of Myrmecocystus]. Type-species: Myrmecocystus creightoni, by original designation.
 * Eremnocystus junior synonym of Myrmecocystus: Snelling, R.R. 1981: 403.