110 PROCEEDINGS OF THE NATIONAL MUSEUM. vol.53. 



ing the animal with its ventral side uppermost so that right and left 

 were reversed. Furthermore by placing the mouth on the same 

 surface as the antennae, whether he called it dorsal or ventral, he 

 has altered the true amount of torsion, that is, the angle between the 

 buccal and anal orifices. 



Hence what he gave as a direct torsion of 135° might well be an 

 inverse torsion of 45°. Again his statement that "the value of the 

 angle of torsion in the adult furnishes very precise specific charac- 

 ters" (p. 203) can not be proved in actual experience. No species 

 shows anything like constancy in either the amount or the direction 

 of torsion. We have just stated that the neck of Pennella filosa is 

 often twisted in corkscrew fashion inside the cyst, but this is by no 

 means always or even usually the case. It is evident, however, that 

 the torsion of these corkscrew specimens will be very different from 

 that of normal specimens, and even amongst the latter there is a wide 

 variation. In most Pennellas the direction of torsion is clearly 

 shown by the twisting of the color pattern on the neck. Interpreting 

 on this basis, the torsion of five specimens of Pennella filosa taken 

 from the same swordfish are: Direct 90°, inverse 180°, direct 3G0°, 

 direct 3G0°, inverse 180°. Manifestly when two torsions are in the 

 same direction and one is four times the other, they do not furnish 

 a " very precise " specific character. 



The only rational conclusion is that torsion in this genus has com- 

 paratively little specific value. We are forced, therefore, in making 

 a key of the species, to rely upon an accumulation of characters, 

 rather than upon any one or two characters alone. 



Furthermore it has been found necessary to exclude the following 

 species from the key for the reasons stated after each : 



Penella anthonyi and P. cettei Quidor. In his attempt to use the 

 mode of branching of the plumose appendages as the chief charac- 

 teristic of the various species Quidor gave figures showing these ap- 

 pendages in the two species here named (1912b, pi. 1, figs. 9, 11, and 

 12). He also gave photographs of the two parasites, half the natural 

 size (pi. 4, figs. 30 and 31), and a few additional facts in his key to 

 the species. But he does not even mention them in his text, and we 

 are left in absolute ignorance of the details of all their appendages, 

 as well as of their general morphology. 



Consequently while they may very likely prove to be new species, 

 we can not accept them upon present evidence. 



Lerneopenna hlainvillii LeSueur (1824, p. 289, pi. 11, figs. 2 and 3). 

 A synonym of Pennella exocoeti (see p. 115). 



Pennatula hocconii Lamartiniere (1798, p. 51, pi. 20, fig. G). Milne 

 Edwards made this a distinct species, but Steenstrup and Liitken said 

 it was not well enough described to warrant separation and made it a 

 synonym of Pennella diodontis. It was the first species (type) of 



