early stages of Latiorina orbitulus. 155 
ancestral species, which must present characters that are 
primitive in several directions, and do not belong to one 
specialised group. The other difficulty is even stronger. 
There are other sections of the Lycaenids which possess 
both myrmecophilous and amyrmecophilous species, and 
their amyrmecophilous species would be equally entitled to 
claim ancestral rank. To keep within Palaearctic species, 
the Theclas are largely amyrmecophilous, yet 7. w-albwm 
has a somewhat ill-developed honey-gland, whilst Neoly- 
caena, Thestor and Callophrys have well-developed glands. 
If then these amyrmecophilous species never had a 
gland, we are driven to believe that these glands originated 
and developed, absolutely de novo, in half-a-dozen, perhaps 
a score or more instances. That such a curious and special 
organ, not found anywhere else, should appear so many 
times over in one group, always with essentially the same 
structure and always in the same place, and accompanied 
in nearly all cases by the remarkable fans of the following 
segment, is altogether in opposition to what we know of 
evolutionary processes, and involves too large a draft on 
our credulity. 
We are driven then to suppose that all the groups, 
some of whose species possess these special organs, have a 
common ancestry, and that those genera and species within 
these groups that are without them have lost them. 
How, in this case, can the fact be accounted for that in 
the species before us, and no doubt in others (certainly in 
the Theclids), the segment that usually carries the gland 
shows no trace of ever having possessed it ? 
When the honey-gland is present, the special long hairs 
(tubercles I and II) of this segment are wanting. There 
is, indeed, reason to believe that the honey-gland results 
from a modification of these particular tubercles. The 
surrounding hairs are also specially modified. In J. 
orbitulus the armature of hairs on the seventh abdominal 
segment takes its place as entirely in accord with that of 
the other segments, and has just the arrangement we 
would expect on the hypothesis that it was an ancestral 
form that had never had a honey-gland. 
The forces available to eliminate the honey-gland in the 
absence of ant attendants are considerable. There is the 
loss of the constantly necessary selection for maintaining 
it, a now useless and indeed wasteful organ must rapidly 
degenerate, but there is further a strong selective force for 
