Blochmann (1887, 1888) discovered intracellular particles (the bacteroids or symbiotes of authors) that resembled bacteria in the fat body of males and females of Blatta orientalis and Blattella germanica (pl. 26), in the ova of these insects, and in their embryos. Bacteroids have since been found in at least 25 species and 19 genera of cockroaches. Presumably such microorganisms are universally distributed throughout the Blattaria. General reviews of the bacteroids of cockroaches and other insects have been published by Glaser (1930b), Schwartz (1935), Steinhaus (1946, 1949), Buchner (1952, 1953), Brooks (1954), and Richards and Brooks (1958). The reader is referred to these papers, and those of authors cited in the list at the end of this section, for discussions of the morphology of the bacteroids, their distribution within the host, and attempts to culture them in vitro.

It has long been assumed, without proof, that cockroaches and their bacteroids form a mutually beneficial association. As it has not been possible to cultivate bacteroids apart from their cockroach hosts, it may be assumed that the host is essential to the continued existence of the microorganism, which also derives from the association other obvious benefits as well. Experiments to show that the host also benefits from the association have centered around rendering cockroaches bacteroid free. Starvation, parasites, electromagnetic radiation, heat or cold, or chemicals have all been used in attempts to eliminate the bacteroids. Of these, only chemical treatment has provided a satisfactory technique.

Few chemicals other than antibiotics have proved to be useful in the elimination or reduction of bacteroids. Yetwin (1932) injected various dilutions of 22 compounds into Blattella germanica. He observed decreases in the bacteroids of the fat body only following injection of methylene blue, but did not pursue this lead further. Gier (in Steinhaus, 1946) observed reduction in the numbers of bacteroids after cockroaches were injected with crystal violet, hexylresorcinol, or metaphen. Bode (1936) reported that injection of irritants such as lithium salts or quinine hydrochloride had no apparent effect on the symbiotes.

Brooks (1957) reared Blattella germanica on diets containing different concentrations of inorganic ions. On a manganese-deficient diet the cockroaches grew poorly and some of their progeny lacked normal bacteroids; about 10 percent of the aposymbiotic generation grew and reproduced on a diet fortified with yeast. Varying the concentrations of other salts in the diets gave results in which the progeny were either aposymbiotic or the fat body was abnormal but the mycetocytes were abundant; all these cockroaches soon died even on fortified diets.

The administration of certain antibiotic drugs has produced cockroaches very nearly free of bacteroids. Brues and Dunn (1945) found that although sulfa drugs had no effect on the bacteroids, penicillin in large doses reduced the number of bacteroids in Blaberus craniifer, but the cockroaches died within a few days. Death was attributed to lack of bacteroids rather than effect of the drug; this is perhaps an unwarranted conclusion in view of the survival of aposymbiotic cockroaches that have been produced by other drugs (Brooks, 1954; Brooks and Richards, 1955). Glaser (1946) found that the bacteroids of Periplaneta americana could be adversely affected or destroyed by sulfathiazole, or sodium or calcium penicillin. Fraenkel (1952) questioned the conclusions of Brues and Dunn (1945) and of Glaser (1946) because of the high mortality in their experiments, and he suggested that the described phenomena "were due rather to direct toxic effects on the host than to loss of the symbionts." Noland (in Brooks, 1954; Brooks and Richards, 1955) confirmed Glaser's results with penicillin and extended sulfa treatments to include Blattella germanica. Every female whose bacteroids were reduced to the vanishing point resorbed her ovaries and was incapable of reproduction. Brooks (1954; Brooks and Richards, 1955) found that administration of several antibiotics did not eliminate the bacteroids from the fat body of B. germanica unless the dose was so high that it caused excessive mortality. Frank (1955, 1956) was able to eliminate bacteroids from Blatta orientalis by injecting or feeding chlortetracycline, oxytetracycline, or penicillin; survival of treated insects was not good and reproduction was poor; the aposymbiotic individuals were smaller than normal. As Richards and Brooks (1958) have pointed out, it is uncertain how much of this difference was the result of loss of bacteroids and how much the effect of the drug. It is obvious that in all these experiments the action of the drugs on the bacteroids was accompanied by equivocal side effects which confused interpretation of the results. The effect on the cockroach of a loss of bacteroids cannot be separated from a possible toxic effect of the drug.

Fortunately Brooks (1954; Brooks and Richards, 1955) obtained completely aposymbiotic offspring from Blattella germanica that had been reared on aureomycin. These bacteroid-free nymphs were practically incapable of growth on a natural diet that was adequate for nymphs with symbiotes. However, the addition of large amounts of dried yeast to the diet enabled aposymbiotic nymphs to mature in two to three times the period required by normal nymphs. Final proof of the function of the bacteroids was obtained by reestablishing them in aposymbiotic cockroaches. The insects that received implants of normal fat body of B. germanica showed a slow, steady gain in weight over the controls (Brooks, 1954; Brooks and Richards, 1956). Obviously, the intracellular symbiotes subserve the normal nutrition of the cockroach. Whether the bacteroids produce only vitaminlike substances, as suggested by Keller (1950), or function in some other way is still to be determined. Brooks (1954) concluded that the amount of vitamin-containing food required for increased growth by aposymbiotic cockroaches is much greater than the known vitamin requirements; hence the factor(s) needed is unknown and present in low concentration, or it serves as a precursor of a second factor(s) whose synthesis is aided by the bacteroids. Brooks (in Richards and Brooks, 1958) has since found that the bacteroids of Blattella germanica "can supply the insect with B vitamins, amino acids and some larger protein fragment."

Gier (1947) stated that the symbiotes of cockroaches are generically all the same. However, as the symbiotes are presumed to have been associated with cockroaches for over 300,000,000 years (Buchner, 1952) they may be assumed to have developed specific differences that link them inseparably to their respective hosts. Ries (1932) transplanted symbiote-containing fat body from Blatta orientalis into the mealworm and larva of Ephestia kÜhniella, and from Blattella germanica and Stegobium paniceum (=Sitodrepa panicea) into B. orientalis. The implants did not become established in the new host, although most of the transplantations were successful in that the hosts survived and the implants remained intact for some time before they were encapsulated by host tissue. Brooks (1954; Brooks and Richards, 1956) transplanted fat body of Periplaneta americana and B. orientalis into aposymbiotic B. germanica. The growth of the cockroaches injected with foreign tissue was not different from that of aposymbiotic controls. Sections of host insects did not contain mycetocytes and no bacteroids were found. Haller (1955a) injected bacteroids or implanted mycetocytes of B. germanica into gryllids, acridids, and locustids. These implants and innoculations were rapidly destroyed by the hosts. But as Richards and Brooks (1958) have pointed out, none of these experiments provide information about the specificity of the bacteroids themselves.

COCKROACHES IN WHICH BACTEROIDS HAVE BEEN FOUND

Bantua stigmosa. Fraenkel (1921)

Blaberus craniifer. Brooks (1954); Brues and Dunn (1945); Hoover (1945).

Blaberus giganteus. Blochmann (1892).

Blatta orientalis. Blochmann (1887, 1888, 1892); Bode (1936); Brooks (1954); Buchner (1912); CuÉnot (1896); Fraenkel (1921); Frank (1955, 1956); Gier (1936, 1947); Glaser (1920); Gropengiesser (1925); Gubler (1948); Heymons (1895); Hollande and Favre (1931); Hoover (1945); Hovasse (1930); Javelly (1914); Keller (1950); Koch (1949); Menel (1907)?; Mercier (1906a, 1907, 1907b, 1907c); Ries (1932); Ronzoni (1949); Tacchini (1946); Wolf (1924, 1924a).

Blattella germanica. Blochmann (1887, 1888, 1892); Bode (1936); Borghese (1946, 1948, 1948a); Brooks (1954); Brooks and Richards (1954, 1955, 1955a, 1956); Fraenkel (1921); Gier (1936, 1947); Glaser (1920, 1930a); Gropengiesser (1925); Haller (1955, 1955a); Heymons (1892, 1895); Hoover (1945); Koch (1949); Lwoff (1923); Milovidov (1928); Neukomm (1927, 1927a, 1932); PÉrez Silva (1954, 1954a); Ries (1932); Rizki (1954); Ronzoni (1949); Tacchini (1946); Wollman (1926); Yetwin (1932, 1953).

Cryptocercus punctulatus. Gier (1936, 1947); Hoover (1945).

Derocalymma cruralis. Fraenkel (1921).

Ectobius lapponicus. Heymons (1892); CuÉnot (1896); Koch (1949).

Ectobius pallidus. Heymons (1892, 1895); CuÉnot (1896).

Epilampra grisea. Fraenkel (1921).

Eurycotis floridana. Gier (1936, 1947).

Loboptera decipiens. Buchner (1930).

Nauphoeta cinerea. Fraenkel (1921).

Nyctibora noctivaga. Gier (1947).

Parcoblatta lata. Gier (1936, 1947).

Parcoblatta pensylvanica. Brooks (1954); Gier (1936, 1947).

Parcoblatta uhleriana. Gier (1936, 1947); Hoover (1945).

Parcoblatta virginica. Glaser (1920); Gier (1947).

Periplaneta americana. Baudisch (1956); Bode (1936); CuÉnot (1896); Gier (1936, 1937, 1947); Glaser (1920, 1930, 1946); Gubler (1948); Hertig (1921); Hoover (1945); Ketchel and Williams (1953).

Periplaneta australasiae. Gier (1936, 1947); Koch (1949).

Platyzosteria armata. Fraenkel (1921).

Polyphaga aegyptiaca. Fraenkel (1921).

Pseudoderopeltis aethiopica. Fraenkel (1921).

Pycnoscelus surinamensis. Bode (1936); Koch (1949).

Supella supellectilium. Brooks (1954).

BACTERIA

Evidence showing that intestinal bacteria contribute to the nutrition of cockroaches is meager. Cleveland et al. (1934) isolated a bacterial organism from the foregut of the wood-feeding cockroach Panesthia angustipennis. The bacterium digested cellulose rapidly in vitro and these workers believe that this cockroach and other related wood-feeding species are dependent on symbiotic bacteria for the digestion of their food.

Mencl (1907) described cell nuclei in "symbiotic," not closely defined types of bacilli that he found in abundance in the digestive tract of the KÜchenschabe, Periplaneta (presumably Blatta orientalis). Unfortunately, he was more concerned about the morphology of the bacteria than the stated mutualistic relationship, so nothing is known of their physiology.

The growth rates of Periplaneta americana and Blattella germanica were retarded when the insects were reared aseptically, which suggests that microorganisms normally found in the digestive tract supply certain necessary dietary constituents (Gier, 1947a; House, 1949). Noland et al. (1949) suggested that microorganisms in the digestive tract of B. germanica synthesized riboflavin since the nymphs reared on a low riboflavin diet accumulated more of the vitamin than could have been ingested in the diet. However, Metcalf and Patton (1942) found little or no bacterial synthesis of riboflavin in P. americana. Noland and Baumann (1951) suggested that methionine, one of the amino acids essential for rapid growth of B. germanica, was synthesized by intestinal microorganisms in the insects.

PROTOZOA

It is probable that with few exceptions protozoa found in the digestive tract are not necessary for survival of the cockroach. However, very few experiments have been performed to determine the importance, if any, of these microorganisms to the host. Cleveland (1925) removed the protozoa from the cockroach (possibly Periplaneta americana) by oxygenation at 3.5 atmospheres. The ciliates Nyctotherus and Balantidium, flagellates Lophomonas and Polymastix, the amoeba Endamoeba blattae, and three unidentified protozoa were killed by this treatment, yet the insects lived normally after defaunation.

Armer (1944) studied the effects of high-carbohydrate, high-fat, and high-protein diets, as well as starvation, on the intestinal protozoa (Nyctotherus ovalis, Endamoeba blattae, Endolimax blattae, Lophomonas striata, and Lophomonas blattarum) in Periplaneta americana. Starvation of the host lowered the incidence or eliminated most of the protozoa, but a high-carbohydrate diet maintained them at a relatively high level. Lophomonas blattarum was eliminated by a high-protein diet, and practically eliminated by a high-fat diet. Lophomonas striata was eliminated from some hosts that were kept on high-fat and high-protein diets. Endamoeba blattae showed a decrease in infection rate when the cockroaches were maintained on high-fat and high-protein diets. The effects of diets on Endolimax blattae were not uniform.

It has been shown by Cleveland (1930, 1948) and Cleveland et al. (1931, 1934) that the wood-feeding cockroach Cryptocercus punctulatus depends upon certain intestinal protozoa for survival; these protozoa utilize as food the wood ingested by this cockroach. The wood is broken down into compounds the cockroach can utilize by the protozoa which elaborate a cellulase and possibly a cellobiase (Trager, 1932). Only molting nymphs of Cryptocercus can pass the protozoa on to the newly hatched young, so that molting and hatching must happen concurrently each year or the young die.

The sexual cycles in species of protozoa in the genera Trichonympha, Saccinobaculus, Oxymonas, Monocercomonoides, Hexamita, Eucomonympha, Leptospironympha, Urinympha, Rhynchonympha, Macrospironympha, and Barbulanympha (fig. 3, B) are induced by hormones produced by Cryptocercus only during its molting period (Cleveland, 1931, 1947, 1947a, 1949-1956a). Perhaps the prothoracic gland hormone of the host may be responsible for initiation of the flagellate sexual cycles (Cleveland and Nutting, 1955). The protozoan sexual cycles may be used as indicators of the onset of molting in Cryptocercus; thus different species of protozoa begin their sexual cycles from 35 days before to 2 days after molting of the cockroach (Cleveland and Nutting, 1954). Hollande (1952) and GrassÉ (1952) have reviewed the roles and the evolution of the flagellates in Cryptocercus and in termites.

The protozoa of cockroaches and termites are clues to the relationship between these two groups of insects. Kirby (1927) pointed out similarities between Endamoeba blattae of Periplaneta and the amoebae of the termite Mirotermes, suggesting that these protozoans were probably derived from an amoeba in an ancestor common to both blattid and termite. Kirby (1932, in Kidder, 1937) found a species of Nyctotherus in Amitermes that resembles Nyctotherus ovalis from domestic cockroaches. The belief that the termites and cockroaches had a common origin is also strengthened by the similarities between the hypermastigotes of both Cryptocercus and termites (Cleveland et al., 1934).

The cockroaches Cryptocercus and Panesthia both feed on wood, but the protozoa found in Panesthia resemble more closely the species in domestic cockroaches than those in Cryptocercus. The Clevelandellidae (from Panesthia) are closely related to Nyctotherus and have probably evolved from common ancestors. However, the separation of the Clevelandellidae from Nyctotherus must have taken place at a later date than the divergence of their hosts, otherwise representatives of that family would probably also be found in Periplaneta and Blatta (Kidder, 1937).

The protozoa of Cryptocercus can be transferred from one individual to another (Nutting and Cleveland, 1954). They can also be transferred to the termite Zootermopsis where they survive only until the host molts; the reverse is also true, Zootermopsis Protozoa can survive in Cryptocercus until the cockroach molts (Nutting and Cleveland, 1954a).