<em>Introduction</em><br/>The data from the literature concerning insect killing <em>Deuteromycetes</em> are discussed. Stress is laid on many fundamental problems and on gaps in the scientific observations.<p/><em>Nomenclature</em><br/>It appears from a bibliographic survey that <em>Metarrhizium brunneum</em> PETCH, <em>Metarrhizium album</em> PETCH and <em>Metarrhizium anisopliae</em> (METSCHNIKOFF) SOROKIN most probably represent a single species. The conidial lengths divide the fourteen isolated strains into two groups; these conform to the 'forma major' and the 'forma minor' of JOHNSTON (1915). <em/> Because of the supposed absence of original type material of <em>M. anisopliae</em> the author fixed a neotype and deposited it in the Centraal Bureau voor Schimmelcultures at Baarn (The Netherlands).<p/><em>Isolation, development and growth, storing</em><br/>Conidiospores of <em>M. anisopliae,</em> isolated from 14 species of insects, were cultivated on cherry-agar, on the medium according to MARTIN <em>(1950)</em> or on a special isolation medium (VEEN & FERRON, 1966). <em/> In the last medium, actidione (cycloheximide) and chloramphenicol is added in order to prevent infections by bacteria and fungi. The fungus sporulates excellently in surface cultures when the Chapoteaut peptone composition is used as nutrient medium. The mass production of conidiospores has been carried out on boiled sterile rice. The formation of blastospores has been tested in shake cultures according to the Adamek (1967) <em/> method; this has been done with five strains of forma major and five of forma minor. Generally a concentration of 10 <sup>8</sup>-10 <sup>9</sup>blastopores per ml. is obtained after 96 hours; one strain, however, sporulated only poorly.<p/>The strains could be successfully stored at + 4°C for at least 20 months after culturing on egg-yolk coagulated at 80°C. Lyophilization of conidial suspensions in skimmed milk did not affect the germinating capacity of the spores for six months.<p/><em>List of host insects</em><br/>Scrutiny of the literature revealed that at least 204 species of insects are susceptible to <em>M.</em><em>anisopliae</em> under natural conditions. These are mostly <em>Coleoptera</em> of which the terricolous species predominate.<p/><em>Germination and development of germtubes</em><br/>Conidiospores of <em>M.</em><em>anisopliae</em> germinate only in contact with water. Intersegmental folds of the host <em>Schistocerca gregaria</em> FORSK. prevented direct observations, but germination of the conidiospores on exposed parts of the cuticle apparently are not affected by the transpiration of the insect. Epicuticular lipids of fifth stage larvae evoke conidial germination, in contrast to beeswax. The composition of the nutrient medium used for sporulation, may likewise exercise great influence on germination of spores when transferred to an extraction of epicuticular lipids of locusts; the various strains of <em>M.</em><em>anisopliae</em> however behave differently. Both in vivo and in vitro the germinated fungus forms structures analogous to 'appressoria'. But in vitro only rarely penetration tubes originate beneath the 'appressoria'. Also the latter structures are not always essential to the development of penetration tubes.<p/><em>Penetration and infection</em><br/>Histological sections show that the infection of second instar locust larvae by conidia of a race of <em>M. anisopliae</em> isolated from <em>S. gregaria</em> takes place after about 90 hours. The site of penetration of the hyphae is the mouth cavity, mainly the proximal part of the maxillary palps. The ambient relative humidity has no influence on this process.<p/>Penetration of exposed integument, however, occurs only if a water film is present.<p/>Experiments on third instar larvae have shown that the fungus has no preference for specific areas of the integument (as for instance the inter-segmental membranes) for penetration. Any area of the integument can be penetrated. It has been observed that both chemical and mechanical activities play a part in the penetration process.<p/>Phagocytosis of the fungus by haemocytes was never observed.<p/>Infection also takes place when conidiospores are injected in the rectum, but obviously this is of no importance under normal conditions when infection through the mouth prevails.<p/>Probably, blastospores are formed in the places where the penetrating mycelium. comes into contact with circulating haemolymph.<p/><em>The last stage of the disease</em><br/>When the insect dies, the amount of fungus material in the body is often relatively insignificant. Shortly before death the insects often behave abnormally. These observations suggest an action of toxins. Therefore, the toxicity of culture filtrates of ten strains of <em>M. anisopliae</em> was assayed on caterpillars of <em>Galleria mellonella</em> L. and larvae of <em>S.</em><em>gregaria.</em> Considerable differences in activity between the strains were found. However, we were unable to establish a correlation between the production of toxins of the different strains and the degree of virulence for <em>S. gregaria.</em><p/><em>Sporulation</em><br/>When locust larvae are kept in very high relative humidity, bacterial infections with a variable mortality are not uncommon. These unfavorable conditions obviously change the physiological condition of the insects. The early stages of an infection by <em>M. anisopliae</em> are also characterized by a reduced vitality of the host. In many cases blastospores as well as bacteria are found side by side in the haemolymph, even before death. A competition between these microorganisms then occurs after death.<p/>Under these conditions the sporulation of the fungus is inhibited, though to a different degree in different strains. This is also confirmed by the observation that bacterial colonies mainly <em>(Aeromonas sp.)</em> may locally suppress the growth of the fungus on a Sabouraud culture medium in Petri dishes. When few bacteria are present, sporulation on insect bodies only takes place in a high relative humidity. A relative humidity of 93 % is already below the limit necessary for sporulation.<p/><em>Concentration effects</em><br/>Experiments with spores in quantities sufficient to kill second instar hoppers have shown that considerable differences in activity exist within the twelve strains studied. The LD <sub>50</sub> of a strain isolated from <em>Polyphylla olivieri</em> CAST. has been determined. The inclination of the regression line representing the relation between the logarithm of the dose and the mortality in probits is fairly large and resembles the values obtained from the insect killing bacterium <em>Bacillus thuringiensis</em> BERL. In similar experiments the food of the hoppers was sprayed with a suspension of spores. These experiments have been performed on first, second and third instar larvae. It appears that the susceptibility of the hoppers to the fungus decreases only slightly during aging, but the incubation period becomes longer in older larvae. Dosage experiments may also be performed <em>per os</em> with blastospores from shake cultures.<p/>CONCLUSIONS<p/>In the past the use of insect killing materials, containing <em>M. anisopliae</em> gave only rather unsatisfactory results. However, from the present investigation it appears that the fungus possesses some features which may render it useful to control the desert locust especially larvae of the first and second instar.<p/>As it has been established that infection in this species may occur via the mouth, it is possible to induce a fungal infection even in air of a very low humidity, as is often found in the habitat of this insect species. The chance of inducing an epizootic, however, seems to be limited because sporulation of the fungus, essential for a spreading of the disease, requests a very high relative humidity.<p/>To determine the possibilities of oral infection in other insect species, it is necessary to study the conditions under which <em>M. anisopliae</em> becomes infectious. Also the doses of spores which are needed to cause high mortality within a short time should be determined. Only after this has been done it will be possible to derive the conditions for practical application of this biological control agent.<p/>Of course technical factors and production costs will be decisive for its eventual use.<p/>Fundamental research is needed in order to select highly virulent strains and to maintain their virulence or to increase it by physical or chemical methods. It is also essential to investigate the specificity of the various strains to harmful and useful insects, as well as their possible effects on vertebrates, though the latter seems unlikely in view of the existing literature.<p/>Technical possibilities and the economics of mass-production will finally be decisive for the use.