Irritating or repellent secretions provide many animals belonging to widely unrelated groups with a more or less potent means of defence....

It will be seen that this method of defence does not rest merely upon a passive unpalatable attribute, but upon an active emission of the unpalatable substance which, since it occurs when the animal is seized or threatened by an enemy, enforces its effectiveness. In its highest development we find different forms whose specialized habits and modified structure enables them to project secretion at the enemy, and thus to discourage attack.

Cott (1940)

There are very few records indicating that cockroaches are unaccepted as food by other animals. Hutson (1943) found that the duck, guinea fowl, and pigeon would not normally eat Pycnoscelus surinamensis, and in his experiments with the chicken eye worm he had to force-feed his birds with infected cockroaches. Lederer (1952) found that insectivorous birds in the Zoological Garden, Frankfurt am Main, either refused hardened (as opposed to teneral) American cockroaches or ate them unwillingly. Carpenter (1925) reported that a monkey (Cercopithecus) failed to feed on cockroaches and suggested that the insects' odor made them repugnant; however, there are a number of positive records of monkeys feeding on cockroaches (see pp. 284-286).

Cockroaches may escape capture by predators through evasive behavior, concealment, protective coloration, mimicry, or secretion of malodorous materials. Nocturnal cockroaches may avoid predators that are active during the day (Crawford, 1934), but nocturnal predators are apparently quite successful in capturing cockroaches. Some cockroaches may be protected by their swiftness, others by their resemblance to vegetation (Williams, 1928). The habit of squeezing into narrow cracks may afford cockroaches some protection.

Burrowing forms such as Pycnoscelus may spend much time in underground cells (Roeser, 1940). Polyphagids rapidly burrow into sand (Fausek, 1906), where they may be protected from predators. Tepper (1893) discovered that a very large Australian cockroach, Geoscapheus robustus, had its fore legs, especially the tibiae, adapted for digging. He observed this species in captivity and in 1894 reported that it appeared to sink into the soil without raising any considerable amount above the surface and that it did not form an unobstructed tunnel. Another large Australian cockroach, Macropanesthia rhinocerus, burrows about two feet below the surface of sandy soil; it also makes nests among pine roots and the nymphs rarely appear above ground (Henson in Day, 1950). Tepper (1893) observed that Australian cockroaches of the genera Epilampra and Oniscosoma buried themselves in loose soil and dust. Baker (in Rehn, 1930) observed that Styphon bakeri is found in humus and rubble in the Dutch West Indies where "It is sluggish in the open, but wedges into the humus quite quickly."

Therea nuptialis, found in India, conceals itself at the roots of fig trees, etc. The small hairs on its elytra retain sufficient dust to conceal it, or at any rate to render it inconspicuous, when not on the wing (Annandale, in Chopard, 1924c). Rehn and Hebard (1914) observed that the nymphs of Blaberus craniifer[13] at Key West, Fla., "were usually found half buried in loose damp earth under boards, where they remained motionless, looking much like lumps of earth (with which they were usually much dusted) until disturbed." Hebard (1917) reported of Monastria biguttata from Brazil that "All of the juveniles are heavily coated with foreign particles" which adhere "to a multitude of closely placed, minute and usually curved spines, which cover the dorsal surface and marginal portions of the ventral surface."

It is apparent from the numbers of predators reported herein that many animals are not deterred by the odorous secretions of cockroaches; these secretions, because they may seem repugnant to man, are often claimed to be repellent to predators. However, Cott (1940) points out that "There are many instances in which protective devices and associated warning colours are known to be ineffectual against certain enemies. But this does not necessarily imply that they are not on the whole beneficial to the species attacked." Certain cockroach secretions may well be repellent to many predators, but as this is a purely negative aspect of the predator-prey relationship little thus far has been observed or published. Potential prey that successfully defends itself against attack is never found in a predator's stomach.

Cockroaches have a variety of glands which secrete odorous materials. Certain secretions, produced by tergal or dorsal glands in males, are involved in sexual behavior; the females feed on the secretion from these glands prior to copulating (Roth and Willis, 1954). However, other secretions which are produced by both sexes are ejected or given off when the insect is disturbed; undoubtedly these are defensive weapons that are used against predators. Very few experiments or observations are on record to show how effective these secretions may be in protecting the cockroach. Although the morphology of some of the glands has been described, relatively little is known about the chemistry of their secretions.

Many species of Australian cockroaches have been reported to emit "disgusting" odors, though the glands producing these secretions have not been described, nor is the chemistry of the compounds known. Cosmozosteria lateralis exposed two orange-red spots on the abdomen while emitting a pungent odor which deterred a collector from capturing it (Shelford, 1912). Another Australian species, Platyzosteria castanea, when disturbed on barren ground tilts forward on the vertex and straddles out the posterior legs, supporting itself in a vertical position on the head and tarsi; in assuming this attitude it will squirt a foetid fluid as a fine spray for a distance of 6 or 7 inches (Shaw, 1914). Spencer (1892) mentions the pungent odor given off by a cockroach which had been accidentally cut in two. Rageau (1956) stated that in the New Hebrides and New Caledonia Cutilia nitida emits, when disturbed, a corrosive liquid with an extremely disagreeable odor.

The adults of Eurycotis floridana emit an odorous fluid when seized (Rehn and Hebard, 1905). The fluid, which may irritate sensitive skin areas, may be ejected as a spray for a distance of several inches. This secretion has been identified as 2-hexenal (Roth et al., 1956), and the ventral abdominal glands which produce it have been described (Stay, 1957). Eisner (personal communication, 1958) has found that the toad Bufo marinus and the frog Rana pipiens invariably spit out adults of E. floridana which they have seized. The odor of 2-hexenal was strongly apparent after these attacks, and the insect was never damaged. However, the lizard Anolis equestris seized and crushed E. floridana before releasing its hold and dropping the insect 5 to 10 minutes later. The blue jay Cyanocitta cristata readily attacked adults of E. floridana and killed them but did not eat the insects until after the odor had dissipated; however, the bird carried nymphs of E. floridana to its perch and ate them. Nymphs of this species do not secrete 2-hexenal (Roth et al., 1956). Recently, 2-hexenal has been tested for its antibacterial activity and has been found to be active against seven species of pathogenic bacteria (Valcurone and Baggini, 1957). Eurycotis decipiens from Trinidad also ejects a fluid which may produce toxic symptoms such as vertigo and nausea (Bunting in Roth and Willis, 1957a).

Large reservoirs of glands similar in appearance and position to those of Eurycotis floridana are present in the adults of both sexes of Neostylopyga rhombifolia and Platyzosteria novae seelandiae. Walker (1904) and Longstaff (in Shelford, 1912) noted that the latter species had a strong odor. Roth (unpublished data, 1957) found that the secretion of P. novae seelandiae when ejected is grayish or milky in color. In the reservoirs of the ventral gland of this insect the secretion is a milky liquid containing floating greenish globules. Both infrared and mass spectrographic analyses show that the secretion is a mixture containing 2-hexenal, the aldehyde that is found in E. floridana. Eisner (personal communication, 1958) observed that the lizard Anolis carolinensis immediately released Neostylopyga rhombifolia without injury, but that Bufo marinus, Anolis equistris, and Cyanocitta cristata ate the insect despite the secretion; several unidentified spiders and the ant Pogonomyrmex badius were not repelled by the secretion of N. rhombifolia.

Dorsal and ventral glands have been found in both sexes of Blatta orientalis and Periplaneta americana (Minchin, 1888, 1890; Kul'vets, 1898; Oettinger, 1906; Harrison, 1906; Liang, 1956). The ventral glands are found in the same general region as those of Eurycotis. We have also found similar ventrally located glands in both Periplaneta australasiae, and P. brunnea. The reservoirs which store the secretion of the ventral glands are smaller in Blatta and Periplaneta spp. than those found in Eurycotis, Neostylopyga, or Platyzosteria.

In Blatta orientalis the dorsal glands can be everted by pressure on the abdomen; the secretion in these glands, according to Haase (1889), has the typical oriental cockroach odor. Although the dorsal glands of the oriental cockroach are usually given a defensive role (Haase, 1889, 1889a; Kul'vets, 1898; Oettinger, 1906; Koncek, 1924), the functions of secretions of these nonepigamic dorsal glands and the ventral glands are still open to question. It is possible that some of the odors produced by cockroaches have functions other than defense or sex attraction. For example, Ledoux (1945) showed that the species odor is largely responsible for the gregarious behavior shown by Blatta orientalis and Blattella germanica. The olfactory stimulus acts over a short distance only, and the source of this odor in the insect is unknown. By washing Blattella germanica in warm chloroform Dusham (1918) extracted a wax which had the odor of the German cockroach. However, there is no evidence to show that cockroaches respond to the same cockroach odors that are detected by man.

Certain cockroaches have recently been found to have odorous secretions which are produced in tracheal glands. In Diploptera punctata the tracheae leading to the second abdominal spiracles of nymphs and adults are modified into odoriferous glands which produce a mixture of 2-ethyl-1,4-benzoquinone; 2-methyl-1,4-benzoquinone; and para benzoquinone; this material is ejected as a means of defense. The offensive odor emitted by adults and nymphs of Leucophaea maderae also issues from the second abdominal spiracles (Roth and Stay, 1958).

Diploptera is capable of ejecting its quinones from either its right or left tracheal gland according to which side of the insect is attacked (pl. 36, A-B). Eisner (1958) found that the secretion repelled the ant Pogonomyrmex badius (Latreille) (pl. 36, C) and the beetle Galerita janus Fabricius when they attacked the cockroach. The spider Lycosa helluo Walckenaer was repelled by large nymphs and adults of D. punctata but young nymphs were usually eaten promptly (Eisner, 1958).

Bordas (1901, 1908) believed that the "conglobate" gland (Miall and Denny, 1886), found in males of Periplaneta americana and Blatta orientalis, was an odoriferous gland used for defense, but Gupta (1947) has shown that in all probability this gland (the phallic gland) secretes the outermost covering of the spermatophore.

What appears to be mimicry occurs in some species of Blattaria. The nymphs of many Panchlorini and Blaberinae vaguely resemble sow bugs (Chopard, 1938). Certain members of the Perisphaerini (e.g., Perisphaerus glomeriformis) from the Malayan region which resemble sow bugs (Annandale, 1900; Hanitsch, 1915) can roll themselves up into a ball thus hiding their antennae and legs (Lucas, 1862). Although these cockroaches are found among dead leaves or under stones, in places in which sow bugs are also found, the benefit to either or both forms is questionable; Annandale (1900) believed that the crustacean and the cockroach, living under similar conditions, developed the same general body shape. Rolling up into a ball is nothing more than an exaggeration of a reflex common to many young cockroaches, that is, an arched position which these insects assume when they immobilize themselves in response to certain stimuli (Chopard, 1938).

There are cockroaches that resemble various Coleoptera and Hemiptera (Belt, 1874; Shelford, 1912; Hanitsch, 1915). Some look like cerambycids, lampyrids, coccinellids, pentatomids, etc. Perhaps the most striking examples are the resemblances of cockroaches in the genus Prosoplecta of the Epilamprinae to beetles of the family Coccinellidae; Shelford (1912) has figured a number of species of Prosoplecta together with the species of beetles which they seem to have taken for models. Williams (1928) mentioned diurnal cockroaches which by a combination of markings, shape, posture, and active flight about vegetation suggest certain wasps.

Unfortunately, practically nothing is known about the behavior of these so-called mimics and models or their relationships with predators in the field. For the most part, the examples are based on a comparison of pinned insects from museum collections (Burr, 1899); for this reason Chopard (1938) believed that not much value should be placed on superficial resemblances of this kind. However, we believe that a lack of knowledge of cockroach mimicry is not a valid reason for rejecting the idea that mimicry, if it occurs, may be of some benefit in the survival of mimetic species. Certainly Cott's (1940) voluminous compilation of the literature on adaptive coloration should make the most skeptic hesitate to conclude dogmatically that these instances of mimicry are merely accidental and meaningless.