444 DISCOVERY REPORTS 



Around the upper limit of the labrum, at the level of the small rigid knob, the " frontal 

 knob" (Figs. 2, 4), between the attachment of the antennules, there is another less- 

 marked chitinous loop which extends backwards to the inner point of attachment of 

 the antennae, and is then produced into the body on either side as a powerful endo- 

 skeletal tube. These slope back for a short distance, and each then divides into a small, 

 outer branch and a larger, flat, inner branch (Figs. 2, 3, 5), which lies more or less 

 in a frontal plane and slopes backwards to become embedded in the endosternite just 

 anterior to the anterior hypostomal apodeme. They have been termed the "Oberlip- 

 penaste " (Miiller, 1894, p. 19). Since they are inward extensions of the fold of cuticle 

 which forms the junction of the median face of the antennae with the labrum, I am 

 calHng them the " antenno-labral apodemes". 



The upper labral loop and the equatorial loop are joined by a thickening down the 

 sides of the labrum, completing a triangle with the endoskeletal apodemes. The whole 

 of the labrum is thus supported by a rigid scaffolding from the adductor muscle. 



A third paired endoskeletal structure exists, which Miiller has termed the " Sternalaste 

 2" (1894, p. 19). It is not, however, of the same nature as his " Sternalaste i " which 

 are true apodemes. It consists of a chitinous strut, which spans the body cavity from 

 the posterior corner of the hypostome to the posterior attachment of the maxillule to 

 the body (Fig. 2). It has presumably been formed by the nipping off of the edge of an 

 infolding between maxillule and maxilla in a manner comparable to the formation of 

 the intersegmental bars, which become transverse tendons in Hemimysis (Manton, 

 1928, p. 414). It should thus be distinguished from an apodeme, the development of 

 which (Manton, 1928, p. 412) suggests that they arise by the gradual sinking into the 

 body of the surface of attachment of a muscle or group of muscles. Hence I have called 

 it the "posterior hypostomal strut". At its upper end it sends forward a small pro- 

 longation which is a true apodeme, serving for the attachment of a muscle which runs 

 to the anterior hypostomal apodeme, and then continues up the side of the body as 

 a cuticular thickening. 



In addition to these paired skeletal structures, there is an excessively thin median 

 apodeme — the frontal apodeme — arising from the ectoderm just above the frontal knob 

 (Figs. 2, 3, 4). It is so delicate that it cannot be seen in preparations of the ventral 

 skeleton. In sagittal sections, however, it can be seen as a triangular fold of cuticle 

 in the median plane which narrows down to a thin hollow strand and passes through 

 a small aperture in the hinder margin of the brain. It then curves upwards and forwards, 

 and spreads out as a triangular plate, to which is attached a thin tendinous sheet, the 

 aortic tendon (Figs. 4, 5). 



The frontal apodeme is a very important structure in relation to the blood system, 

 and will be referred to later. The paired elements, together with the armoured upper 

 lip, function mainly as limb supports. They support certain limbs directly. Thus the 

 mandible, at the lowest point of its attachment, is joined firmly to the equatorial 

 thickening around the labrum just in front of the mouth. The rest of the attachment 

 is thin, the limb merging into the lateral body wall. The articulation of the maxillule 



