AuGusT 29, 1912| 
LEETERS TO THE VEDITOR. 
{The Editor does not hold himself responsible for 
opinions expressed by his correspondents. Neither 
can he undertake to return, or to correspond with 
the writers of, rejected manuscripts intended for 
this or any other part of Natures. No notice is 
taken of anonymous communications.] 
Butterfly Migration in Relation to Mimicry. 
Tue last paragraph in Prof. Poulton’s letter in 
Nature of June 13, referring to Mr. Swynnerton’s 
experience that adult birds possess “a very fair know- 
ledge of the main types of pattern and relative 
edibility of the local butterflies," reminds me of the 
only occasion on which I have seen a_ butterfly 
attacked by a bird during five years’ observation in 
this district. 
While walking in the observatory compound my 
attention was attracted by a Lycznid butterfly of an 
unfamiliar species, probably a migrant from a much 
lower elevation (the observatory is situated in the 
Palni Hills at an altitude of 7700 ft. above sea-level). 
I was watching the mazy flight of the insect in the | 
expectation that it would settle, when I noticed a 
shrike sitting on a post near by, also observing it 
attentively. He evidently had a fair knowledge of 
the local butterflies, and considered this to be some- 
thing new and worth eating, for he suddenly jumped 
from his perch and very cleverly caught the butterfly 
on the wing, a surprising feat for a bird having a 
rather clumsy build and heavy flight. Apparently he 
swallowed the insect entire, for I could discover no 
wings at the spot afterwards. 
The general immunity of butterflies and day-flying 
moths from attack by insectivorous birds is as strik- 
ingly evident here as it is in England or America, 
and seems a serious difficulty in the way of accepting 
the Batesian theory of mimicry. Not only do the 
birds of this district pay no attention to the common 
butterflies, but the latter seem to despise the former. 
I have even seen a small bird frequenting the bracken 
of the uplands chased for a considerable distance by the 
vigorous and somewhat aggressive Argynnis castetsi! 
In contrast with this immunity I have found that 
nocturnal moths, if forced to take long flights during 
the daytime, are very liable to attack, and in these 
circumstances stand a very poor chance indeed of 
reaching a haven of refuge. The watchful birds seem 
ever on the alert to snap up strangers. 
It would seem, then, that unfamiliar lepidoptera 
are much more liable to attack than the common 
everyday kinds. May it not be that the real danger 
to a species occurs during migrations, and that 
mimetic resemblances may afford a real protection 
during such flights? In entering a new district a 
mimetic species would be immune from attack if the 
birds were familiar with the model, even if the latter 
were not unpalatable, while, on the other hand, un- 
palatable species migrating would be liable to attack 
if unfamiliar to the local birds. 
In this district annual migrations occur across the 
Palnis during October and November of a consider- 
able number of species from the plains, including the 
following mimics and models :—Hypolimnas bolina, 
H. misippus, Euploea core, Danais plexippus, D. 
septentrionis, D. limniace, Papilio polytes, P. hector. 
In these mivrations it is noticeable that the mimetic 
species, H. bolina and H. misippus, are very liable to 
have torn wings, suggesting attack by birds, and it 
appears that the models as well as mimics are also 
sometimes attacked. An instance has been recorded 
by H. Leslie Andrewes (Journal of the Bombay 
Natural History Society, xx.. 850), who found evidence 
of systematic onslaught by King crows (Dicrurus) on 
NO. 2235, vor. 89] 
NATURE 
659 
Danaids and Euploeas, also H. bolina 2 and Catop- 
silia crocale. This was near Ootacamund in the Nil- 
giris in October. All the species mentioned by him 
are migrants from the plains, and I believe do not 
normally inhabit the Nilgiri plateau, although com- 
monly seen at that particular season. There is a 
significant absence in the list of the very abundant 
and indigenous D. nilgiriensis, which would be well 
| known to the local birds. 
P. hector (the model of one form of P. polytes 2) 
appears to be specially liable to attack when migrat- 
ing across the Palnis, if one may assume that wings 
on the ground are good evidence of attack by birds. 
This liability to attack of migrants passing over 
the Palnis or Nilgiris appears, however, not of much 
significance when large areas are considered. A 
mimic such as H. misippus Q possesses, so to speak, 
a passport over the whole of the plains of India, 
Persia, Arabia, &c., owing to its close resemblance 
to the very abundant and widely distributed plains 
butterfly D. chrysippus. The facility thus afforded for 
dispersal would surely be an important factor in the 
life of the race. J. EvEerRsHED. 
Kodaikanal Observatory, South India, July 9. 
Parallel Mutations in Oenothera biennis. 
In a culture of a particular strain of O. biennis, L., 
a series of forms has been observed which constitute a 
parallel series to the well-known mutants from 
O. Lamarckiana, Ser. O. biennis, unlike O. 
Lamarckiana, has small flowers and a short style, 
rendering the flowers rather strictly self-pollinating. 
The particular race in question I received from the 
Madrid Botanical Garden. It has typical O. biennis 
flowers, as mentioned above, but the foliage closely 
resembles that of O. Lamarckiana. These plants were 
grown, to the number of 131 this year, at the John 
Innes Horticultural Institution, Merton, in connection 
with my other Oenothera cultures, which number in all 
more than 10,000 individuals. 
While in the majority of the plants in this culture 
the foliage resembles O. Lamarckiana or O. rubri- 
nervis, several have leaves corresponding to the 
mutants, there being six laevifolia, one lata, and pos- 
sibly one gigas. With larger cultures probably other 
mutant types will also be found. The peculiar 
characters of the lata foliage are even correlated with 
sterility of the anthers, as in the lata from 
O. Lamarckiana, though the flower otherwise is that 
of O. biennis. 
Though the foliage characters of these O. biennis 
forms are not identical with those of the Lamarckiana 
mutants, yet they differ from each other in correspond- 
ing ways, and thus form a parallel series. 
The interesting question as to the origin of this 
strain of O. biennis cannot be answered at the present 
time. Even if they originated through crossing (as 
seems probable), their flowers are now self-pollinating, 
so that each individual, with occasional exceptions, 
must represent a ‘‘pure line.”” The most probable 
assumption is that, as in the case of O. Lamarckiana, 
the aberrant forms all originated from one type 
having Lamarckiana-like foliage. Of the mutant 
types in this culture, the O. biennis lata at least has 
evidently talken its origin directly from one of the other 
types, since it produces no pollen. It has probably 
arisen through such irregularities in the distribution 
of chromosomes during the meiotic processes as I 
have described for the O. Lamarckiana series of forms, 
and the presumption is that some of the other mutant 
types have had a similar origin. This is in harmony 
with my hypothesis that the mutation phenomena in 
O. Lamarckiana are not due merely to hybrid splitting, 
