A Critique on the Phase Theory of Locusts

ion, refers to the characters of concrete individuals in which the effects of phase variation cannot be completely separated from those of other sources of variation. Thus phase variation is superimposed upon variation due to independent environmental and genetic factors affecting the same characters, though usually to a much smaller degree. Further, the degree of mutual stimulation, and hence the extent of phase variation, may itself be modified by environmental and internal factors acting independently of population density.] 4. In a more specific sense, a phase is one of three such types designated respectively solitaria, transiens, and gregaria. The phase solitaria is the type characteristic of populations howing no trace of gregariousness, and whose immediate antecedents also showed no gregariousness. The phase gregaria is best defined as the type characteristic of swarms, i.e., populations showing clearlymarked gregariousness, and behaving substantially as collective units. The phase transiens is the type characteristic of populations with a history of gregariousness intermediate between that of solitaria and gregaria. 5. Non-gregarious grasshoppers, and other nongregarious animals, cannot, therefore, possess a phase gregaria, even when they occur in dense populations; but they may show phase variation. If they show slight gregariousness, without forming definite swarms, they may be conceded a phase transiens. 6. The categories "congregans" and "dissocians" are not phases, because they are not defined on the basis of a difference in mutual stitnulation, population density, or gregariousness, and do not necessarily display distinctive physical or other characters. If one wishes to indicate the direction of phase change, such terms as "gregarizing individuals" and "degregarizing individuals" might be used, although it is necessary to remember that there is never any guarantee that an observed trend in phase transformation will be maintained in the same direction for any length of time. 7. "Phase" is essentially a characteristic of individuals. However, a population may be said to have a phase, in the specific sense of ph. solitaria, transiens, or gregaria, if it is entirely homogeneous for the phase in question. 8. A measure of what may be called the "phase status" of a population may be obtained in the form of means and standard deviations for suitable phase characters of the component individuals, or by a frequency measure. In the case of a morphometric ratio, any such mean must be the mean of the individual ratios, and not the ratio of the totals of the absolute measurements, as recommended by successive International Locust Conferences. The phase status of a popuPHASE THEORY OF LOCUSTS 403 lation may also be estimated by the sexual dimorphism ratio and other measures which, while depending upon individual phase characters, are not themselves measurable in the individual. 9. The degree of correlation between different phase characters, both physical and biological, is low, so that the indications of phase given by different characters may be markedly confficting. This makes it impossible, for example, to draw reliable inferences as to behavior from the physical characters, and has led to general confusion as to what is implied, in any particular instance, by the phase terminology. 10. As a consequence, we are faced with the alternatives of (a) reconciling ourselves to using the phase categories in a vague way to indicate types with ill-defined constellations of characters, or (b) defining the phases of each species in a rigid, if somewhat arbitrary manner, by means of ranges of values of the best morphometric character for differentiating the phases in that particular species. This choice is by no means an easy one, but alternative (b) is to be preferred, since it does provide a completely objective basis for defining the phases by using that physical character, open to quantitative estimation, which is the most intimately associated with the cause of phase differences. Adopting this course, it is clear that only adult insects, with the parts to be measured undamaged, can be classified into phases. Nymphs and damaged adults cannot be identified with certainty. While this is unfortunate, the situation is no different from that in which material of all stages is placed under alternative (a), where, since the phases themselves are not rigidly defined, no specimen or series can be certainly identified. [Since the whole field of locust ecology and epidemiology can be studied without reference to phases, the phase terminology could no doubt be dispensed with altogether if we had only the requirements of this field to consider. But the phases have a taxonomic, as well as a biological significance. Taxonomy is concerned with the full range of physical variation of a species, and well-marked infraspecific forms need to be differentiated and named.] 11. In determining the ranges of the best morphometric haracter for use in defining each phase, the following procedure is desirable: Numerous large series of specimens hould be obtained for a number of different localities and seasons, and representing the full range of gregariousness from sparse non-swarming populations with non-swarming antecedents up to the largest and densest swarms with swarming antecedents. These series should be grouped, purely on their degree and history of gregariousness, as "non-swarming with non-swarming antecedents," swarms," and "intermediate." Series means, group means, and the interseries variance within each group should be calculated for each major morphometric haracter. The best indicator of phase may then be selected as that character which gives the most significant differences between the groups. This best indicator should be a phase character proper, and not, for example, the sexual dimorphism ratio, which cannot be applied to individual insects. Using the series means of the best indicator, the range of means corresponding to each of the three categories of gregariousness hould be determined. Overlap will be found, not only between the ranges for adjacent categories in the gregariousness equence, but sometimes also between the range for non-swarming populations with non-swarming antecedents, and the range for swarms. Arbitrary limiting values should then be selected so as to give three contiguous, non-overlapping ranges, in which a maximum proportion of the series means will fall into the phase to be expected from the degree and history of gregariousness of the series. The phases in the particular species concerned are then rigidly defined by these ranges of the best indicator. The phase status of any series can then be determined as solitaria, transiens, or gregaria on the basis of its mean for the selected character, and any individual can be assigned to its phase on the basis of its individual value for the character. 12. It remains desirable to have some means for indicating the general phase affinities of nymphs, etc., which cannot be definitely classified under the system proposed. Sometimes it may be best simply to describe in a few words what the significant characters of such individuals are. Thus one might speak of "black first-instar hoppers." Something more than this is required to indicate the general ensemble of characters, and the terms "solitarioid," "gregarioid," and "transientoid" might be used for that purpose, not italicized, not accompanied by the word "phase," and necessarily of somewhat uncertain connotation. For example, one might say: "The swarm of yellow, gregarioid adults, of which no specimens were taken, produced typically gregarioid hopper progeny. Stragglers from this band became transientoid by the final instar, and 404 THE QUARTERLY REVIEW OF BIOLOGY gave rise to ph. transiens adults." The same terminology could be used for species in which phase studies are not sufficiently advanced for rigid definition of the phases to have been possible. 13. The "outbreak process" is more readily analyzable into operative processes than into stages in time. The essential process is the "vicious circle of swarm formation," which is a purely behavior phenomenon. The classical physical phase differences arise as secondary consequences of the outbreak process, and play no important part in promoting either further phase transformation or further development of the outbreak process. They are, at most, pointers to the stage that has been reached in the outbreak process, and somewhat unreliable pointers at that. 14. The definition of the phases on the basis of physical characters, coupled with the uncertain correlation between the physical characters and behavior, calls for an independent classification of individuals and populations on a behavior basis. The unambiguous terms "swarming" and "nonswarming" are proposed for this purpose, the latter being synonymous with "solitary-living." The dual classification then permits uch combinations as "swarms of ph. solitaria," "non-swarming gregarioid hoppers," etc. 15. The fundamental cause of outbreaks of locusts is multiplication, and the "irregular periodicity" of outbreaks is primarily due to changes in the population level of the outbreak areas, whose ecological characteristcs are particularly well adapted to permit survival under unfavorable conditions and rapid multiplication under favorable conditions. Several features of the outbreak areas that make them favorable for locust multiplication also promote concentration, and thus set in train the development of gregariousness and transformation i to the ph. gregaria. In the course of these changes migration is probably stimulated; certainly it is made immediately obvious, and practically significant, on account of the gregariousness of the insects and their numbers. 16. The factors of abundance, gregariousness, and migratory capacity are all reflected in the word "swarm," and the essential process that goes on in the