it be carefully dissected from its surrounding connections as it passes upwards along with the cords uniting the first (sub-sophageal) and second post-oral ganglia, it will be found, under the microscope, as has been observed by others, to possess the well-marked anatomical peculiarities of an ordinary tracheal vessel. Passing abruptly backwards for about 0·07", in close proximity to the ganglionic cords and below them and the œsophagus, it gives off from its upper surface a branch (trachea ramalis) (fig. 1, b) about one-third the calibre of the main vessel; and the two continue their course parallel to each other for about 0.03". Near to this point a bifurcation takes place in both, the lower and larger vessel (trachea saccularis) (fig. 1, a) dividing somewhat anteriorly to the other. If the course of the two branches of the trachea saccularis is traced, they are found to pass round the nervous cords, one on either side, until they are about 0.25′′ in length, when each suddenly expands into a pyriform sacculus-the so-called 'salivary bladder'-(0.35" in long diameter). The corrugated inner membrane of the trachea is gradually lost near the orifice of the sac, while the outer membranes of the vessel are expanded over its surface, and form an exceeding delicate hyaline covering with an internal fibro-cellular coat. I have never found a liquid in these sacculi, nor have I seen them in other than a collapsed state. During its passage around the nervous cords each of the branches of the saccular trachea receives a large nerve, which passes along between the outer coats of the vessel and sends off branches to the sacculi and the dendritic' bodies to be hereafter described.

The ramal trachea after its bifurcation pursues a similar course to the saccular, the branches running in juxtaposition with those of the saccular trachea for about 0.15". They then commence to divide dichotomously, and continue so to do until the ultimate leaflets of the dendritic bodies are attained. The ramal trachea in its whole length and the saccular in certain parts are invested with a hyaline membrane, between which and the proper tunics of the trachea is a space filled with nucleated elliptic bodies (about the size of lymph corpuscles) and a granular yellowish-white material. This hyaline and corpuscular investment is of varying thickness, but is far more developed upon the ramal than upon the saccular trachea. En

veloping the finest branches of the ramal trachea it accompanies them throughout their somewhat convoluted course until they have arrived at their ultimate ramifications, where it appears to form the chief substance of the cordate leaflets of the dendritic bodies (fig. 2) into which the ramal tracheæ expand.

The two delicate arborescent bodies thus formed, designated in modern hand-books' the 'salivary glands,' are, in the adult insect, about 0'45" in long diameter, or upwards of one-third the entire length of the animal. The comparative size is much less in the young insect. By their upper and inner surface they are applied to the under part of the œsophagus, and are retained in their position to a great extent by the ramifications of the œsophageal tracheæ over their surface. These œsophageal tracheæ pursue a course distinct from the ramal trachea, and nowhere enter the substance of the bodies in question. Between the two secondary branches of the ramal trachea, on either side, is placed the corresponding sacculus of the saccular trachea. A delicate net-work of fibres connects the sacs with the leaflets and the latter peripherally with one another. These fibres were at first supposed to be bands of connective tissue, but may possibly be nerve fibres. They anastomose frequently with each other at short intervals, and contain in their interior occasional deposits of pigment matter with oval and stellate nuclei and granules. The fibres can be traced passing beneath the external hyaline coat common to the sacculi and leaflets, and they then seem to be assimilated to the inner coats of the sacculi and leaflets.

We

What can be the functions of these dendritic bodies? are forced to set aside the hypothesis that they are in any way connected with the production of saliva for the following

reasons:

1. The position of the orifice of the common trachea upon the posterior and under surface of the lingua, and therefore external to the true oral cavity.

2. The so-called 'common duct' presents simply the appearance of an ordinary trachea, both as regards the spirally corrugated arrangement of its internal coat, which of itself

1 Cf. Rolleston, Forms of Animal Life, p. 200 (1870). Huxley, An Introduction to the Classification of Animals, p. 57 (1869), et complures alios.

points to its function as the transmitter of a gaseous fluid1 rather than a liquid, and the nucleated appearance of its outer

coats.

3. It can be readily shewn that all the branches of the ramal and saccular tracheæ open into this common vessel, wherefore any fluid in one part will probably find its way into any other. Thus, if we shew that the 'common duct' transmits a gaseous fluid (such as air), all portions of this system will probably contain that fluid.

4. If the so-called 'salivary receptacles' were intended for the retention of a liquid secretion, we should expect to see some of this fluid occasionally within them. In upwards of thirty dissections made by me, these sacs have, upon opening the thorax, always been found collapsed and apparently empty.

5. That the common duct' together with the 'receptacles' contains an elastic fluid is proved by the fact that it is possible to inject them by using the method employed to inject the airtubes of the body generally: that is, by placing the insect immersed in a suitable liquid beneath the exhausting receiver of an air-pump.

6. The collapsed condition of the sacculi is such as would of necessity result upon opening the thorax, supposing these delicate bodies contained air previously, or at all events a gaseous fluid freely communicating with the air.

7. We have no reason for supposing this fluid is other than air.

I have thus attempted to shew how unfitted these sacculi and vessels are for the retention and transmission of any liquid secretion, and how well adapted they are, on the other hand, for the function of aëration. If the dendritic bodies, therefore, are of a glandular nature, they belong to the category of ductless glands.

1 Excluding the so-called 'salivary ducts,' I can find no instance in insects of a vessel with a spirally corrugated inner membrane, formed for the transmission of a liquid secretion. The false trachea' observable in the proboscides of some insects differ both anatomically and physiologically from trachea. (Cf. Lowne, Anatomy and Physiology of the Blow-Fly, 1870.)

246 MR HOLLIS. SALIVARY GLANDS OF THE COCKROACH.

DESCRIPTION OF THE FIGURES,

FIG. I. Labium of the cockroach, shewing the origin of the common trachea.

a. One division of the bifurcating saccular trachea;

b. Ditto of the ramal trachea;

c. Anterior surface (dorsum) of lingua; d is placed over the orifice of the common trachea;

ee. Sublingual palpi;

ff. Labial palpi;

g. One of the divisions of the true bifid labium.

FIG. II. A portion of the dendritic bodies, shewing the ramifications of the ramal trachea and their expansion into leaflets. At 'a' the investment is seen to be continuous with the substance of the leaflet. (This specimen had been steeped in glycerine for upwards of six years, and did not shew the dichotomous divisions of the trachea distinctly. The network of (nerve ?) fibres is however shewn at 'b, b.')

FIG. III. A portion of the common trachea near the oral orifice. The spiral corrugations of its internal coat are apparent, and at the circumference (a, a) the hyaline outer coats can also be seen.

FIG. IV. Under (or posterior) surface of the lingua; a, the common trachea; b, the sulcus; c, the orifice of the common trachea; the softer membranes appear to be attached to a firm chitinous framework.

OBSERVATIONS ON PHYSIOLOGICAL CHEMISTRY. BY JAMES BLAKE, M.D., F.R.C.S., San Francisco, California.

THE object of the following observations is to consider the bearing of some recent chemical discoveries on physiology, and to call the attention, not only of physiologists, but of chemists, to certain facts which I think show that in living substances we possess reagents which may be made available for investigating the chemical properties of other compounds. Before, however, entering on the direct subject of my communication, I would offer a few remarks on the method that is being followed in studying the chemistry of living compounds. Owing to the complex nature of these investigations it is most important that a right method should be pursued in conducting them. Here, as in every other branch of scientific research, the only true path to pursue is by gradually proceeding from the more simple to the more complex, and thus arriving by slow but certain steps from the known to the unknown; and yet in no other department of scientific research is this simple rule so completely ignored as in physiology. How large a share of physiological investigation is devoted, for instance, to researches on the nervous element, one of the most complex of living substances, and one whose properties are probably the most dissimilar from those substances with the chemistry of which we are best acquainted. After making so unfortunate a selection as regards the field for investigation, physiologists have, I believe, been equally at fault in the choice of their reagents, these having been selected from amongst substances with the chemical and physical properties of which we are most ignorant, fully justifying the assertion of Drs Brown and Fraser at the beginning of their interesting memoir "On the connection between chemical constitution and physiological action," &c. (Journ. of Anat. and Physiol. Vol. II.), where they state, "Unfortunately we know next to nothing of the constitution of the majority of those substances, the physiological action of which has been most carefully investigated." What would be thought of a chemist, who,