Tepper (1893) made the broad statement that the majority of Australian and Polynesian cockroaches appear to be wholly carnivorous, eating other insects, eggs, and larvae. He stated that, because of their voracity and cannibalistic tendencies, the carnivorous species lead more or less solitary lives so that one rarely meets several in close proximity; they are never very numerous at any time because the stronger devour the weaker in the absence of other prey. Tepper stated that Australian species of Ischnoptera hunt for their prey among the foliage of shrubs, and that Australian species of Cutilia [= Drymaplaneta, Hebard (1943)] run about actively on the surface, or ascend shrubs and trees in quest of living insects and therefore are highly beneficial. Tepper (1894) also stated that Geoscapheus robustus ate earthworms, grubs, and caterpillars. Froggatt (1906) and Marlatt (1915) attributed to Tepper the statement that cockroaches, like Epilampra notabilis, which are found out-of-doors in Australia, are carnivorous and feed on caterpillars and other soft-bodied insects; but Froggatt (1907) believed that this alleged behavior needed confirmation.

A number of observations have been recorded which indicate that sometimes cockroaches may be predatory. According to Ealand (1915), nymphs of the cockroach Pseudomops cincta fed on the Argentine ant Iridomyrmex humilis. In the laboratory, Eurycotis floridana has been observed to catch and devour the wasp Anastatus floridanus which parasitizes the eggs of Eurycotis (Roth and Willis, 1954a). Parcoblatta pensylvanica was observed devouring a larva of Polistes sp. in its cell in a deserted wasps' nest (Rau, 1940). Brigham (1866) saw a cockroach kill and eat a centipede four or five inches long. Annandale (1910) described the destruction in Calcutta of termites by Periplaneta americana. During a heavy rain storm many termites flew into the dining room and were set upon by the cockroaches which seized them with their mandibles and began to gnaw their abdomens. If disturbed, the cockroaches carried the termites away in their mandibles without using their legs to seize, hold, or carry the prey. Sometimes only the abdomen, but other times the whole body with the exception of the wings, was devoured. Perhaps this observation led Allyn (Anonymous, 1937) to theorize that, first, cockroaches could eradicate termites from houses, and then the blattids in turn could be eliminated. Falls (1938) has pointed out the unfeasibility of this idea. Blattella vaga has shown some tendency to eat plant lice (Flock, 1941a). Certain small cockroaches found beneath cane leaf-sheaths, in the Philippine Islands, preyed in part upon leafhoppers (Uichanco, in Williams et al., 1931).

Takahashi (1924) stated that the American cockroach will eat the eggs of the hemipteron Cantao ocellatus (Thunberg). Cunliffe (1952) observed mite-infested cockroaches (Blatta orientalis, Blattella germanica, and/or Periplaneta americana) dislodge and eat the mite Pimeliaphilus podapolipophagus. Sonan (1924) reported that cockroaches (P. americana and P. australasiae) devoured the egg clusters and first instar larvae of Prodenia litula and the first instar larvae of Attacus atlas which were being reared in the laboratory. Lederer (1952) stated that Periplaneta americana ate reptile eggs in the aquarium at Frankfurt am Main. Pettit (1940) stated that cockroaches "are said to have destroyed a large colony of dermestids used to skeletonize carcasses at the University of Kansas."

DeFraula (1780) believed that his silent "gryllon" [obviously Blatta orientalis from his drawings; see Willemet (1784)] was the enemy of the chirping species of cricket, because after the cockroach became established in his home he no longer heard crickets chirping. Gilbert White (1905 ed.), writing in England in the late 18th century, stated that "Poda says that these [Blatta orientalis] and house crickets will not associate together; but he is mistaken in that assertion"; however, in August 1792 White noted that "Since the blattae have been so much kept under, the crickets have greatly increased in number." For several years Jolivet (1950) had observed changes in a mixed population of Blatta orientalis and Acheta domesticus in an old kitchen in France. He suggested that the cyclical fluctuations in the relative abundance of the cockroaches and crickets might be caused by reciprocal predatism with one species more susceptible than the other at certain stages. Mallis (1954) has stated that crickets prey on other insects as well as on one another. LhÉritier (1951) had also observed crickets becoming rare in bakeries in France, having been superseded everywhere by B. orientalis; however, he doubted that Jolivet's hypothesis was the correct explanation and suggested that the higher optimum temperature requirements of crickets might be the regulating factor. Lederer (1952) stated that the number of crickets decreased in the aquarium buildings at Frankfurt am Main as the population of American cockroaches increased.

Platyzosteria novae seelandiae was found under the bark of trees in New Zealand devouring bugs (Walker in Shelford, 1909b).

For years it has been believed that cockroaches feed on bedbugs (Cimex lectularius L.) and this statement has been repeated in many reference works and articles. Ealand (1915) stated that cockroaches devour bedbugs with avidity. Even today similar statements are to be found in the literature. "In the old sailing ship days, they [cockroaches] were often welcomed by crews because of the belief that they would eradicate a population of bedbugs. This belief was based on scientific fact, as cockroaches are known as predators of bedbugs" (Monro, 1951). Cockroaches will often "help rid a house of bedbugs by devouring all the little parasites they can capture" (Gaul, 1953). The basis for this belief may have originated with a statement by Webster (1834) who wrote that bedbugs disappeared aboard "H.M. Sloop Chanticleer" when cockroaches made their appearance. Newman (1855) reported the observations of a friend who claimed to have seen a cockroach seize a bedbug in an infested boardinghouse in London. In 1920 Purdy reintroduced cockroaches into a house from which they had been exterminated, in order to control the bedbugs which had become established. According to a popular account by Lillingston (1934) African natives are said to ask sailors for a cockroach or two to be used to hunt bedbugs.

In Siberia, Burr (1926, 1939) found Blattella germanica and bedbugs inhabiting the same room. Mellanby (1939) studied the populations of an animal house in which bedbugs and cockroaches occurred in large numbers; the bugs apparently were not attacked and their numbers increased greatly over a period of a few weeks (Johnson and Mellanby, 1939). Wille (1920) placed starved B. germanica with bedbugs for 20 days, but the cockroaches failed to attack the bugs. In India, captive adults and nymphs of two species of house cockroaches would not touch living bedbugs or their eggs (Cornwall, 1916). In laboratory experiments Gulati (1930) found that Periplaneta americana ate young bedbugs which had soft, blood-filled abdomens; adult bedbugs with harder exoskeletons sometimes were rejected. The maximum number of bedbugs eaten by a cockroach was 3 out of 12 during a period of 48 hours. Johnson and Mellanby (1939), also in laboratory experiments, were unable to show that bedbugs can be controlled by Blatta orientalis or that bedbugs are eaten to any extent by them. The existing evidence indicates that there is little basis for the often repeated statement that cockroaches destroy bedbugs in nature. As Lorando (1929) pointed out, assassin bugs, cockroaches, and red ants can hardly be considered as practical factors in bedbug control, though he did recommend the use of spiders.

According to Martini (1952), cockroaches prey on mosquitoes and sand flies but we have been unable to find any original sources for these statements; the only reference we have found in which cockroaches and Phlebotomus are mentioned together is a paper by Whittingham and Rook (1923); they fed ground-up cockroaches to larvae of Phlebotomus papatasii. Wharton (1951) reported that cockroaches and other predators attacked mosquitoes knocked down by insecticides and affected the number recovered.

Cockroaches will on occasion attack and bite animals other than insects. In an earlier paper (1957a) we discussed about 20 reports of cockroaches biting man. The injury is usually confined to abrasion of the callused portions of hands and feet but may result in small wounds in the softer skin of the face and neck. We failed to include the following reference in the above-mentioned paper. Sonan (1924) had his toes and breast nibbled by cockroaches on Hiyakejima Island during sleep. He had previously learned from a policeman that Periplaneta americana and P. australasiae nibbled people on that island, but he had hardly believed it before he experienced the biting himself.

INTRASPECIES PREDATION

Those who have reared cockroaches in the laboratory have undoubtedly seen cannibalism occur in the cultures. Cannibalism has been observed among the common domiciliary species of cockroaches as well as laboratory colonies of Leucophaea maderae (Scharrer, 1953), and Blaberus craniifer[12] (Saupe, 1928). Edmunds (1957) reported that cannibalism was common in a laboratory colony of Periplaneta brunnea and that egg capsules deposited by a female were often eaten by the other cockroaches.

Periplaneta americana occasionally ate other cockroaches and their oÖthecae and also attacked members of their own species (Lederer, 1952). Griffiths and Tauber (1942) recorded the killing of male American cockroaches by females of the species: "One female was especially vicious and attacked each new male as he was introduced into the container. Most of such males had molted less than 2 days previously. Older males were more capable of defending themselves against attacks of these cannibalistic females." Even though adequate food may be present, females of Periplaneta americana may eat their own eggs (Klein, 1933). Some females may regularly eat their oÖthecae as soon as they are dropped (Griffiths and Tauber, 1942). To be completely eaten an oÖtheca generally must be attacked before it has hardened. If a hole is eaten in one side of the capsule, the cockroach may devour the eggs and leave a portion of the oÖtheca. Frequently only the keel or a part of the keel is eaten and when this occurs the eggs fail to hatch and usually do not complete development because of the rapid loss of water (Roth and Willis, 1955). When adults of P. americana and P. australasiae were deprived of food, both males and females ate newly deposited eggs and, finally, the females ate the males (Sonan, 1924).

Parcoblatta virginica in laboratory cultures also may eat part of its oÖthecae; in this species only the soft end of the recently deposited oÖtheca was eaten (Roth, unpublished data, 1957).

Cros (1942) observed oÖthecae-bearing females of Blatta orientalis attack and kill males of the same species which were attempting to mate; these males were then eaten by the females. Cros also observed injured and recently molted nymphs of B. orientalis to be eaten by others of the same species.

Pettit (1940) noted that cannibalism in his culture of Blattella germanica occurred only when the insects were molting. Adult insects attacked the molting cockroaches more often than did the nymphs. However, nymphs after the fourth instar occasionally set upon other molting nymphs. First-to third-instar nymphs rarely victimized their mates. The victims were all older than third instar; the later stadia were progressively more subject to attack, and molting adults suffered the greatest mortality. No direct correlation was noted between population density and cannibalism.

German cockroaches may attack newly molted nymphs of their own kind and cause them to deflate (Gould and Deay, 1938). LhÉritier (1951) has observed the hatching nymphs of B. germanica being devoured by their congeners even before they have left the oÖtheca.

Nauphoeta cinerea in laboratory cultures will eat newly hatched young of the same species (Roth and Willis, 1954; Willis et al., 1958). In Hawaii, in nature, N. cinerea may kill and eat the cypress cockroach, Diploptera punctata (Illingworth, 1942; Fullaway and Krauss, 1945).

Bunting (1956) stated that species of Neoblattella are omnivorous with carnivorous and cannibalistic tendencies. An adult female Panchlora sp. was killed and eaten by Neoblattella sp. in captivity. A male, provisionally identified as N. celeripes, was killed and partly eaten by two other males of the same species.

The factors influencing the extent of cannibalism among cockroaches are not completely known. According to Wille (1920) hunger was not the cause of cannibalism in Blattella germanica. Wille claimed that the tendency toward cannibalism increased at high temperatures and decreased at low temperatures. Pettit (1940) also noted this effect. Gould and Deay (1938) stated that under crowded laboratory conditions, when there was a scarcity of food, cannibalism among Periplaneta americana was common. The injured cockroaches and those unable to molt were often eaten. Adair (1923) made similar observations. Undoubtedly, conditions of crowding, availability of food, temperature and other factors all influence cannibalism, but practically no experimental work has been done on this subject.

It is interesting, in comparison with the above positive examples of cannibalism, that both Saupe (1928) and Roeser (1940) observed no cannibalism during extensive studies with Pycnoscelus surinamensis. In fact, Roeser stated that there was never a case of cannibalism in spite of long hunger periods imposed on both nymphal and adult insects.