Besides preying on their own species or on other blattids, cockroaches exhibit additional symbiotic relationships among themselves. These relationships are (1) the familial associations of parent and offspring, (2) gregariousness, (3) intraspecies fighting, (4) interspecies compatibility, and (5) interspecies antagonism. There are some inconsistencies between observations made on the same species by different workers, which only further observation and experimentation will explain. Some of the reported observations are unique; this is especially true for the feral species. Because of the paucity of information, it is impossible at this time to make valid generalizations about some of these interesting relationships.

FAMILIAL ASSOCIATIONS

The females of many species of cockroaches insure varying degrees of protection to the developing young in their ways of disposing of the oÖtheca after it has been formed. The extent of this association between the mother and her developing progeny varies from the minimum amount of time spent by oviparous females in concealing their oÖthecae, to the duration of embryogenesis in the so-called viviparous species, a period of over a month or more.

Haber (1920a) observed a female of Periplaneta americana chew a groove in a piece of pasteboard into which she attempted to deposit her oÖtheca. The oÖtheca failed to adhere to the shallow hole and fell to the floor. After several futile attempts to replace the oÖtheca in the hole, the female finally left the egg case on the floor of the cage and coated it with an oral secretion to which she attached bits of trash. During this operation she chased other females away when they ventured near the site. Qadri (1938) described the behavior of the female of Blatta orientalis in concealing her oÖtheca in a hole that she dug in sand; she deposited the egg case in the hole, coated it with saliva and sand, and then refilled the pit. Rau (1943) described in detail how females of P. americana and B. orientalis covered their oÖthecae with wood dust or sand in holes they had prepared in the substrate. Both species placed a sticky oral secretion in the holes and then deposited their oÖthecae therein. After coating the oÖthecae with more sticky secretion, the females adjusted the oÖthecae so that the keels were uppermost and then carefully concealed the oÖthecae with the excavated debris. Both females spent over an hour in the act. Rau (1924) previously reported that of 90 oÖthecae deposited by B. orientalis in jars containing earth and trash, 36 were placed in crevices or excavated holes, and 38 were hidden by being covered with dirt stuck to them with saliva; only 16 were left uncovered.

Edmunds (1957) described oviposition by Periplaneta brunnea. Some females spent from 30 to 40 minutes secreting from the mouth a frothy substance that was smeared on the substrate; the egg capsule was deposited in the secretion and covered with additional froth, which hardened into a very strong cement. Some females spent as long as two hours coating the capsule after it was deposited. It was not stated whether the oÖtheca was otherwise concealed. The female remained with her body over the oÖtheca for several hours and drove away other cockroaches which approached.

Sonan (1924) observed that Periplaneta americana and Periplaneta australasiae spent from 40 minutes to an hour covering their oÖthecae, and that if the females were frightened away from this activity, they returned again to complete it. As well as excavating holes in the substrate in which to deposit its oÖthecae, P. americana also avails itself of readymade crevices of appropriate size (Ehrlich, 1943). Species of Epilampra in Malaya were said by Annandale (1900) to deposit their oÖthecae in crevices in floating logs just above the water line. However, Shelford (1906) stated that four genera (including Epilampra) of the subfamily Epilamprinae are "viviparous," in which event the females would carry their oÖthecae within their bodies during embryogenesis and would not place the oÖthecae in crevices in logs.

The female of Cryptocercus punctulatus was observed to make a groove in a piece of wood, then carry her oÖtheca 6 inches from where she had dropped it and place it in the groove; she covered the oÖtheca so that only a portion was visible (Cleveland in Cleveland et al., 1934). Dr. W. L. Nutting (personal communication, 1954) collected a number of oÖthecae of C. punctulatus in the field and found each one almost completely sealed off with bits of wood in a deep groove in the roof of a chamber in a log. The keel of the oÖtheca was visible but the rest was well camouflaged. He stated that "The adult pair usually frequents the chamber at this time, while their broods of previous years occupy neighboring galleries."

Berland (1924) observed a female of Loboptera decipiens filling a hole (the abandoned nest of a hymenopteron) with earth that she carried in her mouth; he later found her oÖtheca behind the earthen barricade which she had erected.

In summary, the following species of oviparous cockroaches have been observed concealing their oÖthecae (only those references not previously cited are given): Blatta orientalis; Cryptocercus punctulatus; Ectobius sylvester (Harz, 1956, 1957); Epilampra sp.; Eurycotis floridana (Roth and Willis, 1954a); Loboptera decipiens; Balta scripta, Methana curvigera, Methana marginalis, and Methana caneae (Pope, 1953a); Pelmatosilpha marginalis, Pelmatosilpha purpurascens, and Nauclidas nigra (Bunting, 1956); Periplaneta americana (Haber, 1919; Adair, 1923; SeÍn, 1923; Nigam, 1933; Gould and Deay, 1938; Rau, 1940a); Periplaneta australasiae (Girault, 1915b; Spencer, 1943; Pope, 1953); Periplaneta brunnea (Roth and Willis, unpublished data, 1958); Periplaneta fuliginosa (Gould and Deay, 1940); Periplaneta ignota (Pope, 1953); Supella supellectilium (Flock, 1941). Undoubtedly other oviparous species that drop their oÖthecae long before the eggs hatch also make some attempt to conceal the oÖthecae by placing them in crevices or covering them with debris.

Sometimes the oÖthecae are deposited but not concealed. Hafez and Afifi (1956) reported that in Egypt Supella supellectilium attaches its oÖtheca to a suitable substrate with a gummy oral secretion but leaves the egg capsule otherwise exposed. We (1954) have noticed similar behavior in laboratory colonies of this species and of Blatta orientalis, as have Gould and Deay (1940). Cornelius (1853) stated that the female of B. orientalis takes care of the safety of her offspring to the extent of usually dropping her oÖthecae in places which are dry and raised above the ground, although rarely one also may find some oÖthecae scattered on the ground. For lack of suitable material females of Periplaneta americana sometimes did not conceal their oÖthecae (Nigam, 1933). Frequently in laboratory colonies P. americana merely drops the oÖthecae loosely in sand or food "in contrast to P. australasiae, which almost always went to considerable trouble to fasten their eggs securely and to conceal them with debris" (Pope, 1953). If conditions under which Nauclidas nigra is kept are not suitable, the female will drop her oÖtheca anywhere (Bunting, 1956). Rau (1940) stated that the female of Parcoblatta pensylvanica does not conceal her oÖtheca. However, Gould and Deay (1940) stated that this species deposits its oÖthecae loosely behind bark. Ellipsidion affine and Ellipsidion australe attach their oÖthecae to bark or the underside of leaves but apparently make no attempt to conceal them (Pope, 1953a).

The females of most of the above species have no further familial association with their offspring. The eggs hatch with no attention from the mother who is probably not even in the vicinity at that time. The young apparently do not react to the presence of the parent, as such, after hatching. This is not unexpected, as several additional oÖthecae may have been deposited by these oviparous females before the eggs of the first oÖthecae hatch. However, a different behavior is encountered among species that do not form a second oÖtheca until after the eggs of the first have hatched (see below) and in the so-called colonial species.

Shaw (1925) reported that in Australia both Panesthia australis and Panesthia laevicollis appear to live in families, and that one usually finds a pair of adults associated with from 12 to 20 nymphs in different stages of development; he continued, "it is only where the molts are very abundant that one loses sight of this familial habit." Tillyard (1926) also stated that the Australian species of Panesthia live in burrows in soil in strict family communities of a pair of adults and 10 to 20 nymphs. A related colonial species, Cryptocercus punctulatus, lives in both sound and rotten logs in colonies consisting of a pair of adults and 15 or 20 nymphs, probably representing two or three broods (Cleveland et al., 1934; Cleveland, 1948). Chopard (1938) has cited this association as an example of gregariousness, which it may well be; however, the presence of only one pair of adults in each colony suggests a more intimate relationship.

Among species of Blattella and certain other genera with similar reproductive habits the female carries her oÖtheca clasped in her genital cavity with the posterior portion projecting behind her. Each normal oÖtheca is carried for approximately the duration of embryogenesis and is not dropped until, or shortly before, hatching. We have seen (1954, fig. 65) newly hatched nymphs of Blattella vaga crawl over the body of the mother who stood quietly near the dropped oÖtheca; this female raised her wings and some of the nymphs crawled under them onto the dorsal surface of her abdomen. The nymphs seemed to feed on the grease covering the mother's body. The association was short-lived, however, and soon the nymphs scattered. Pettit (1940) stated that when hatching of Blattella germanica occurs in the open (on a table top), the nymphs may remain near the capsule only a few minutes. Ledoux (1945) found that newly hatched nymphs of B. germanica remained together without shelter in a single, sparse group. If the nymphs were separated by blowing on them, the group quickly reassembled, usually in the same spot. Ledoux showed that this gregarious grouping of first-instar nymphs was not necessarily a familial association by placing nymphs from two oÖthecae together. In groups of 8 to 12 nymphs there was a perfect intermingling of the offspring from the two different females.

It is among the so-called viviparous cockroaches that the greatest number of observations have been made of postparturient associations between female cockroaches and their offspring. The females of these species carry their oÖthecae in brood sacs within their bodies until embryogenesis has been completed. This behavior ensures protection of the young from desiccation and attack by parasites (Roth and Willis, 1955a). (See Roth and Willis, 1958a, for an analysis of oviparity and viviparity in the Blattaria.) Shelford (1906, 1916) reported that he had captured a female of Pseudophoraspis nebulosa in Borneo with numerous young nymphs clinging to the undersurface of her abdomen. He also recalled that there was in the Hope Museum (Oxford) a female of Phlebonotus pallens to which the following label was attached: "'Ceylon ... carries its young beneath its wing covers. 1878.'" Pruthi (1933) found in South India another female of P. pallens which was carrying over a dozen young nymphs on her back beneath her wings. In his paper Pruthi reproduced a photograph of this specimen with the light-colored nymphs in place on the back of the female. Hanitsch (1933) reported having seen a museum specimen from Luzon, Philippine Islands, of the apterous female of Perisphaerus glomeriformis with nymphs still clinging to her undersurface; he also reported having seen a museum specimen of a female of Ellipsidion variegatum from Australia with four young clinging to the upper side of the apex of her tegmina and six to the oÖtheca which projected beyond her body. Presumably this specimen was giving birth when captured. Gurney (1954; personal communication, 1958) stated that specimens of Perisphaerus sp. from Mindanao and Luzon have been found with young nymphs clinging to the middle and hind coxae. The first-instar nymph has an elongate face and specialized galeae. Karny (1925) also observed that at the slightest alarm the young of some species of Phoraspidinae creep under the dome-shaped front wings of the mother.

The newly hatched young of Leucophaea maderae have also been seen congregated under the mother on several occasions. SeÍn (1923) stated that after being born, the nymphs of this species gather under the mother and accompany her at night in her excursions in search of food. PessÔa and CorrÊa (1928) reported that "During the first days the free larvae hide under the adult cockroach which becomes restless and active in contrast to its usual slow gait." Wolcott (1950) stated that "They are not only gregarious, but the mother broods over her young, and together they sally forth at night in search for food, until they are of such a size as to mingle with their elders."

The African mountain cockroach Aptera fusca has been observed during late summer and early winter in familial groups beneath loose bark, under stones, and in dead leaves (Skaife, 1954): "Each party consists of a number of black young ones, together with one, two or more adult females and perhaps a winged male or two. Later on they scatter and live more or less solitary lives." In Malaya Karny (1924) often found phoraspidine females between leaves surrounded by about 20 young nymphs. He stated that one also often found females of Perisphaerus armadillo surrounded by pale, yellowish-white young; similarly he had observed that Archiblatta hoevenii was found mostly in colonies made up of mothers and their young. The duration of these associations is not known.

Saupe (1928) noticed that the newly hatched nymphs of Blaberus craniifer (see footnote 12, p. 322) collected together under the body of their mother and stated that this is as pronounced a case of brood care as Zacher had observed with Pycnoscelus surinamensis. Nutting (1953) stated that "A degree of maternal solicitude is exhibited by this roach [B. craniifer], for many times I have observed the female to remain motionless for an hour or more with her unpigmented brood clustered around and beneath her body." We, too, have observed similar behavior in laboratory colonies of B. craniifer and Leucophaea maderae.

Chopard (1950) noted that after hatching the young of Gromphadorhina laevigata remained grouped around the female for some time; the mother stood motionless, high on her legs, with her thorax curved up to make room for the brood which hid under her body. We (unpublished data, 1958) have seen young nymphs of Gromphadorhina portentosa also stay near their mother for some time after birth; the mother at this time produced a characteristic hissing sound when she was only slightly disturbed by the movement of our hand near her and her brood. The sound is produced as air is expelled through the second abdominal spiracle. We have seen recently hatched nymphs of Nauphoeta cinerea crawl beneath the mother, even under her wings, where they remained about an hour (Willis et al., 1958). Bunting (1956) observed a female of Blaberus discoidalis collect a mound of debris into which she inserted the tip of her abdomen; he found young in the mound later the same day. This female showed no maternal care for the young after birth. Whole families of cockroaches may be found in bromeliads in Brazil (Ohaus, 1900). Hebard (1920) observed a colony of adults and young of Dendroblatta sobrina on a tree trunk in the Panama Canal Zone.

Whether any of the above associations exemplify maternal care for the newly hatched young is questionable. The behavior of the mother, beyond placing her eggs in a suitable location, seems to be entirely passive. The first-instar nymphs are the active partners in these associations, and they may merely be seeking shelter under the nearest convenient object rather than under the mother as such. More extensive studies of some of these relationships will be needed before claims for maternal care, as suggested by Scott (1929), can be substantiated.

GREGARIOUSNESS

Casual statements that cockroaches are gregarious are often encountered in the literature. There has been some argument to the effect that large numbers of these insects seeking the same environment in a limited space would appear to be gregarious, whereas there is probably no true social tendency (Rau, 1924). Reactions of cockroaches to certain stimuli in the environment undoubtedly do result in aggregations of individuals. However, as Chopard (1938) has pointed out, it is difficult to assign the respective parts played in assembling by the attraction of the milieu and by gregarious instincts. Chopard (1938) also stated that Orthoptera with a gregarious tendency are found rarely isolated; one finds them, on the contrary nearly always collected in the same shelters, close together, as if conscious of a need for contact between themselves. He continued further that one can be tempted to attribute the assembling to taxes but that interattraction equally plays an important role; for example, if one places a large number of cockroaches in a container and offers them similar shelters composed of cardboard tubes, one finds that nearly all the individuals will assemble in one of the tubes, ignoring the others. Pettit (1940) claimed that in Blattella germanica gregariousness seemed to depend on the mutual attractiveness of body secretions as well as a thigmopositive behavior and love of warmth.

Ledoux (1945) has studied experimentally gregariousness and social interattraction in Blatta orientalis and Blattella germanica. He also found that the cockroaches tended to collect in shelters containing other cockroaches. He concluded that group formation is not the result of chance, but is a social phenomenon, and that interattraction is mainly olfactory, conditioned by (1) positive chemotaxis to odors emitted by the cockroaches themselves, (2) positive hygrotaxis, and (3) thigmotaxis. He found also that large groups are not stable and tend to break into smaller groups.

Gregariousness in the Orthoptera varies in intensity according to the species and within a species according to the age or physiological state of the insects (Chopard, 1938). This is well exemplified by several of the blattid species discussed below.

Gregarious groupings of cockroaches have been observed most frequently among the domiciliary species. A few examples will suffice. Gal'kov (1926) observed heavy infestations of undetermined cockroaches in workers' living quarters in the Ural region: "In the corners near the stove, the cockroaches covered the walls in a dense carpet." After fumigating he collected about 135,000 dead cockroaches from one barracks and about 475,000 from another. We have reviewed a few other examples of heavy infestations in our 1957(a) paper.

Periplaneta americana was observed by Gould and Deay (1938) in an old meat-packing building in Indiana. Adult cockroaches were present in large numbers between closely placed beams, but the nymphs were more common in cracks between bricks. Clusters of several hundred cockroaches were seen on the open walls of the cold, dark hide room. Gould and Deay stated "American roaches of all sizes live together in perfect harmony. Young nymphs have been noted in clusters underneath adults and crawling over the adults as they wander about in rearing jars." In the monkey house of the Hamburg zoo, P. americana spent most of the day in the cellars resting on the walls in groups of about 200 individuals (Brecher, 1929). Lederer (1952) noted that in closed, dark, heated spaces under the aquarium at Frankfurt am Main, P. americana rested in groups of 20 to 30 individuals; he stated that it was remarkable that the "herd" divided itself into groups each of which usually contained insects of the same age or stage of development. Eads (1954) found P. americana in 40 percent of 762 sewer manholes in Tyler, Tex.; 13 percent of 670 of these manholes were heavily infested with 100 or more cockroaches in each. Other heavy sewer infestations have been reviewed in our 1957(a) paper.

Ehrlich (1943) has stated that Periplaneta americana exhibits social behavior. For instance, cockroaches of various ages inhabit a fairly large space jointly; the adults and older nymphs sense approaches with their antennae and warn and protect the young by a beating of wings and by body movements. There is complete utilization of the available living space; the imagos drive older nymphs from their resting places, and the older nymphs drive out the younger ones, until all cracks, depending on their size, are occupied by various age groups of different sizes. In his experiments Ehrlich observed that in cages with no hiding places the cockroaches would group together; when given a choice of small and large shelters, P. americana hid only under the larger ones that could shelter more insects. Finally, the cockroaches ceased to bite and fight each other when they crowded together in the face of danger.

Of Blatta orientalis Marlatt (1915) stated "This species is notably gregarious in habit, individuals living together in colonies in the most amicable way, the small ones being allowed by the larger ones to sit on them, run over them, and nestle beneath them without any resentment being shown." Haber (1919) also observed that this species is often noticed "huddled together, the younger ones crawling over, around, and beneath the older ones."

Wille (1920) observed that nymphs of Blattella germanica remained almost constantly in groups during the first and second instars, but less so during the third instar. He believed that the aggregations of young occurred because they could occupy narrow crevices where the larger insects could not penetrate. At usual room temperatures the older nymphs and adults lived completely isolated, but at certain temperatures they gathered together in large, tightly pressed groups.

Supella supellectilium is said to be gregarious (Gould and Deay, 1940). The smaller nymphs aggregate in small groups in rearing containers, but the older ones remain separate from one another (Hafez and Afifi, 1956). Leucophaea maderae is sociable and rarely found alone; in their favorite hiding places, hills of these cockroaches can be seen hanging together (SeÍn, 1923). Wolcott (1950) also stated that L. maderae is gregarious. Annandale (1900) observed that in the "Siamese Malay States" large colonies of Periplaneta australasiae conceal themselves in hollows of bamboo logs from which houses are built. Moulton (1912) stated that he was astonished at the large numbers of P. australasiae and Symploce cavernicola that he saw swarming on the sides of caves of Mt. Jibong, Borneo.

Rehn and Hebard (1905) stated that in Key West, Fla., Eurycotis floridana fairly swarmed under the coquina boulders in the woods, in groups of a dozen containing both young and adults; Pycnoscelus surinamensis was very abundant in the same type of habitat. Caudell (1905) also found the young of E. floridana with the mature individuals. Hebard (1917) in his discussion of Lattiblattella rehni again mentioned finding frequent colonies of E. floridana in Florida. He also found many specimens of Blaberus craniifer under boards on the ground at Key West. He found Parcoblatta lata numerous under bark of dead pine trees in Alabama. However, Dowdy (1955), in an ecological study of oak-hickory forest in Missouri, stated that "Parcoblatta [sp.] were never recorded as being gregarious, in fact they were mostly solitary. However, in some cases two were found together." Yet Blatchley (1895) stated of Parcoblatta pensylvanica that in the winter in Indiana "One cannot pull the loose bark from an old log without dislodging a colony of from ten to a hundred of the nymphs of various sizes." Males of Parcoblatta virginica were said to be often gregarious beneath loose bark and under chunks and rubbish (Blatchley, 1920).

Rehn and Hebard (1927) quoted observations made earlier by Hebard on Byrsotria fumigata in Cuba: "I found the specimens under flat stones, sometimes in colonies of 3 or 4 mature specimens and numbers of immature individuals in all stages of development." These observers also reported that Aspiduchus borinquen was found in Puerto Rico in a limestone cavern by thousands in the grass and on the walls. J. W. H. Rehn (1951a) stated that a related species, apparently Aspiduchus cavernicola, was seen in great numbers on the side walls and roof of a cave in Puerto Rico, but it was not possible to collect any of these and, we infer, confirm the species. Rehn and Hebard (1927) in their account of Simblerastes jamaicanus reported finding it in numbers in a termite nest. Pemberton and Williams (1938) stated that Diploptera punctata is of gregarious habits in Hawaii. Saupe (1928) observed a strong "herd instinct" in all age groups of Blaberus craniifer. Bunting (personal communication, 1956) stated that large nymphs and adults of Blaberus discoidalis "congregate in narrow cracks or on the underside of some low object. The younger nymphs keep in close communities of approximately the same age." Sonan (1924) stated that in Formosa(?) Salganea morio is usually found in groups of six or seven in decayed trees. Species of the genus Litopeltis may be found in small groups as they are somewhat gregarious (Rehn, 1928).

The physiological or psychological effects of gregariousness, or lack of it, are interesting aspects of the basic phenomenon. Landowski (1937) studied in Blatta orientalis the effect on development and growth of the transition from life in complete isolation to life in groups. He kept nymphs in groups of 1, 2, 4, 8, and 16 in jars of identical size and shape. Landowski found that (1) mortality increased with the size of the group and with age, as each animal occupied more of the available space. [Presumably these factors are less detrimental in nature where the group is unconfined.] He further found that (2) life in complete isolation extended the time required to produce an adult insect; and (3) the mean weight of the adult insect was, generally, in inverse proportion to the number of nymphs raised together; isolated insects usually attained the greatest adult weight.

Similarly, Griffiths and Tauber (1942a) found that isolation extended the period of nymphal development in Periplaneta americana. As most of their isolates died before reaching maturity, these workers concluded that the American cockroach does not thrive when individually isolated and that several individuals must be together for optimum development to occur. Pettit (1940, 1940a) observed that isolated nymphs of Blattella germanica take longer to mature than those reared in groups. Wallick (1954) found indications in B. germanica that there is an inverse relationship between population density and individual weight; as the population decreased the weight increased. He also noted an inverse relationship between population density and life expectancy in this species.

We (Willis et al., 1958) have confirmed the above observations that Blattella germanica, Blatta orientalis, and Periplaneta americana complete nymphal development in less time when reared in groups rather than individually. We (loc. cit.) also found that nymphs of the following additional species matured more quickly when reared in groups: Eurycotis floridana, Periplaneta fuliginosa, Supella supellectilium, Nauphoeta cinerea, and Pycnoscelus surinamensis; only a very slight decrease in the average length of the developmental period was found in grouped nymphs of Leucophaea maderae.

Wharton et al. (1954) observed that virgin adult males of Periplaneta americana that had been individually isolated upon emergence were almost wholly unresponsive to the sexually stimulating, female odor for a test period of four weeks. Similar males of comparable age that were kept in groups reacted strongly from the sixth day on. Removal of reactive males from the group inhibited the reaction in these isolates, but the response returned when the insects were regrouped. We (1952) had similarly observed that no isolated male of Blattella germanica was ever seen to give a courting response without having received some form of external stimulation. Yet when numbers of males were kept together isolated from females, on several occasions the males became active and a few individuals gave a courting response. As the sexual stimulus is received by the male of B. germanica through contact rather than odor, as in P. americana, presumably it was mutual contact between the grouped males that released the courting activity.

Cloudsley-Thompson (1953a), in his studies of diurnal rhythms in Periplaneta americana, observed a steady decline in total activity in successive 24-hour cycles: "When two cockroaches, even of different species (P. americana and P. australasiae) were kept together, this depression did not appear to set in so readily." The associates apparently kept each other active.

Isolated females of Periplaneta americana can be conditioned to run a simple maze with less time and fewer errors per trial than when paired or when a member of a group of three (Gates and Allee, 1933). There was less activity, and accordingly fewer errors per minute, among cockroaches tested as pairs and groups of three than as isolated individuals. This observation should not be contrasted with that of Cloudsley-Thompson (1953a), cited above, because the intervals during which activity was observed were quite different.

In the above account we have presumed that aggregations of some species are indications of gregariousness. However, until gregariousness has been proved experimentally for each species, we concede that reactions to environmental stimuli might be sufficient to bring about some of the observed groupings without any interaction between individuals.

In concluding this section we note that Tepper (1893) stated that carnivorous cockroaches in Australia lead more or less solitary lives, and that one rarely meets several together in close proximity. Takahashi (1940) observed that in Formosa Blattella humbertiana does not have a tendency to throng together. Rau (1947) stated that the adults of Ischnoptera deropeltiformis showed no tendency toward gregariousness, but in the laboratory newly hatched young lived close together under bark and remained together throughout the nymphal stages. We wonder whether this gregariousness was not imposed by the restricted quarters of the cage. As mentioned above, Dowdy (1955) did not find Parcoblatta sp. to be gregarious in the field.

INTRASPECIES FIGHTING

Fighting occurs among cockroaches of the same species over food or shelter or between males. Saupe (1928) observed late-instar nymphs of Blaberus craniifer attack each other and even adults. Additional records cited in the section on intraspecies predation (p. 322) imply fighting within a species. Rau (1924) saw a male of Blatta orientalis attack another male in copula and bite away a large portion of its wing. Two other males in the container had their wings badly torn overnight, presumably as a result of fighting.

Ehrlich (1943) stated that individuals of Periplaneta americana that are feeding will ward off intruders by spreading their wings and pushing with their hind legs. However, the intruder will approach again and again biting the feeder in the legs and wings. Frequently the odor of approaching food was sufficient to cause the cockroaches to fight and bite each other. Biting and fighting also occurred when individuals of this species defended their daytime hiding places. A position of attack is assumed when two antagonistic individuals of P. americana meet (Ehrlich, 1943, fig. 14). The insects raise their bodies slightly above the ground, by extending their legs, and they stretch their heads forward horizontally so that their mouth parts protrude; when the insects jump at each other, they may wound each other severely in the soft parts of the body. Fighting between sexually excited males resulted in injury to their legs, wings, cerci, and other parts of the body. Frequently an insect that could no longer defend itself was killed. Lederer (1952) also made similar but less extensive observations on fighting in this species.

Pettit (1940) quoted Woodruff as stating that nymphs of Blattella germanica, apparently healthy and perfectly normal, would do battle for no apparent cause other than a chance meeting, and that occasionally the fight was to the finish, the loser being eaten. Pettit could not substantiate such voracious attacks, although he saw nymphs engage in fights lasting about two seconds during which one would be driven off by vigorous bites on legs or cerci. Small nymphs of B. germanica tended to ignore each other, but third-and later-instar nymphs would engage in "quarrels" of short duration when two met. Pettit noted that males of B. germanica that were crowded together quickly set upon, but did not always kill, other cockroaches introduced into their cage. When he isolated a dozen males in a small cage, they became quarrelsome and three of the group were killed and partly eaten. After several days the surviving males had taken positions so that each was equidistant from his neighbors. Some of these males attacked other males and a female that were introduced, by biting their legs and cerci. Females under similar conditions were much less aggressive, although Pettit saw some females that roved about biting all large members of the group that were within easy reach.

We have frequently observed aggressive behavior between males of Nauphoeta cinerea, which resulted in torn wings. The males would wrestle with each other rolling over and over.

INTERSPECIES COMPATIBILITY

We agree in essence with Chopard (1938) who stated that it is improper to speak of associations apropos of the ecological distribution of Orthoptera. He continued that it is clearly evident that different species of Orthoptera, which are found grouped on a territory more or less narrowly limited, have no interdependence among them. Their grouping results uniquely from almost similar reactions to the different factors which characterize this limited milieu. There is neither interdependence nor interaction; the grouping is a false biocoenose, born under the action of the environment, and does not survive a modification of this milieu.

However, as there are numerous examples of mutual toleration between different species as well as examples of incompatibility, the subject has more than academic interest even if no true ecological significance. On the other hand, further study may show that certain of these associations are definitely ecological, particularly among the feral species. As might be expected, most of the following examples pertain to domiciliary cockroaches.

Dozier (1920) occasionally found Periplaneta americana with Eurycotis floridana in decaying stumps, beneath loose bark of decayed trees, and beneath corded wood. Adair (1923) stated that in his house in Egypt Periplaneta americana, Blatta orientalis, and Blattella germanica were found together in a cupboard. Sambon (1925) found B. orientalis and B. germanica side by side but not fraternizing in a home in Italy. Gould and Deay (1938) observed that apartments over stores were infested with both B. germanica and P. americana, but did not indicate whether these occupied the same microhabitat. Gould and Deay (1940) observed that in the Purdue University greenhouse Periplaneta fuliginosa was found "under benches, boxes, pots and other objects in association with the American roach." Dr. L. A. Hetrick (personal communication, 1954) wrote us that several summers before he had had a mixed infestation of cockroaches, which included Periplaneta australasiae, Periplaneta fuliginosa, and Pycnoscelus surinamensis, in his chicken shed.

Eads (personal communication, 1955), in response to our inquiry about the mixed populations of cockroaches that he had reported infesting sewers in Texas (Eads et al., 1954), stated that "Each of the ten colonies of B. orientalis found in Tyler manholes were associated with larger colonies of P. americana. True breeding colonies of B. orientalis appeared to be present since all the developmental stages were taken. The same situation existed with the P. fuliginosa and the two species of Parcoblatta. Larger colonies of P. americana were associated with the other species in each case. From our limited observations the two species always appeared to be perfectly compatible." Eads et al. (1954) had found Periplaneta fuliginosa in three manholes, Parcoblatta bolliana in one manhole and Parcoblatta pensylvanica in one manhole. We assume that the groups of each species were spacially discrete so that they were recognizable as colonies. Dr. T. A. Olson (personal communication, 1958) has observed two or more species of cockroaches in a single structure but never in mixed colonies. Each species was separated physically from the others. Olson concluded that cockroaches of different species do not mingle freely unless forced to do so by some special environmental condition. Pettit (1940) found B. germanica and P. americana similarly separated in the same building or even in the same basement laboratory.

Perkins (1899) found Lobopterella dimidiatipes generally living in company with the young of Periplaneta australasiae in Hawaii. Rehn and Hebard (1914) in Florida found P. australasiae abundant with Periplaneta americana on a quarter-boat. They also noted that the forficulid Marava [= Prolabia] arachidis (Yersin) appeared in numbers in a kitchen after dark accompanied by swarms of P. americana. These workers also found Leurolestes pallidus in a fruit store in Key West "where the species was common in a pile of old burlap bags and in cracks under the stands which it shared with one fairly large colony of Blattella germanica, occasional specimens of Holocompsa nitidula, a few specimens of Periplaneta americana, and one specimen of Supella supellectilium." They also found H. nitidula with Blaberus craniifer "between old boards in a woodshed, where nymphs were more numerous than adults."

Rehn and Hebard (1914) stated of Supella supellectilium in Florida that "The females were all taken in cupboards where Blattella germanica was found in swarms." The association in human habitations of S. supellectilium and B. germanica has been reported also by Sein (1923), Puerto Rico; Shaw (1924), Australia; Mallis (1954): "German and brown-banded roaches were often found in the same crevice."; Anonymous (1958), Texas; and Anonymous (1958a), Georgia. Gould and Deay (1940) stated that other species of cockroaches, especially B. germanica, may be found with S. supellectilium in the same part of a building. Yet Shaw (1925) stated that "when Supella supellectilium Serv. invades places already occupied by Blattella germanica L., it tends to oust the latter."

Blaberus discoidalis has been found in homes or in fruit debris in Puerto Rico in company with the more common, domiciliary species Leucophaea maderae, but never in abundance (Sein, 1923; Wolcott, 1950). Illingworth (1915) in Hawaii found Symploce hospes associated with Nauphoeta cinerea, Graptoblatta notulata, and Diploptera punctata.

Hebard (1917) found Aglaopteryx diaphana in a bromeliad on a forest tree in Jamaica together with Nyctibora laevigata and numerous Cariblatta insularis. He also found numerous Aglaopteryx gemma under signs on longleaf pines in Alabama with occasional specimens of Parcoblatta lata. In Virginia he found Parcoblatta uhleriana in a decaying chestnut log with Cryptocercus punctulalus. In Florida he found Latiblattella rehni with Eurycotis floridana and, more rarely, with Periplaneta australasiae under bark of pine trees. In Key West he found Symploce hospes in the cupboard of a hotel with swarms of Blattella germanica and a few Supella supellectilium.

Rehn and Hebard (1927) in their study of West Indian blattids reported finding Neoblattella proserpina in epiphytic bromeliads in Jamaica in company with Neoblattella eurydice and Neoblattella dryas. They also list most of the associations cited by Hebard (1917).

Ramme (1923) reported that he found in Germany four species of Ectobius (lapponicus, lucidus, pallidus, and sylvester) living together in an area about 50 m. by 200 m. Although he had stated that his specimens of E. lucidus were a distinct species in 1923, Ramme (1951) later decided that they were a form of E. sylvester, E. sylvester f. lucidus.

Dow (1955) reported trapping Blattella germanica, Periplaneta americana, and Periplaneta brunnea in houses and privies in south Texas. At our request Dr. Dow (personal communication, 1958) analyzed his records to determine whether there were indications of associations between these species, with the following results:

As stated in my published note, the roaches were at first classified to genus only. The 83 Periplaneta subsequently identified to species represented 28 different collections, 11 from houses and 17 from privies, all in Pharr, Texas. Tabulation of the data shows first that P. americana was taken only once in a house and that P. brunnea was taken only 4 times in privies. Of course this distribution greatly reduces the probability that they would be caught together, and it is not surprising that P. americana was trapped alone in the single house collection. P. brunnea, however, was trapped with P. americana 2 of the 4 times it occurred in privy collections.

To investigate the occurrence of Periplaneta with and without Blattella, an analysis has been made of 560 trap collections taken in 40 houses and 40 associated privies in Pharr, Texas, in weekly intervals (from May 14 to June 23 [1948]). In the houses, Periplaneta and Blattella were caught in the same jar 26 times, Periplaneta alone 12 times, Blattella alone 83 times, and neither genus 159 times. In a fourfold table, the value of chi-square (14.7) is significant and indicates that the frequencies are not proportional. The number of times Periplaneta and Blattella actually occurred together (26) is, however, much larger than the expected number calculated from the row and column frequencies (14.8). In the privies, Periplaneta and Blattella were caught in the same jar 9 times, Periplaneta alone 50 times, Blattella alone 18 times, and neither genus 203 times. In a fourfold table, the value of chi-square (1.95) is not significant but the same type of disproportion is evident and the expected frequency of both genera in one trap is 5.7, lower than the actual frequency of 9. Both immature and adult roaches are included in this analysis.

The above evidence would be more satisfactory if based on more extensive data. There is also a possible objection in that the traps were operated for at least overnight, during which time one species could theoretically supplant another. Of course, it is doubtful that there is anything involved here like territory (in the ornithologists' sense). On the other hand, it is well to consider that Periplaneta and Blattella are both likely to be more abundant in the same type of favorable location and that this factor might offset in part some direct antagonism between the species.

The only known specimen of Ischnoptera podoces was captured in company with the type series of Cariblatta nebulicola, in dead leaf litter in Jamaica (Rehn and Hebard, 1927). In Florida Periplaneta australasiae was often taken in company with Pycnoscelus surinamensis and Eurycotis floridana (Blatchley, 1920).

INTERSPECIES ANTAGONISM

In contrast to the presumably amicable associations mentioned above, other observations in the literature seem to indicate that some species of cockroaches are incompatible when they attempt to occupy the same habitat niche. Marlatt (1915) stated "Rarely do two of the domestic species occur together in the same house. Often, also, of two neighboring districts one may be infested with one species, while in the other a distinct species is the commoner one. The different species are thus seemingly somewhat antagonistic, and it is even supposed that they may prey upon one another, the less numerous species being often driven out." Phelps (1924) stated "Roaches of different species are rarely found together, although roaches of the same species live together on very amicable terms."

In 1859 Darwin (1887) stated that "In Russia the small Asiatic cockroach [Blattella germanica?] has everywhere driven before it its great congener [Blatta orientalis?]." Yet in France Girard (1877) suggested that the oriental cockroach be introduced into a restaurant infested with the German cockroach as the best way to expel the latter, because the more robust species drives away cockroaches of smaller size. Wille (1920) in Germany found usually only one species of cockroach in a house. Yet when he placed B. orientalis and B. germanica together, there were no reciprocal attacks even by hungry individuals. Wille concluded that because of their greater speed, smaller size, greater number of eggs, and faster development, the German cockroaches eat the available food and so make the environment unfavorable for the oriental. However, he noted that cases may be seen in which the opposite is also possible. Laing (1946; British Museum [Natural History], 1951) observed that in the British Isles B. orientalis seems to have lost its dominant position to B. germanica in recent years; it was stated that these species are not as a rule found together and that the greater rapidity of breeding and ability to climb of B. germanica, as well as the layout of modern buildings, are some of the factors that favor the spread of B. germanica. Ledoux (1945) found that first-instar nymphs of B. germanica and fourth-instar nymphs of B. orientalis, adults of B. germanica and sixth-instar nymphs of B. orientalis, as well as adults of both species, did not form mixed groups. However, when he combined fifth-and sixth-instar nymphs of B. germanica with fourth-and fifth-instar nymphs of B. orientalis, which are all practically of equal size, sometimes he would find mixed groups, but generally the groups were distinct. Lucas (1912) stated that Burr had found B. germanica and B. orientalis swarming within a rubbish heap in England; presumably both colonies were breeding and multiplying and one species was not detrimental to the presence of the other.

Shaw (1925) claimed that Supella supellectilium tended to oust Blattella germanica, but Pope (1953) thought it doubtful in Queensland. Wolcott (1950) stated that "The larger and more powerful domestic cockroaches, Periplaneta americana (L.), P. australasiae (F.) and P. brunnea Burmeister have very definitely fallen behind in Puerto Rico in competition with the little German roach." PessÔa and CorÊa (1928) observed that other species of cockroaches were rare in Brazil in houses that were infested with Leucophaea maderae. Lederer (1952) noticed that in the reptile house of the aquarium at Frankfort am Main Blatta orientalis was obviously kept down by Blattella germanica, even before the appearance of P. americana. However, B. germanica was not driven out of the reptile house by P. americana although the populations of each fluctuated for about 22 years after the American cockroach had settled there; both species occupied separate resting places. Lederer further observed that within four years of the introduction of P. americana into the crocodile house, none of the original infestation of B. orientalis could be found; a small colony of Pycnoscelus surinamensis in the reptile house was apparently also driven out by P. americana. Chopard (1932, 1938) stated that the oriental cockroach does not exist in company with P. americana which very probably destroys it. Pettit (1940) kept B. germanica and P. americana together in a cage for several weeks but neither species gave any indication of feeding on the other.

Froggatt (1906) stated that "It is probable that the advent of the larger and more formidable American cockroach into Australia has led to the retirement or destruction of our indigenous species" [presumably Periplaneta australasiae]. Tillyard (1926) noted that this statement is incorrect as neither species is native to Australia. Yet Shaw (1925) stated that in Australia "When both species live together in the same places, australasiae Fabr. will probably be found gradually to displace americana L." Local fluctuations in the relative abundance of these species could be a basis for such dissimilar observations. However, MacDougall (1925) observed that in the plant houses of the Royal Botanical Garden, Edinburgh, the Australian cockroach seemed to have overcome the American which had been more numerous in former years.

In conclusion, we emphasize that many of the above observations are merely tentative impressions gathered by workers who have watched many species of cockroaches in nature. Obviously, additional observations coupled with appropriate experimentation will be needed to disclose the true structure of each presumed association and to resolve apparent discrepancies. Although we are greatly indebted to the cited authors for their contributions to the known information, we anticipate that future results of cleverly designed laboratory experiments will do much to dispel the uncertainty that still surrounds our knowledge of the relations of the Blattaria to each other.