BY E. J. TILLYARD. 285 



set any arbitrary limits to the various Orders, for each type would merge by 

 small degrees into the next. As matters stand, the chain of types already dis- 

 covered is practically complete enough for us to indicate the courses of the 

 various lines of evolution without any doubt; the more complete it may become, 

 by discovery of new intermediate types, the more difficult it will be to uphold 

 any one of these new fossil Orders as a separate entity. Yet we may not place 

 these types within recent Orders, without leaving it to be inferred that we be- 

 lieve them to have belonged morphologically to such Orders in other characters 

 besides the wing-venation. Belmontia and Parabelmontia were plainly, from 

 their venation alone, not true Scorpion-flies, but more generalised insects, pro- 

 bably combining the more archaic characters of the true Meeoptera with those 

 of the Megaloptera. 



Genus Parabelmontia, n.g. (Plate xxxiii., fig. 2; Text-fig. 3.) 



Foreicing. — Costal space narrow, with humeral (fern) and distal {dc) vein- 

 lets present, and Sc forked distally into Sci and Sc2. Ei a strong, straight, un- 

 branched vein, connected with Se2 distally by a short cross-vein, and with R2 by 

 another one, close to the former. Rs arising from R at about one-fifth of the 

 wing-leng-th, and dividing into Ra+s and R4-1-5 slightly before half-way. R2-f3 

 divides into R2 and Rs at a level only slightly distad from that at which R4-1-6 

 divides; R2 is simple, but R3, though ending simply on the wing-margin, divides 

 to form a small elongated cell, which is closed again not far from the margin, by 

 fusion of the two branches. R4 and R5 both divide again about half-way along 

 their lengths, and their branches run free to the margin, R4b being at the apex 

 of the wing. The radial cell (re) may be considered closed, much as in Bel- 



Text-fig. 3. Parabelmontia permiana,n..s.e:tsv. (x 6). For lettering, see p. 260, 



montia, by the cross-vein placed distally between Rs and R4a, and is also crossed, 

 about its middle, by another cross-vein. M is fused basally with R, but leaves 

 it below hm, and shortly afterwards divides into Mi-4 and Mg, the latter being 

 a very strongly formed convex vein forming the upper arm of a large cubito- 

 median Y-vein closely resembling that of Belmontia; the lower arm, or basal 

 piece of Cui, is broken near its middle at, a point where a cross-vein descends 

 from it on to Cu2. Mi-4 divides slightly before the level of the first dichotomy 

 of Rs, and each branch again divides into two, of which the upper branch in 



