feeding caterpillars of Selenia tetralunaria, some on trees in the open air and others on cut food indoors. Those treated in the former way completed their development in the summer as usual; but the housefed larvæ, though they pupated in good time, did not become moths until the following spring. Thus, a normally double-brooded species was made single-brooded by being fed indoors, whereas the high temperature of a house should have had just the opposite effect to this. Mr. Merrifield finds in the condition of the food the explanation of this anomaly. He believes that feeding on cut branches wanting moisture, and with little circulation of sap, must have a retarding effect on the development of the insects.

An interesting account of the way in which the caterpillar of Acronycta alni makes its cocoon is given by Mr. M. Fitzgibbon ("Entomologist," vol. xxv., p. 40). The habit of this larva is to select a rotten stick, which it proceeds to hollow out, in order to form a shelter wherein it may winter in its pupal stage. The noteworthy point of the description is the use of the caterpillar's strong bristles in sweeping away the sawdust formed by excavation. These bristles are flattened at the tip, and seem specially adapted for this function. In the same paper the colouring of this caterpillar is discussed. In its early stages, it is supposed to be imitative of a bird-dropping, but after the last change of skin it becomes conspicuously adorned with warning-colours. Like many other larvæ, it hides from its enemies until it is too large to be hidden any longer, and then it tries to frighten them away.

An important contribution to Insect Embryology has been lately made by Dr. R. Heymons, of Berlin ("Zeitschr. für Wissensch. Zoologie," vol. liii., p. 434, &c.). This is a paper entitled "Die Entwicklung der weiblichen Geschlechtsorgane von Phyllodromia (Blatta) germanica, L.," which gives a very exhaustive account of the origin and formation of the ovaries and oviducts in that household pest known as the German Cockroach. This insect is a smaller and paler species than the Common Cockroach (Periplaneta orientalis), which it is said to be supplanting in many places in Britain, where it has succeeded in establishing itself. In his introduction, Dr. Heymons points out that technical difficulties have hitherto prevented a complete study of this subject. The eggs of the cockroach are enclosed in a chitinous capsule, and the yolk-mass contained in each egg is of a nature unfavourable for microscopic investigation of the early stages of development. These obstacles have now, however, been successfully overcome. The cockroach is an insect of a primitive and generalised type, and its development is not complicated by secondary metamorphic adaptations; hence it is specially valuable for embryological research.

The primitive germ-cells arise from the mesoderm, from the ordinary cells of which they may be distinguished by their larger size. They first appear towards the hinder end of the blastoderm, before the formation of the somites. They are also formed, however,

from the walls of the somites as these are developed, and, in the younger stages of the embryo, are always to be found in the second, third, fourth, and fifth abdominal segments. Thus, whether formed from the primitive layer, or from the cœlomic epithelium, the origin of the germ-cells is clearly mesodermic.

This result is contradictory to that lately obtained by Cholodkovsky, who believes that the germ-cells are probably formed from small yolk-cells. Dr. Heymons, however, has no doubt that the Russian observer fell into error owing to the imperfection of his method of preparation, especially as he was unable to distinguish the germ-cells before the stage at which the somites begin to form.

Certainly the mesodermal origin of the germ-cells seems by far the more probable of the two alternatives, and, after a review of the results obtained by workers in other insect groups, Dr. Heymons concludes that, among insects generally, though sometimes in a modified way, the mesoderm gives rise to the generative organs. In the origin and segmental arrangement of the germ-cells he sees a similarity to what is observed in the Annelids.

As development proceeds, the somites become largely filled up with the fat-body, and the germ-cells begin to take up a dorsal position. The differentiation of sex is now evident. Epithelial cells of mesodermal origin arrange themselves around and amongst the germ-cells. The epithelial cells ventral to the germ-cells form the oviduct, those dorsal to them arrange themselves in linear series to form the thread-like ends of the ovarian tubes, whilst others enclose the germ-cells and so form the ovarian tubes. Special attention is called by Dr. Heymons to the fact that, throughout the course of the development, the germ-cells are always quite distinct from the epithelium of the ovary.

During the larval stages, changes take place in the nuclei of the germ-cells, and they begin to assume the characters of ova.

The thread-like prolongations of the ovarian tubes appear to be of use only during the embryonic and larval periods, and their function is to enable the ovaries to change their position in the body of the insect. They are all continuous with a single thread of cells which stretches to the pericardial septum.

Dr. Heymons states that the contents of the ends of the ovarian tubes are ordinary epithelial cells and germ-cells. These portions of the tubes in different insects have been called "Endkammer" by various authors who have described a mass of undifferentiated cells as found in them. In Phyllodromia, however, the germ-cells are, as stated above, all along distinct from the epithelial cells.

Some statements have lately been made with regard to the effect produced by the parasitic Stylopidæ on their hosts, which are Bees, mostly of the genera Andrena and Halictus. Mr. R. C. L. Perkins ("Entom. Monthly Mag." [2], vol. iii., p. 1) gave as the result of his investigations on stylopised bees that the internal organs of the hosts

were hardly affected by the presence of the parasites, and that the males certainly, and the females probably, were still capable of reproduction. These results are extremely surprising, being in direct contradiction to those obtained by Shuckard, Westwood, and Perez. The last-named of these, a few years ago, gave a very full account of stylopised bees (" Actes Linn. Soc., Bordeaux," vol. xl., p. 21). Mr. F. V. Theobald, however, confirms, to a great extent, the work of these older observers ("Entom. Monthly Mag." [2], vol. iii., p. 40), and makes it likely that the specimens examined by Mr. Perkins were of a very exceptional nature. The Stylops, lying above the viscera, obtains its nourishment by osmosis through the abnormally thin walls of the alimentary canal, which, being forced down on the reproductive organs, renders them abortive. They do not, as a rule, contain ripe ova or spermatozoa. Indeed, Mr. Perkins admits that he could not find the former in any stylopised female.

The habits of another class of parasites are described by Herr C. Verhoeff ("Zool. Anzeiger," vol. xv., p. 41). These are bees of the genus Stelis which have long been known to lay their eggs in the nests of other bees (Osmia). Former observers had noticed the eggs both of the parasite and of the host in the food-mass in the cell, and at a later period had found only the larva of the Stelis. Hence it had been inferred that the Osmia grub died of starvation. It seems, however, that its death is really due to violence. The Stelis lays her egg before the Osmia, and at the further end of the food-mass, whilst the Osmia lays her egg on the surface of the food-mass. Hence the two eggs are at opposite ends of the food; and both grubs feed for a time without conflict. Later, however, the Stelis, which, being older, is larger and stronger than the Osmia, attacks the latter, digs its mandibles into the head of its victim, kills, and devours it. The time occupied by the Stelis larva in eating up the Osmia is one to two days.

It seems doubtful if this habit is one to which the term parasitism can properly be applied.

G. H. CARPENTER.

IX.

The Relationship of the Carboniferous Plants,

THE

Sigillaria and Stigmaria.

true nature of Stigmaria has long been a fruitful subject of discussion among palæophytologists. Brongniart, who appears to have first established and defined the genus, describes the specimens so named as stems bearing leaves, which ultimately disarticulated and left the surface marked by scars arranged in a quincuncial order.1 He seems to have had some knowledge of the internal structure of Stigmaria, and on the strength of this and other features placed it, with some reservation, amongst the Lycopodiaceae, pointing out that it had many points of analogy with the existing genus Isoëtes, and suggesting that it might be considered as a gigantic aquatic Lycopodiaceous plant, in fact, a sort of arborescent Isoëtes. The subsequent discovery by Binney and others of fossil Sigillaria and Lepidodendra, whose trunks were continued downwards at the base into large, spreading, and bifurcating structures recognisable as Stigmariæ,2 soon gave rise to another view of their morphological nature, viz., that they were in reality the roots of these trees, and not the stems of independent plants as had been generally supposed. From that time to the present, opinions have been divided as to which of these two views was the correct one, and as new facts have been brought to light, the advocates on either side have sought to interpret them in accordance with the opinions they hold. After carefully weighing the whole of the evidence available five years ago, Count SolmsLaubach came to the conclusion3 that the best explanation of the known facts was to be found in the hypothesis that the Stigmariæ were rhizomes adapted to a soft pulpy environment, whose appendages, though probably foliar in their nature, were functionally organs of absorption.

On the assumption that this conclusion was correct, the further question presented itself as to how the aerial Sigillaria with its Stigmarian rhizomes was developed. Was it to be supposed that

1 Prodrome d'une Histoire des Végetaux fossiles," 1828.

2 For details of these discoveries, &c., see Williamson, Stigmaria ficoides, Palæonto graphical Society, 1887; also Solms-Laubach, "Fossil Botany," Clarendon Press. 3 "Fossil Botany," pp. 291–293.

the Sigillaria gave rise to the Stigmaria, or vice versâ? To this question, so far as I am aware, no satisfactory answer was given, and its solution was left over until further investigations should throw fresh light upon it.

Some such light appears to have been obtained by Professor Grand' Eury, who has recently written an elaborate memoir on the Géologie et la Paléontologie du Bassin houiller du Gard, in which the subject is dealt with. Unfortunately, this memoir will only be accessible to a very limited number of workers. Printed under the auspices of the colliery companies of the district dealt with, only 125 copies have been struck off, and these, it is understood, will not be obtainable by purchase, but disposed of in another way. M. Zeiller, however, in presenting, on behalf of the author, a copy of the memoir to the Société géologique de France, has given a careful and somewhat lengthy summary of the palæontological portion, and has thus ensured for the main results a much wider publicity than they would otherwise have obtained. So far as these bear upon the morphological character of Stigmaria and its relation to the aerial stem of Sigillaria, they are embodied in the following paragraphs, for the substance of which I am indebted to M. Zeiller's excellent summary.

In the first place, it may be noted that Professor Grand' Eury, like Count Solms-Laubach, has come to the conclusion that Stigmariæ are rhizomes which floated in water or spread themselves out on the surface of mud. In this condition they were able to live on indefinitely without giving rise to an aerial stem. But his observations have led him much further than this, and have enabled him to indicate, at least in a general way, the mode in which these creeping rhizomes gave origin to aerial stems. The first step in the process is the knotting up of the Stigmaria, so to speak, in order to form a large bulb. M. Zeiller does not state whether this bulb-formation takes place at the growing tip of Stigmaria or at some other point, or whether both cases are possible. However this may be, the connection is said to be of such a nature that the bulbs are united to the Stigmaria by a vascular axis. At first, the bulbs present four swellings at the base, but these afterwards elongate, and ultimately show the cruciate disposition so often recorded as characteristic of the bases of the stems of Sigillaria. At first, neither the upgrowing stem nor the rootlike branches carry any appendicular organs, and consequently their surfaces are not marked with the usual characteristic scars. Thus the lower part of the bulb, in developing and branching, takes the form of Stigmariopsis, while at the same time the stem begins to raise itself vertically. At the base of the stems arising in this way, and which are often swollen into the shape of a bottle, M. Grand' Eury has observed neither true foliar scars nor any trace of a vascular bundle, but merely a number of glands arranged in pairs. Thus, the base of the aerial stem has the characters of the fossils known as Syringodendron, or rather those properly so-called, viz., those of the type of Syr. alternans, which