166 THE AMERICAN NATURALIST. [VoL. XXXIV. 
parent stock and becomes free-swimming. The parent stock, 
as soon as the stolon is separated, regenerates the lost segments 
and in like manner develops a second and possibly a third or 
fourth bud. (In some forms of Autolytus (4. varians) the 
regeneration of new segments takes place before the shedding 
of the mature stolon, and thus a chain consisting of, at times, 
as many as eight stolons in different stages of development, 
becomes attached to the parent stock. In Autolytus cornutus, 
however, as in Procerza, the stolon matures and becomes sep- 
arated before an addition of new segments takes place, so that 
the parent stock of this form never bears more than a single 
stolon at a time.) During the process of separation the stolon 
undergoes such changes as are of service to it in the change of 
its life from among hydroids to that of surface swimming, the 
most conspicuous of which are the modification of the para- 
podia and the development of the swimming setze. The sexual 
differences, as will be seen in comparing the free stolons, are 
also very conspicuous and appear quite early in the develop- 
ment of the stolon. 
The free-swimming male stolon of an Autolytus was first 
described by Oersted, in 1843, as a new species of annelid to 
which he gave the name Polybostricus, while the free female 
stolon of another species (Autolytus prolifera) was similarly 
described by J. Müller, in 1853, under the name of Sacconereis. 
It was not until 1862, however, that Agassiz, in observing the 
separation of the stolon in Autolytus cornutus, demonstrated 
for certain the relation of the parent stock and stolons. 
Alternate generation for annelids was first suggested by 
De Quatrefages in 1843, after observing a part of the budding 
process in a syllid (Sy//is monilaire) previously described by 
Savigny ; later again by Krohn, in 1852, for Syllis prolifera, 
but the first complete description of the process was given by 
Agassiz, in 1862, in a paper entitled * On Alternate Generation 
of Annelids and the Embryology of Autolytus Cornutus." The 
development of the stolons of this syllid and their subsequent 
separation he followed stage by stage, and established the 
identity of the separately described Polybostricus and Saccone- 
reis without a doubt. The stolons he described as sexual indi- 
