HOMALOMYINÆ Dorso-central bristles 2 in front of suture, 3 behind; anal vein short, axillary curved and continued beyond end of anal; arista nearly always almost bare; eyes bare. Small or rather small species; abdomen shortish, broad, flat, and with characteristic triangular dorsal spots or modifications thereof.

CENOSINE: Dorso-central bristles 1 in front of suture, 3 behind; anal and axillary veins straight and abbreviated; eyes of male never nearly touching, always bare; abdomen tubular, usually spotted.

In January, 1888, Mik called my attention to a distinction in the claws of Mydæinæ and Homalomyinæ.

Polietes hirticrura, Meade; confirmed: Kilmarnock.

I need not repeat Meade's additions, unless by way of criticism, as he has been kind enough to send me nearly all for examination.

Limnophora albifrons, Rond.; the specimen was Hydrophoria socia, Fln., .

Hydrotea parva, Meade; I think this is H. glabricula, Fln., which was previously reputed.

Pegomyia hyoscyami, Pz., and Winthemi, Mg., may be erased, the exponents being very unsatisfactory.

Homalomyia triangulifera, Meade (nec Rond.), is H. lepida, W., which by some extraordinary chance I omitted from my list, but noticed the omission just in time to include it among the reputed. H. nigrisquama, Meade; confirmed.

Canosia scrupulosa, Zett.; the specimens belong to Spilogaster, as at present constituted. C. pictipennis, Lw.; Meigen's name of costata is more generally accepted, even though he mistakenly called it a Sapromyza.

Pteropacila lamed, Schrk. In including this species in my list I fell into the same error into which Haliday had fallen more than fifty years before. Our British species is the little-known Toxoneura muliebris, Harr. (fasciata, Mcq.). Consequently P. lamed should be excised from my list, and T. muliebris added from the reputed list and placed next to Palloptera under the Lonchæidæ.

Oscinis rapta, Hal.; this should be added to the British species, and O. rapta, Mcq., excluded from the reputed species.

These are a few notes upon species concerning which I have arrived at some definite conclusion; but there are hundreds of others upon which I am in doubt. When I published my list I thought I fairly understood the genus Tanypus with 19 species; I now think I possess 40 species. In Simulium I admitted two species; I now think I possess about ten species. A vast amount of work remains yet to be done, but it had better be done with caution.

Sussex Lodge, Newmarket.

CONTRIBUTIONS TO THE CHEMISTRY OF INSECT COLOURS.

BY F. H. PERRY COSTE, F.C.S.

II. A PRIORI ANTICIPATIONS.

It will perhaps conduce to clearness if I now quote briefly a few extracts from the notes that I drew up for my own guidance before commencing these experiments. They are not without a moral-seen in the light of later and experimental knowledge; and, anyhow, my readers will thus be travelling along the same route that I myself pursued.

Now the results that I anticipated were to have a twofold significance, viz., from the phylogenetic and the chemical standpoints. Phylogenetically, I anticipated finding illustrations of the evolutionary doctrine of the origin of species, and of the various stages represented by each species in a natural group or genus,-at least so far as regards colouring,-just similarly, for instance, as Weissmann found in the markings of Sphinx larvæ.* Furthermore, I expected to obtain two wholly different lines of results. with these colours, according to the nature of the reagents used. I anticipated obtaining (a) progressive, and (b) retrogressive, modifications; and it also seemed possible that by such results some light might be thrown on classification. And, lastly, similar markings in different colours on allied species might be expected to yield a common colour (either progressively or retrogressively) under the influence of the same reagent.

Chemically, I hoped to compile a table of the colours in their biological rank, and noted it as very interesting, when the apparently same colour occurred in different groups, to determine whether it would be affected in the same way (cf., for instance, the yellow of the under wing in the Triphænæ and in Heliaca arbuti). Only after writing the above did I recollect that Grant Allen (reasoning, however, not from the results of experiments, but from observation alone) had drawn up a table of colours, in their biological rank, for flowers; and it struck me as very interesting to see how far my results might accord with his. I expected that in insects yellow would be found above white (not below, as in flowers), and below red. But what of black (probably above all these), blue, brown, &c.? That remained for experiment to show.

Nevertheless, I wrote,-very likely a linear arrangement of all the colours may be impossible; for as to the ranks of colours, I doubt if there be a universal unity; but rather suspect that a colour may rank below a second colour in one group, and above it in another. I imagine this to be the case from a consideration of the distribution of colour on various insects; broadly speaking,

*

'Essays in Descent' (Meldola's translation), I., 223, &c.

I take it that the ground colour of a wing is the lowest in rank, and primal, whilst bands and markings on it show more highly evolved colours. If so, in Pieris and Colias, black is above white and yellow; but in Vanessa atalanta, below white and red; and So on. This were certainly to assume that the rank of the colours corresponds with their historical order of development. It remains to be seen whether this be a legitimate assumption. Moreover, as regarding the identity of ground and primal colour, one must remember that a high colour may commence as a band, and end by covering nearly the whole wing, in which case the original ground colour would eventually appear secondary; hence arises a necessity for caution in making such assumptions.

[In view of the experimental results afterwards obtained, I may be permitted to point out that I had thus foretold beforehand the priority of white to yellow (antithetically to Mr. Cockerell's views); and of yellow to red, in order of development. On the other hand, it will appear how utterly fallacious were all my à priori speculations on the rank and behaviour of black; and also, I fear, on the relations of band and spot colours to ground colour.]

Now, as to progressive modifications, I prophesied by far the most interesting and vitally important results from the use of such reagents as the organic acids. For my reagents I divided into two main classes:-1st, those that are naturally present in the insect itself, or in the plants that it feeds upon; and 2nd, those altogether foreign to the insect and plant. In the first class were included such reagents as formic, acetic, malic, citric, succinic, oxalic, benzoic, and tartaric acids, &c.; also, by an extension, such oxidising reagents as potassic permanganate, seeing that oxidation perhaps plays an important part in the evolution of colours; although an objection might of course lie here against such reagents on account of possible chemical action other than oxidation. In the second class I included all such reagents as strong mineral acids (sulphuric, nitric, hydrochloric, &c.); strong alkalis,-ammonia, sodic hydrate, potassic hydrate; and any salts that might be tried, as, e.g., potassic ferro-cyanide, potassic iodide, argentic nitrate, &c. Evidently none of these could affect the insects in nature,-they are all foreign reagents; therefore I considered that no progressive modifications could be looked for from the use of these reagents, but merely retrogressive; for it seemed likely that by destroying the present colourings they would reveal the earlier stages. Subsequently, however, I noted that possibly these progressive and retrogressive reagents were too sharply divided, since "unnatural" reagents might produce the progressive modifica tions in experimenting, although in nature such modifications have been brought about by other means.

Now regarding this progressive modification of colour,although carefully guarding myself against expecting too much, or admitting so sweeping an anticipation as that of producing artificially all the stages in coloric development displayed in nature, seeing that such are probably due to the action of more causes than simply the metabolism of the insect; due, for instance, to variations of temperature, atmospheric and climatic conditions generally, yet I considered that by the external application of such already mentioned organic reagents, as in very small and dilute quantities are present in the insects themselves, I might produce progressive modifications of colour similar to those that have occurred already in many species. I hoped thus to complete changes already commenced in an insect itself, or displayed by closely allied species: for instance, to change the white of Pieris into the black, already present as spots and tip markings; the white of Euchloë cardamines into orange, like the tip; the pale yellow of Colias hyale into the deep yellow of Edusa; these yellows into black, like the borders; and the yellow of Gonepteryx rhamni into the red of G. cleopatra. Especially, too, in those species where the males are brilliantly coloured and the females dull, did I anticipate being able to change the female tint into the male (cf., Lycænæ, Euchloë, &c.).* How utterly fallacious and doomed to disappointment were all these anticipations, will subsequently appear.

Then as to the retrogressive modifications,-here, at any rate, the event proves me to have been on somewhat safer ground in my predictions. I have already explained that I expected to find such reagents as the strong mineral acids and the caustic alkalis convert the latest evolved colours backwards to a lower colour, thus giving me the original (sic) type colours; and this both as regards the several colours of one species, and the range of colours in a genus of many species. These results would form, therefore, an exact counterpart to those that I fallaciously hoped for in progressive modification. For instance, in Pieris I expected to turn all the black into white; in Euchloë cardamines, the orange back to white; in Gonepteryx rhamni, the red spots to yellow; and in Vanessa atalanta, the red into the ground colour, black.

Supposing these anticipations to be fulfilled, the question next occurred to me whether it might be possible by diluting a given reagent, or by taking a less strong one, to retrace retrogressively the colour evolution step by step. For example,-I thought, a powerful acid may turn Colias all white at once; but it would be far more interesting if (by using weaker acid) one

* In this way, too, I hoped to throw some light on various isolated observations regarding the effect of change of diet in producing coloric variation: as, e. g., that Arctia caia is far darker when the larvæ are fed on walnut leaves.

could first turn its black into yellow, and then the yellow to white; although, I should add, that in this particular case I fancied the black had never come through yellow at all, but direct from the Pieris black, and that from white; but, at any rate, in such an instance as that of G. cleopatra and G. rhamni, I hoped to reduce the red flush and spots, respectively displayed by them, to the ground yellow, and then this yellow to the type colour of the genus, as shown by the lowest forms. Again, in species displaying so many colours as the Vanessa do, it would be most interesting to find the lineal order of these, although, no doubt, some would be collateral, and not all unilineal.

Then, again, regarding the statement I have already quoted (p. 155, above), that the same markings differently coloured in allied species might be expected to yield a common colour on treatment with retrogressive reagents, it appeared to me that a crucial instance would be afforded by the species of Catocala, e.g., nupta and fraxini. Was the blue of fraxini, I asked, evolved via the red of nupta (or vice versa), or are not these two colours more probably collaterally divergent? In the first case one might expect this result:--fraxini, blue reduces to red: this to "x" nupta, red reduces to "x"; but in the second case, both the blue of fraxini and the red of nupta might be reduced by a common reagent (or by different reagents) to a common colour. An analogous instance to that of these Catocala is afforded by the two species, Euchloë cardamines and E. eupheno, in which latter the ground colour is bright yellow, and the tip orange. By thus applying this principle to all genera that display the same markings in different colours, I hoped to discover the actual order of coloric evolution in each species, and the genetic relationship to all the others.

Such were my expectations as regards retrogressive reagents. It will appear that they have been only partly realised, since I have not succeeded in destroying highly evolved colours, step by step, but have never obtained more than one retrogressive change.

That the above-stated opinions may be viewed in their proper light, I must again remind my readers that I have simply been quoting to them from notes written out before I had made a single experiment. I trust that I have not been unduly prolix in so doing, but it seemed to me that greater coherence and unity would thus be given to my paper than could obtain, did I omit all reference to my previous anticipations, and plunge directly into the experimental results. Moreover, the speculations I had indulged in of obtaining those progressive modifications appear to me to carry a moral and a warning, only illustrating once more the utter futility of relying on any à priori hypothetical views, however probable they may seem,-without subjecting them to the test of experiment. Hypotheses certainly are