M\\\ DIATOMS OF THE ALBATROSS VOYAGES. 363 



These he called Donkinia. fl The third genus to be separated was Rhoicosigma 

 by Grunow, 6 including all species with strongly bent frustules (seen in zonal view) 

 and with valves sometimes, though not always, dissimilar. These distinctions have 

 been accepted by Van Eeurck, Peragallo, and l)«- Toni. < leve rejects them, «•■ 

 Toxonidea, and accepts this, as he admits, on dubious grounds. 1 think nothing is 

 gained by breaking up the generic unity of this sharply marked genus. The char- 

 acteristics, as Cleve points out, c are variable in each of the alx I or Bub- 

 generic division they can be, a1 least at present, made use of; bul beyond this they 

 only increase the difficulty of assigning diatoms that have the clear < ryrosigma quali- 

 ties. 



('lev*' has, however, gone further than others toward obliterating the worth of this 

 genus by dividing il into two genera on the sole basis of the direction of the Btriation. 

 He includes all forms with transverse-longitudinal striae in Gyrosigma Bass. I 

 and the others, with striae in three directions (excepting Toxonidea i in Pleurosigma 

 (\Y. Smith' Cleve. Ii is probably true, as he -ays. that then' are no transitional 

 forms between these two kinds; indeed, it would not be easy to conceive how there 

 could be, for lines in two directions could hardly change to lines in three directions 

 without assuming new angles, and only one of the three lines could possibly remain 

 in the original direction of one of the two lines; and this is jusl what we find, both 

 the two-line and three-line forms having one of the lines transverse. The lack of 

 transition is. therefore, to my mind an empty argument. If, now. we look at the 

 striking forms, called by Ralfs Donkinia, a genus rejected by Cleve. wo find Bpe< ies 



of both two-line and three-line sculpture that are far I lose to put into separate 



genera, on any reasonable interpretation of a genus; thus J>. recta (Donk.) Gran. 

 and I), carinata (Donk.) Ralfs."' Whoever also compares such two-line forms 

 (Pleurosigma) UttordU (W. Smith) Griff. & Henf. and G. (Pleurosigma) acumii 

 (Kiitz.) Grun. with such three-line form- as G. Pleurosigma latum <irun. and G. 

 (Pleurosigma) affiru <irun. must see that it i- artilioial to place them in two genera. 

 I do not therefore look upon Cleve's division of this genus as having the merit of the 

 earlier on< 



Gyrosigma aestuarii Breb. I Griff. & Henf. Micr. Diet. ed. 3. 350. pi. 11. /. 85. 1875. 

 Nm icula aestuarii Breb.; Kin/.. Sp. Alg. 890. 1849. 



Ph urosigma aestuarii W. Smith, Synop. Brit. Diat. 1: 65. />/. SI. f. 275. 1853. Cleve 

 & Grun. Sv. Vet. Akad. Ilandl. 17-: 52. 1880. Perag. Le Diatomiste l 4 - 5 : 12. 

 pi. 5.f. 12-1S, 15. 1891. Cleve, Sv. Vet. Akad. Handl. 26-: 12. 1894. Pritch. 

 Bist. Infus. ed. 4. 916. 1861. Rabh. Fl. Eur. Alg. 1: 234. 1864. Griff. & Henf. 

 Micr. Diet. ed. 3. pi. 15. f. 85 (/. SScI L875. Van Heur. Synop. pi. 18. f. 8. L881. 

 H. L. Smith. Sp. Diat. Typ. qo. 394. L874 (not Cleve, Bih. Sv. Vet. Akad. 1": 

 13. pi. .'./. 19. 1873=/'. normanii Ral 

 Pleurosigma angulatum W. Smith, variety; Van Hour. Synop. L15. L881; Treat. 



Diat. 251. pi. 6.f. 258. 1896. De Toni, Syll. Alg. 2: 232. L891. 

 Pleurosigma candidum Schum. Schrift. Phys. Okon. Ges. Konigsb. 8: 59. pi. .'./. 



57. 1867. Perag. Le Diatomist* 1 ; - : L2. pi. 5.f. 11. L891. 

 The do-ire of H. L. Smith, Van Heurck, and De Toni to make this a variety of 

 P. angulatum \V. Smith i- no, to bo commended. Though smaller than the latin-. 

 the markings are decidedly coarser, the ends narrower and less gracefully tapered, 

 and in the type form slightly but sharply benl Bidewise. 

 Found at station 2823, Gulf of California. 



a Pritch. Bist. Enfus. ed. 4. 921. pl.8. f. 9. 1861. 



^Sv. Vet. Akad. Ilandl. 17-: ~>s. 1880. Le Diatomist< 1 : 33. pi. 10. f. : 8. L891. 

 'Sv. Vet. Akad. Handl. 26-': 32, 112. 



& Trans. Micr. Soc. Lond. n. b. 6: pi L858. Le Diatomiste 1 : pl.9.f. 



6, 8. L891. 



