Most species of the aphid genus Pemphigus migrate between Populus spp. (primary hostplants) and the roots of herbaceous dicotyledons (secondary hostplants), and only 4 species (spyrothecae, monophagus, mongolicus & siphunculatus) are known to be non-migratory on the primary host. We proposed that the non-migratory life cycles of these Pemphigus species have evolved through generation packing (GP): in the migratory ancestor of a non-migratory Pemphigus species a mutation that caused alate .emigrants to Iarviposit in the gall (instead of on the secondary host) occurred and was fixed, and the "emigrants" were later transformed into apterae. Morphological data for P. spyrothecae (Aoki & Kurosu 1986) and P. monophagus (Aoki & Kurosu 1988a) accord well with what the GP hypothesis predicts.As a result of GP, 2 types of (lst-instar) larvae coexist in the gall of a non-migratory species. Larvae of one type are original gall inhabitants, and those of the other type had once lived on the secondary host, and were packed into the gall. Since, in P. spyrotheeae (and probably also inP. monophagus), only lst-instar larvae of the former play a defensive role, we called them 'soldiers.' Unlike soldiers of CoIophina and cerataphidines (see e.g. Aoki 1987), they are not completely sterile. It6 (1987) attacked this particular usage of 'soldiers' by us.We are not sure that discussions about usage of a word are "basically fertile" (cf. Popper 1945). As is often said (e.g.,by Dawkins 1986), words are our servants, not our masters. For different purposes it is convenient to use words in different senses. To avoid the misunderstanding that we are Humpty Dumpties (cf. Sober 1984, pp. 70-71), we will explain one purpose of our 1986 paper briefly.It is assumed to be a general model of evolution that new behavior develops first, then structural reinforcements follow (e.g., see Mayr 1970, p. 363).One of us (Aoki 1987) thought that pseudoscorpion-like soldiers of cerataphidines on the secondary host have evolved in this way.First, division of labor occurred in the 1st instar: some lst-instar larvae made a speciality of colony defence.Later, they were transformed into bizarre pseudoscorpion-like soldiers.In our 1986 paper, on the other hand, we tried to show that soldiers can evolve in a different way: in P. spyrothecae, basic morphological differentiation between the 2 types of larvae had already been completed before they coexisted in the gall. For the purpose of contrasting the 2 ways of caste differentiation, we found it convenient to call the thick-legged lst-instar larvae (i.e., defenders) of P. spyrothecae 'soldiers.' They may not be a perfect example of aphid soldiers, as It6 pointed out.However, if our story of P.spyrothecae is accepted, it will be easy to imagine how a species with perfect soldiers can be derived from a spyrothecae-like one. Suppose, for instance, that the fundatrix of P. spyrothecae or P. monophagus gained the ability to produce larvae of the 2 types by a generation slip, as the aptera of P. spyrothecae actually di...