In the Navy [Japanese] a seaman who has captured 300 cockroaches will be granted one day special shore leave. They call it "shore leave for cockroaches." The purpose is to promote extermination of cockroaches in a warship because, on the one hand, any warship suffers from numerous cockroaches, and, on the other hand, any seaman likes shore leave.... The formalities for a shore leave for cockroaches are as follows. A seaman keeps cockroaches which he captured (mainly B. germanica, because P. americana and P. australasiae are seldom found in Japan) in a bottle or in a bag until the number reaches 300. Then he brings them to the deck officer to get the confirmation that he has actually captured more than 300 cockroaches. If the deck officer confirms it, the seaman goes to a cabin where a petty officer reports that the

deck officer confirmed the number of cockroaches. The petty officer signs the seaman's name, name of division, rank, and date to be on shore leave in the log book for cockroach shore leaves. The petty officer brings the log book again to the deck officer to get his approval and then goes to the commander for the final approval. In the Navy, they have another special shore leave for rats. In this system, a seaman gets one day shore leave for one rat. The formalities for the latter are the same as for the former, and there is a log book for the rat shore leave in the petty officer's quarters. The author took advantage of these systems frequently.

Sonan (1924)

Little is known of the effects of predatism and parasitism on natural populations of cockroaches. Many statements in the literature are very general; yet there are a few data on egg parasites (e.g., Tetrastichus hagenowii) which suggest that, in the absence of parasites, populations of domestic cockroaches might be much larger than they are in certain areas. We have summarized the literature on natural control and also that on the use by man of predators and parasites in the biological control of cockroaches. However, because of the paucity of information, we have been unable to evaluate the effectiveness of biological control in reducing the numbers of pest cockroaches. This is an area that might reward further investigation.

INVERTEBRATES

Scorpions.—In Puerto Rico, cockroaches are probably the principal food of the scorpions which live in old houses, on tree trunks, etc. (SeÍn, 1923). The staple diet of scorpions in Arizona is the small cockroach commonly known as the water bug (Stahnke, 1949); in the part of Arizona where he resides, Stahnke (personal communication, 1953) says that the "water-bug" is most generally Supella supellectilium although Blattella germanica is also found, but less abundantly.

Spiders.—Jefferys (1760) mentioned a large spider which was protected in the Antilles and especially on Guadeloupe because it hunted down and fed on cockroaches; the spider was reputed to be common in every house. Sir Hans Sloane (1725, in Cowan, 1865) reported that residents of Jamaica kept spiders in their houses to destroy cockroaches. Takahashi (1924) reported that, in the Taihoku area of Formosa, human habitations contained large numbers of spiders which caught and ate cockroaches. Smith (in Marlatt, 1915) reported that Brazilians encourage large house spiders because they tend to rid the house of "other insect pests." In British Guiana tarantulas were kept in a bungalow to control Periplaneta and Pycnoscelus (Beebe, 1925a).

Ants.—A Madam Merian noticed that ants cleared houses of cockroaches (Kirby and Spence, 1822). A small reddish-yellow ant, called PucchuÇiÇi by Peruvian Indians, pursued and destroyed a cockroach called Chilicabra which was a pest in native huts (Tschudi, 1847). Schwabe (1950b) found swarms of ants attacking living Pycnoscelus surinamensis and stated that ants are probably the chief enemy of this cockroach in Hawaii. Wallace (1891) stated that in Africa a band of driver ants may enter a house and clear it of cockroaches and other arthropods. In British Guiana, Beebe, (1925) found that several times a year army ants cleared the laboratory of all cockroaches and tarantulas.

Wasp egg parasites.—Matsumura (1917, in Asano, 1937) proposed that parasitic wasps such as Evania and Brachygaster be protected in Japan as the natural enemies of cockroaches. In one area in France, 20 percent of the oÖthecae of Loboptera decipiens were parasitized by Zeuxevania splendidula (Genieys, 1924). Edmunds (1952a) found that 12 percent of 459 oÖthecae of Parcoblatta collected during December through April of 1950-51 in Ohio were parasitized; evaniids accounted for about 7 percent of the parasitization. Additional collection data in 1951-52 Edmunds (1953a) showed that 8.7 percent of 320 wood-cockroach oÖthecae were parasitized; 2.8 percent of these parasites were evaniids; almost 13 percent of the egg capsules collected showed evidence of previous parasite emergence. Cameron (1957) reported that oÖthecae of Periplaneta americana collected in Saudi Arabia were 29 percent parasitized in March and 25 percent parasitized in October by Evania appendigaster. Sonan (1924) found 1 of 65 oÖthecae of P. americana and P. australasiae parasitized by E. appendigaster in Formosa.

Cottam (1922) stated that the increase of Supella supellectilium in Khartoum was checked by a wasp egg-parasite that was later identified as Anastatus tenuipes (see p. 246) (FerriÈre, 1930, 1935). In this country, this wasp seemed to be effective in decreasing the numbers of Supella in certain areas in Arizona (Flock, 1941).

In Formosa, Tetrastichus hagenowii was an important parasite of cockroach eggs (Maki, 1937). Sonan (1924) reported 30 percent parasitization of 65 oÖthecae of Periplaneta americana and P. australasiae collected in Formosa. In Bangalore, India, the natural parasitization of randomly collected oÖthecae of P. americana varied from 21 percent (of 495 oÖthecae), July 1947-June 1948, and 43 percent (of 288 oÖthecae), July-December 1948, to 57 percent (of 178 oÖthecae), July-October 1949 (Usman, 1949). Cameron (1955) obtained T. hagenowii from oÖthecae collected in Trinidad, B.W.I., and Saudi Arabia; some 15 percent of the oÖthecae of P. americana and P. australasiae collected in October in Trinidad were parasitized; a later collection (March) was 34 percent parasitized; a small sample of P. americana oÖthecae was 65 percent parasitized. The oÖthecae collected in Saudi Arabia in March were 20 percent parasitized. Plank (1947) found that the eggs of the American cockroach in Puerto Rico (probably in laboratory cultures) were so heavily parasitized by T. hagenowii that he had to use P. australasiae for experimental purposes; in 1950 Plank stated that more than 50 percent of American cockroach oÖthecae were parasitized.

Fahringer (1922) stated that Prosevania punctata could be used to eradicate cockroaches, but he did not test his hypothesis. Marlatt (1902) felt that the usefulness of Evania appendigaster in biological control was impaired by Tetrastichus acting as a hyperparasite (see footnote 6, p. 236). However, Wolcott (1951) stated that in Puerto Rico E. appendigaster is quite abundant and is a factor of considerable importance in controlling cockroaches. Kadocsa (1921) stated that Brachygaster minutus and Evania appendigaster were not important in the biological control of cockroaches. These general statements are not supported by experimental evidence.

It is likely that the smaller wasp egg parasites are more effective than the evaniids in controlling cockroaches. Only one evaniid develops in a parasitized oÖtheca, but many individuals of the other wasps develop in one oÖtheca and the number of females that emerge is usually large. However, Cameron (1957) concluded that, with a parasitism rate of 25 to 29 percent and three to four generations a year, against one or less for the host, Evania appendigaster in the areas where it is established is a valuable control agent.

The use of specific egg parasites to control cockroaches has not been attempted extensively. Cros (1942) liberated a species of Tetrastichus (=Eulophus sp.; see p. 254) in his home in Algeria to control the oriental cockroach; as far as we know, he did not report the parasite's effectiveness in reducing the cockroach population. According to Zimmerman (1948) Comperia merceti, when accidentally imported, practically wiped out Supella supellectilium in parts of Hawaii; he claimed to have controlled the brown-banded cockroach in a store building with this parasite. In some parts of Honolulu, almost 100 percent of the oÖthecae of this cockroach were parasitized (Zimmerman, 1944). We (1954b) ran some simulated field tests in which we liberated Tetrastichus hagenowii in rooms artificially seeded with oÖthecae; from 28 to 83 percent of American cockroach oÖthecae and 56 percent of oriental cockroach oÖthecae were parasitized during these tests.

Evania appendigaster was introduced from Hawaii into Canton Island in 1940 against Periplaneta americana, and it has become established (Dumbleton, 1957). This parasite was also successfully introduced into Samoa (Dumbleton, 1957).

Cockroach-hunting wasps.—An earnest attempt has been made to establish in Hawaii wasps that prey on cockroaches. Just how effective these wasps are in controlling cockroaches is still unknown. Dolichurus stantoni was introduced from the Philippines in 1917 and spread to several of the Islands (Swezey, 1920, 1921; Williams, 1944). Bridwell (1920) stated that as a result of this introduction there was a great decrease in cockroaches of the genus "Phyllodromia." A number of Podium haematogastrum from Brazil were liberated in Honolulu (Williams, 1925) but did not become established (Williams, 1928). The effectiveness of Podium was questioned by Williams (1928) who observed that Podium "destroyed innumerable Blattidae, which nonetheless swarmed in their neighborhood, and I must confess from my observations on the various cockroach-hunting wasps that the blattid more than holds its own alongside its enemy."

Introductions of Ampulex have proved more successful. Ampulex canaliculata was introduced into Hawaii from the United States (Williams, 1928a, 1929). Williams also introduced A. compressa into Hawaii in 1940, and the species was reared in large numbers for distribution (Pemberton, 1942). A. compressa has since become established on most of the Islands (Pemberton, 1945a, 1947; Williams, 1946; Van Zwaluwenburg, 1950). The thousand of A. compressa now found in the Hawaiian Islands are all descendants of three wasps captured in Noumea, New Caledonia (Williams, 1944). According to Williams (1941), the number of cockroaches was noticeably reduced at the University of Hawaii poultry farm, where some A. compressa were released. Pemberton (1953) believed that this wasp has become sufficiently abundant to be of definite value. Simmonds (1941) recommended importing A. compressa into Fiji for cockroach control. A. compressa was introduced from Hawaii into Guam in 1954 against Periplaneta americana and into the Cook Islands in 1955 against Periplaneta spp.; it is not yet known whether the parasite became established in either place (Dumbleton, 1957).

VERTEBRATES

... on conserve avec soin les crapauds dans les maisons, et que les dames les tolÈrent, mÊme sous leurs robes, en raison de leurs continuels services, car ils se promÈnent sans cesse À la recherche des Kakerlacs.

Girard (1877)

Toads.—Bufo marinus was first introduced into Puerto Rico from Barbados in 1920 to reduce several major insect pests including cockroaches (Leonard, 1933). It was introduced from Puerto Rico into Hawaii by C. E. Pemberton in 1932 where it rapidly became established; it has since been distributed throughout the Pacific area. B. marinus is one of the world's largest toads; it attains a body length (exclusive of the hind legs) of 7 to 9 inches (Oliver, 1949) and has been kept alive for more than 11 years in captivity (Pemberton, 1945). Alicata (1938) placed giant toads in a fenced area in Hawaii containing an infestation of Pycnoscelus surinamensis; after 24 hours the toads were dissected and each was found to have eaten from 11 to 25 cockroaches. Illingworth (1941) found that 40 to 90 percent of 53 stools of this toad in Hawaii contained remains of P. surinamensis. Alicata (1947) recommended the maintenance of B. marinus in poultry yards to reduce the population of P. surinamensis, the vector of the chicken eye worm.

Toads have also been recommended for controlling cockroaches in houses (Meech, 1889; Sweetman, 1936). Girard (1877) cited a note in a French newspaper which stated that toads were kept in houses in Cuba to control the American cockroach.

Tree frogs.—Tree frogs enclosed in a room overnight were said to effectively clear it of cockroaches (Marlatt, 1915); on sugar plantations in Australia, these amphibians were encouraged in houses and kept as pets because they hunted and devoured large brown cockroaches (Froggatt, 1906).

Birds.—In Guadeloupe, Dutertre (1654) claimed that all the fowls of the country were fond of small cockroaches and lived on practically nothing else. In Hawaii (Zimmerman, 1948) and in the Lesser Antilles (Ballou, 1912) cockroaches are eaten by poultry whenever the birds can find them. In Puerto Rico, Wetmore (1916) stated that owls kept in houses feed extensively on cockroaches; the stomach of one owl which had been kept in a native house was filled entirely with cockroaches. In British Guiana, Beebe (1925) found that cockroaches were eaten by 27 species of birds.

Reptiles.—H. (1800) claimed that two lizards cleared his house of the "true brown cockroach" and suggested that lizards be used for cockroach control because the reptiles are docile and harmless. On Arno Atoll geckos and night-feeding skinks eat large numbers of cockroaches (Usinger and La Rivers, 1953). According to Wolcott (1924) the number of cockroaches eaten by lizards is surprisingly large considering the nocturnal habits of these insects. Beebe (1925a) kept geckos in a bungalow to help control Periplaneta and Pycnoscelus.

Mammalia.—Cowan (1865) stated that in England hedgehogs were often kept domesticated in kitchens to destroy cockroaches. This writer also stated that a lemur was kept on board ship to destroy cockroaches.

Large numbers of the American and Australian cockroaches were eaten by the mongoose in Hawaii (Perkins, 1913).