8 ANTONY GEPP AND ETHEL S. GEPP. 



distinct from our Antarctic plants that we are compelled to separate these latter off 

 under a new specific name. Agardh's plant appears to have stem and branches terete, 

 and the branches become gradually thicker towards their base as they are traced down- 

 wards into the stem. In D. liarveyana the stem and branches are flat when dry, 

 elliptical when moist, and the branches gradually become attenuated towards the base 

 before joining the stem. Setchell and Gardner (Algse of N.W. America, 1903, p. 247) 

 also keep these plants separate, maintaining that the Antarctic D. media Kiitz. (Tab. 

 Phyc. IX., p. 40, t. 95, II.), which equals D. harveyana, is not the same as the North 

 Pacific P. media (Ag.), Grev., which they call P. aculeata, var. media J.Ag. Under 

 the latter name they cite No. 353 of Tilden's American Algse, a specimen of which we 

 have examined. It bears a descriptive label saying " stem and branches traversed 

 (axially) by an articulated filament surrounded by a layer made up of larger (air 

 cavities) and smaller cells." This we take to be the structure of typical D. media Ag. ; 

 and we do not find it to be identical with the structure of D. liarveyana. For in P. 

 liarveyana a transverse section of the elliptical stem (fig. 13a) shows, in addition to the 

 primary axial articulated filament, two or more parallel articulated tubes cut across and 

 situated at irregular intervals more or less on the longest diameter of the ellipse, 

 running longitudinally up and down the stem, each of the tubes being ringed round by a 

 sheath of small cells (pi. iii., figs. 13b and 14b). The actual position and inter-relations 

 of these tubes are at once seen upon examination of a fairly thick median longitudinal 

 section (fig. 14a), cut parallel to the flattened surfaces of the complanate- thallus. 

 Sufficiently good sections are quickly obtained by the rough-and-ready method of 

 shaving off the opaque cortex and then slicing down the flat medulla with a sharp 

 scalpel, the material being held gently between finger and thumb. The primary 

 filament with its septa and sheath of minute cells is then visible in the axis of the 

 thallus. At intervals it puts out on one or other side a lateral tube which bends 

 upwards, and after running subparallel to the axis for about 4 mm. turns outwards into 

 the next lateral branch above (fig. 14a). There are also two or more thinner secondary 

 tubes parallel to the axial filament, sheathed, and dividing usually at long intervals ; 

 these divisions anastomose with one another here and there, and occasionally with 

 the main filament. We think it possible that these tubes are the " einzelnc nicht 

 regelmassig vertheilte grossere Zellen, umringt von um die Halfte kleineren " 

 described by Reinsch (Flora 71, 1888, p. 190) as characteristic of his South 

 Georgian D. aculeata, var. eompressa, which plant would then be a synonym of 

 P. harveyaiKi. 



One of the forms recorded from the southern oceans is P. viridis var. dimtttns 

 Hook, et Harv., specimens of which are preserved in the British Museum. Though 

 they somewhat resemble P. liarveyana in habit, they differ from it in structure. The 

 structure is comparable with that of typical P. viridi*. A figure (fig. 16) is given here 

 of a transverse section of the thallus of P. viridis var. dixtttiiN, for comparison with 

 that of I), harveyana. 



