380 THE ENTOMOLOGIST’S RECORD. 
this is interesting, more especially as I have never seen a /ilipendulae 
near the vicinity. Thirty-seven and a half per cent. are therefore pro- 
gressing towards the confluence of the spots, though the number of 
extreme specimens is very small indeed. Separated from this large 
colony by perhaps half an acre was another little colony, no doubt a 
recent extension from the other, where I took thirteen specimens, of 
which six are of the orobi form, six are typical with the median spots 
confluent, and one has all the spots confluent. In size only two can 
be considered large, the others are average, and the confluent specimen 
is smallish, 7.e., it is the smallest of the thirteen. I only visited the 
colony two or three times, as I did not discover it for perhaps a fort- 
night after I had been working my original colony. There remains 
the near extension of the parent colony, only separated by a stream 
and a very dense hedge; here I took fourteen specimens, all of which 
are large, eight of these are ovvbi and six are typical. Some three or 
four miles away from here I found another thriving colony of the 
species. Separated as it was by hill and dale as well as by a large 
trout stream, I have no doubt of its being an entirely independent 
colony. I only discovered it three days before I left the district, and 
it was at the very end of July, so that specimens are nearly all passé, 
but I took them for the sake of comparison. Of forty-one specimens 
twenty-three are orobi, sixteen are typical, and two have the median 
and terminal spots confluent ; the majority of the specimens are of a 
small average size, not more than half a dozen could be considered 
large, and all of these are with confluent median spots. 
It would be well to compare these results with captures from three 
other British localities where I have taken it. 
In 1909 I found it fairly common not far from Shanklin, in the 
Isle of Wight. Here the locality was quite different from its Devon- 
shire haunts, it was an old disused gravel pit surrounded entirely by 
corn land. Here I took thirty-five specimens, twelve of these are small 
specimens of the form o7vbi, nine are small typical specimens, and one 
small with median and terminal spots confluent. The smallest example 
I have seen is among these typical specimens, a beautiful little insect 
with very broad dark borders to the secondaries, almost inseparable 
from the race syracusiae. Kleven are fine specimens with large con- 
fluent median spots, of which four have prolongations on to the basal 
spots, two others are small with all the spots confluent, and one has 
distinct traces of a sixth spot. Here, however, filipendulae was abun- 
dant and active in the near vicinity. 
Another locality where I found the species was in the Cotswold 
Hills in the year 1900. It was, however, rare, and I did not discover 
a definite colony, and none of my captures had confluent spots. 
I also took the species in 1901, on the same ground as Lycaena 
arion in Cornwall, here—a sheltered spot—there was a small colony, 
but unfortunately I was intent on arion and only took five specimens 
of trifolii, though looking back I remember the colony well. Of the 
five, however, two are typical and large, three are orobi of average size. 
‘It would appear, therefore, that it is the habit of the species to 
develop a small and a large race, and this quite irrespective of the 
abundance of its larval food. I can attest to this in its Devonshire 
and Isle of Wight localities. 
That it is usual for it to develop two well-marked forms in abund- 
