He also notes that certain common species which, in the nymphs at least, appear to exude the characteristic anal fluid when disturbed, nevertheless are no.t attractive to ants. He found the following species of Membracidae attended by ants in the vicinity of Ithaca, New York: Theliabimaculata (Fabricius), Telamonaampelopsidis (Harris), T. unicolor Fitch, Cyrtolobus vau (Say), Atymna castanese (Fitch), Ophiderma pubescens (Emmons), Vanduzea arquata (Say), Entylia bactriana Germar, and Publilia concava (Say).

The following ants were actually observed by Funkhouser taking the secretion from the membracids: Formica truncicola' subspecies obscuriventris (Mayr), Formica exsectoides Forel, Camponotus pennsylvanicus (DeGeer), Crematogaster lineolata (Say), and Prenolepis imparls (Say). All these ants seemed to make no distinction between the various species of tree-hoppers listed above and the mutual behavior of these insects was much the same in all the cases studied: "The ants stroke their charges with their antennae, whereupon the membracids give off from the anal tube a liquid that issues in bubbles in considerable quantity. The anal tube of the membracid is capable of great evagination especially in the nymphs, in which it is long and cylindrical and usually tipped with a fringe of fine hairs. The honeydew is eagerly taken from the end of this tube by the ants. In many species the adults as well as the nymphs are sought, and the ants seem to be as attentive to one as to the other but the adults have not been observed to excrete the liquid to the same extent as the nymphs." (Funkhouser, 1917, p. 403.) The liquid sought by the ants "is colorless and transparent, rather heavy and somewhat sticky. When first exuded it is inclined to be frothy, due no doubt to bubbles of air which emerge with it, but it quickly clears on settling. It is practically tasteless even in comparatively large quantities, and many attempts to distinguish a sweet taste have proved unsuccessful. The term honeydew, therefore, commonly applied to the fluid, is hardly a descriptive one. It is very likely, of course, that the liquid may contain sugars not detected by the human tongue, and this would seem to be indicated by the fact that fermentation appears to begin if the substance is left exposed. No chemical analysis of honeydew has been, made." (Op. cit., p. 404.)

Miss Branch (1913, pp. 84-85) states that young Enhjlia sinuata seemed unable to molt successfully without the presence of ants. This fact led her to believe that the ants are necessary factors in the life of an individual membracid. Funkhouser's experiments, however, gave no support to this theory. Tree-hoppers of many species were reared in the field and in the insectary, with and without ants, and no difference was noted in the length of the instars or success of the molting process.

Kornhauser (1919, p. 546) gives the following account of the manner in which Thelia bimaculata (Fabricius) is attended by ants. This membracid feeds on the sap of the common North American locust tree, Robinia Pseudo Linnaeus. It deposits its eggs in slits in the bark,

where they remain during the winter, hatching in early June. The first, second, and third instars occur on the branches, constantly attended by ants:

In my principal collecting fields [at Cold Spring Harbor, New York], Formica Iruncicola Nylander subspecies obscuriventris and Cremastogaster lineolata Say were the chief ants associated with Thelia. When tapped by the antennae of the ants, the Thelia nymph or adult exudes from the anal tube a drop of clear fluid which is taken by the ant with great alacrity. Toward the middle of June, the ants build collars about the bases of the locust trees, and inside these collars in the cracks of the bark are to be found hundreds of Thelia nymphs of third to fifth instar, quietly feeding and undisturbed by the numerous ants in attendance. In this moist situation, protected from many of their enemies, the nymphs thrive. Formica builds the protecting collar of leaves, twigs, and bits of wood; Cremastogaster builds of sand grains cemented together. When one breaks the collar, many ants swarm out and attack the intruder, Formica biting one's fingers ferociously, while others grab the Thelias and drag them into underground passages. These pugnacious ants seem to have complete mastery of the Thelia nymphs.

Membracidae are sometimes carried by ants into their formicaries (Enslin, 1911, pp. 19-21; W. M. Mann, 1915, p. 162), but they usually die soon, probably due to lack of food.'

Lamborn (1914) has described in detail several cases of trophobiosis between ants and coccids, membracids, jassids, and psyllids in Southern Nigeria. Regarding Leptocentrus attifrom Walker, a tree-hopper which is invariably ant-attended in its mature and larval stages, he writes as follows: "The solicitude of ants for the larvae has a very definite object, for they are extremely partial to the fluid excreted at the anal extremity, and I remember seeing a Camponotus akwapimensis variety poultoni with the caudal whip of a membracid larva actually in its mouth." (Lamborn, 1914, p. 495.) I have on several occasions, in the Belgian Congo, collected ants which were in the act of attending tree-hoppers: so, for instance, in April 1912, at Elisabethville, Katanga, a number of worked of the common Pheidole megacephala subspecies punctulata (Mayr) were

'Additional information concerning the relations between Membraeida? and ants is given by Belt (1X74), Mrs. Rice (1893), Green (lflOO). Froggatt (1902, p. 717). Baer (1903), Buckton (1903. p. 262), Poulton (in Buckton, 1903), Distant (1908, p. 209). Enslin (1911), Miss Branch (1913), Kershaw (1913), Lamborn (1914), and others.

busily engaged in licking the sweet excretions of some of these hemipterous insects feeding on a bush; again, at Welgelegen, Katanga, Myrmicaria eumenoides subspecies opaeiventris variety congolensis (Forel) was found attending membracids fixed on the calyx of a malvaceous plant (Bequaert, 1913, pp. 427 and 428). Bell-Marley at Durban, Natal, observed that the common South African tree-hopper, Oxyrhachis tarandus (Fabricius), attracts great numbers of "small red ants." (Distant, 1908, p. 209.)

The nursing of scale insects by ants has repeatedly been noticed by Cockerell, Newstead, King, and others.1 A rather interesting phase is offered in the case of various ants which keep coccids inside the swellings of myrmecophytes. Zimmermann found Lecanium tenebricophilum Green at Buitenzorg, Java, t ogether with ant s in living branches of Erythrina lithosperma Blumc (Green, 1904, p. 204). In southern Europe, Crematogaster Scutellaria (Olivier) and Camponotus pubescent; (Mayr) often become harmful to olive trees by the care they bestow upon scale insects (Peragallo, 1882). Keuchenius (1914a and b) holds the view that (Ecophylla smaragdina is very noxious to coffee plantations through its habit of keeping and protecting in its nests the green coffee scale, Lecanium viride, one of the most serious pests to the coffee tree. Gowdey (1917) also mentions that the root form of Pseudococcus citri, a parasite of coffee, orange, lemon, and cacao in Uganda, is attended by the ant Pseudolasius gowdeyi Wheeler.

Most of the wood-boring ants either accommodate themselves to pre-existent galleries made by other insects or attack dead wood only. Occasionally they find their way into houses. Forel (1874) and R. Brim (1913) have described cases in which populous colonies of the European Camponotus ligniperdus and C. herculeanus had excavated the beams, window-sills, and other wooden parts of buildings. Certain carpenter ants of temperate regions (Camponotus ligniperdus, C. herculeanus, C. pubescens, and others) extend their burrows into healthy wood (Forel, 1874); they may thus become very destructive in forests, the more so since they attract woodpeckers, which bore large access-holes through the perfectly healthy outside layers of the tree in order to feed on the carpenter ants and their brood. 8. A. Graham (1918) describes how carpenter ants of an unidentified species are responsible for great damage to stand

lIn India the lac-producing coccid, or lac insect. Tachardia lacra (Kerr), is frequently attended by ants, Crematogaster autmuda (Mayr) and Camponotux compresiua (Fabricius), which may become a source of regular annoyance to the lac grower. In their eagerness to obtain "honeydew " the workers often nip off the white filaments, the two anterior of which are connected with the respiratory apparatus of the lac insect, the coccid being killed consequently (G. R. Dutt. 1911').

ing white cedar in Minnesota, at least twenty per cent of the trees cut showing ant injury to the stump. In this case, so far as observed, the ant never attacks a sound tree, but always gains entrance through a wound or decayed spot. When a colony has been established in a tree, the ants usually work well above the rotten area into the sound heartwood, honey-combing the tree with longitudinal galleries until there is often only a thin outer shell ot solid wood. From the main nest they cut openings to the outside, frequently following a knot, through which the sawdust can be cast and through which the inhabitants may pass to and fro. Ants which make their galleries in the bark (such as many species of Leptothorax) usually do not burrow beyond the external dead layers and occasion little or no damage, except in cases where the bark itself is of economic value: Camponotus hercideanus vagus (Scopoli) and Crematogaster scutellaris (Olivier) are credited with destroying the bark of corkoaks in southern Europe and North Africa (Maceira, 1904; Emery, 1908; Seurat, 1901; A. Krausse, 1913 and 1919).

Harvesting ants have often been accused of depredations in cereal fields, but these charges are apparently much exaggerated. Emery (1891, pp. 176-177), it is true, has observed in Italy that species of Messor actively engage in carrying off grain during the harvest. It does not seem, however, that the damage thus done could be very serious, since harvester ants collect mainly seeds of weeds and wild grasses. Yet in certain regions of North Africa, where colonies of Messor are very numerous, the grain these ants store away may amount to an appreciable portion of the harvest. Ducellier (1912) estimates that, in Algeria, Messor barbarus collects 50 to 100 liters of wheat from each hectare. J. Perez (1903, pp. xxxiv-xxxv) has recorded cases in which Messor barbarus stole freshly sown carrot-seeds and also the ripe seeds of coriander in a vegetable garden near Bordeaux. Similarly, Koningsberger (1908, p. 99), in Java, blames Plagiolepis longipes (Jerdon) with stealing planted seeds of tobacco.

A few species of ants are commonly found in houses, boats, and ships; they are spread by commerce to considerable distances, and rapidly become cosmopolitan.1 Such domestic species in the Belgian Congo include, among others, Monomorium pharaonis (Linnfeus), Tetramorium simillimum (F. Smith), and especially the many forms of Pheidole megacephala (Fabricius); the last-named is the famous house ant of Madeira (O. Heer, 1852 and 1856), which has now established it

^onisthorpe (1915. pp. 334-350) has given an interesting account of the exotic ants which have been introduced into Britain. His list includes fifty-one species, but only a small number of these have established themselves there; they are most commonly found in hothouses.

self everywhere in the tropics and subtropics.1 In the Congo, the large workers of a form of Camponotus maculatus can also frequently be seen at night in houses in search of food. They are particularly fond of sweets, of which they may absorb considerable quantities, their gaster then becoming greatly distended. At Khartum, cases of oedema of the eyelids have been ascribed to the bites of ants (Chalmers, 1918). Yet even these domestic ants should not be considered wholly noxious, because many of them are to a large extent carnivorous, thus destroying great numbers of roaches, larvs e of flies, and other indoor pests (see Illingworth, 1913 and 1917). Perhaps the most dreaded of these house ants are the fire-ant, Solenopsis geminata (Fabricius), a very pugnacious species with a severe sting, and the Argentine ant, Iridomyrmex humilis Mayr, which is becoming a serious nuisance in many subtropical countries.2

By far the majority of ants afford to the vegetation a very effective protection, destroying a large number of phytophagous insects.3 Foremost in this respect are the driver ants (Ecitonini in America, Dorylini in the Old World tropics), with their populous colonies and wandering habits, and also the many, highly carnivorous Ponerinae.4

The wandering armies of South American Ecitons have been described by H. W. Bates (1863, p. 354), Belt (1874, p. 17), and many others. Perhaps Richard Spruce's account (1908, II, pp. 370-373) gives the clearest idea of the usefulness of their operations and it is interesting. enough to be quoted at length:

Ecitons or foraging ants (called Cazadoras in Peru) seem to be true wandering hordes, without a settled habitation; for a certain number of them may always be seen carrying pupae, apparently of their own species; but they sojourn sometimes for several days whenever they come upon suitable food and lodging. . . .

The first time 1 saw a house invaded by Cazadoras was in November 1855, on the forest slope of Mount Campana, in the Eastern Peruvian Andes. I had taken up my abode in a solitary Indian hut, at a height of 3,000 feet, for the sake of devoting a month to the exploration of that interesting mountain. The walls of the hut were merely a single row of strips of palm trees, with spaces between them wide enough to admit larger animals than ants. One morning soon after sunrise the hut was suddenly filled with large blackish ants, which ran nimbly about and tried their teeth on everything. My charqui proved too tough for them; but they made short work of a bunch

lPheidole megacephals1 has of late been replaced as house ant in Madeira by the Argentine ant, Iri'lomyrmex humilisl

2Ants more commonly found in or near houses in India are, according to Assmuth (1907, p. 302), Prenolepiis longicornis (Latreille) and Monomorium pharaonis (Linnaeus).

3The activity of ants in destroying noxious insects was discussed in detail by H. Stitz (1917) in a recent paper. Delpino (1875, p. 89) expressed the view that "the ants are the chief equilibrating and moderating factors affecting phytophagous insects," perhaps a somewhat overdrawn statement.

4Some of the African Ponerinae are almost exclusively termitophagous. See, for instance, the accounts of the habits of Megaponera f<rtens (p. 65), Paltothyreus tarsatus (p. 62), and others.