1897.] Zoology. 443 
is reduced to four. While the Condulegaster show a striking concen- 
tration of the armature, which is limited to two pairs of teeth. In fact, 
the original radially symmetrical arrangement is seen to have been 
transformed to a bilateral one. The structures are best studied in the 
larvee, for they become considerably reduced and obscured in the imago. 
Summarizing his results in the form of a phylogenetic tree the author 
considers the agrioninze and the petalurinæ to have arisen from the 
primitive form, calopteryginz. Then from the petalurinz there arose 
three branches, two of them terminated by the eschnine and the gom- 
phinæ respectively ; the third passed off to one side as a low branch 
that formed the cordulegastrine. From this form there arose one 
branch that soon divided and finally gave rise to the corduliinz and the 
Libellulinz, the highest of the dragon-flies. The author’s conclusions 
differ from those of Calvert in that the cordulegastrine form a link 
between the two forms represented by the subfamilies corduliine and 
Libellulinz, and that represented by the petularinz, instead of an inde- 
pendent branch. 
The Regeneration of an Antenna-like Structure Instead 
of an Eye.—The regeneration of a structure very much resembling 
that animals antennula on the stump of an eye stalk of Sicyonia sculpta 
is well worth recording along with the regeneration of a well-formed 
lens from the iris in Triton, as described by Wolff and also by Müller 
whose paper was noted in the NATURALIST some time ago (p. 72). 
The regeneration of such a structure is described by C. Herbst’ in © 
several out of eighty-five specimens from which he cut the eye. Only 
six of the eighty-five remained alive at the end of five months after the 
operation, but all but one of these showed evidence of a regenerated - 
structure. Some seven other cases were secured during the five months 
by fixing the animals recently dead, or about to die, so that he lad 
twelve good cases showing a regenerated structure. Similar experiments 
had previously been performed upon the eye stalk of Palæmon, with like 
resuits. 
The accompanying figures represent the three groups into which 
Herbst divides the regenerated structures according to their degree of 
perfection. In the first (fig. 1) there is shown only a small protuber- 
ance (n) having little evidence of segmentation. In the second group 
(fig. 2) there is developed a large process provided with two hairy areas 
and a secondary two-jointed lobe that Herbst likens to a crustacean 
flagellum (fr). In the third group (fig. 3) a large antenna-like struc- 
1C. Herbst. Festschrift der Naturf. Gesel. Zurich, 1896, pp. 435-54. 
