_U.S.A._--Four live and 14 dead specimens from 4 aircraft, Miami (Welch, 1939). One live and 15 dead specimens (adults?) from 16 aircraft; 8 oöthecae from 7 planes; 147 live and 83 dead nymphs from 108 planes, Miami (Denning et al., 1947). Southern U.S. (Hughes, 1949). Five live and 7 dead specimens, Miami (Porter, personal communication, 1958). IV. CLASSIFICATION OF THE ASSOCIATIONS Asano (1937) classified the natural enemies of cockroaches into two types as follows: 1. Enemies that feed mainly on cockroaches (certain ripiphorid beetles and certain chalcid, evaniid, and ampulicid wasps). 2. Organisms which, in their search for food, devour cockroaches that may be encountered (certain species of scorpions, spiders, ticks, centipedes, Strepsiptera, ants, birds, rats, and "parasitic bacteria"). Cameron (1955) arranged the associates of cockroaches in two groups as follows: Group A. Parasites and predators. 1. Parasites: Hymenoptera (Evaniidae, Eulophidae, Eupelmidae, Encyrtidae, Pteromalidae, Cleonymidae) and Coleoptera (Ripiphoridae). 2. Predators: Hymenoptera (Ampulicidae), Hemiptera (Reduviidae), Coleoptera (Dermestidae), Arachnida (Araneae, Acarina). Group B. Parasites and symbionts. 1. Protozoa (including examples of both parasites and commensals). 2. Nematoda (including both primary "parasites" and secondary parasites). 3. Bacteria (including the mutualistic bacteroids). 4. Algae [_Arthromitis_ (= _Hygrocrocis_) _intestinalis_; see p. 124]. Asano's arrangement, although essentially true, is limited; Cameron's system is divided into arthropods (group A) and lower forms (group B), but does not include higher animals. Both attempts at classification need amplification; this we have endeavored to do below. In classifying the biotic associates of cockroaches, we were immediately confronted with a problem in semantics. The concepts parasitism, predatism, and symbiosis have all been used with various shades of meaning by different authors. The problem is not solved merely by accepting as authoritative specific definitions, however apt they seem to be, because, unfortunately, these concepts are not mutually exclusive. For example, among the entomophagous insects, as Sweetman (1936) has pointed out, there can be no definite line of separation between parasitism and predatism: the two intergrade, only the extremes being quite distinct. In fact, Andrewartha and Birch (1954) generalized these relationships by calling both categories predatism. These authors divided natural populations of associated organisms into nonpredators and predators. Although this simplifies their presentation of the general principles of ecology, for our purpose more narrowly defined terms have proved useful.