Now if this theory as to the causes of deterioration in disused organs be correct, it follows that rudimentary organs can only occur in species with sexual reproduction, and that they cannot be formed in species which are exclusively reproduced by the parthenogenetic method: for, according to my theory, variability depends upon sexual reproduction, while the deterioration of an organ when disused, no less than its improvement when in use, depends upon variability. There are therefore two reasons which lead us to expect that organs which are no longer used will remain unreduced in species with asexual reproduction: first, because only a very slight degree of hereditary variability can be present, viz. such a degree as was transmitted from the time when sexual reproduction was first abandoned by the ancestors; and, secondly, because even these slight degrees of variability are not combined, or, in other words, because panmixia cannot occur. And the facts seem to point in the direction required by the theory, for superfluous organs do not become rudimentary in parthenogenetic species. For example, as far as my experience goes, the _receptaculum seminis_ does not deteriorate, although it is, of course, altogether functionless when parthenogenesis has become established. I do not attach much importance to the fact that the Psychids and Solenobias—(genera of Lepidoptera which Siebold and Leuckart have shown to include species with parthenogenetic reproduction)—still retain the complete female sexual apparatus, because colonies containing males still occasionally occur in these species. Although the majority of colonies are now purely female, the occasional appearance of males points to the fact that the unisexuality of the majority cannot have been of very long duration. The process of transformation of the species from a bisexual into a unisexual form, only composed of females, is obviously incomplete, and is still in process of development. The case is similar with several species of _Cynipidae_, which reproduce by the parthenogenetic method. In these cases the occurrence of a very small proportion of males is the general rule, and is not confined to single colonies. Thus Adler[195] counted 7 males and 664 females in the common _Cynips_ of the rose. In some Ostracodes, on the other hand, the males appear to be entirely wanting: at least, I have tried in vain for years to discover them in any locality or at any time of the year[196].