138 CARL SKOTTSBERG 



for the petals in the species from Masatierra were described and figured with 

 a distinctly longer nail than in the species from the continent or from New 

 Zealand. In 1908 I only got scanty material, but the Masafueran Sophora was 

 found with pods. As these were smooth and entirely unwinged, I separated 

 the island forms as subsp. fernandeziana (Phil.) from the ordinary 5. tetraptera. 

 This time I have a large material at hand, having also studied the collections 

 in Santiago and Kew. 



All plants from Juan Fernandez differ from Chilean or New Zealand Sophbras 

 in the smaller number of leaflets, the smaller flowers, the longer nails of the 

 petals, and in the unwinged pod. It is noteworthy that there is, in New Zea- 

 land, a species with small flowers, narrow-winged pod and only 2 — 4 pairs of 

 leaflets, 5. prostrata Buchan. In some respects, this comes nearer to the Juan 

 Fernandez forms, but is, of course, quite distinct. 



Of PHlLIPPl's species, Edw. masafuerana is easily distinguished, as seen 

 from my description and figures. All specimens collected in Masafuera (Gua- 

 JARDO! — the type — , Johow! and the writer) are exactly like each other 

 and rather unlike all forms from Masatierra. The difference between Edzv. 

 Reedeana and fernandeziana appears to be slight. My figures were prepared 

 from PHlLIPPl's type material; they show a certain difference in the shape of 

 the petals. PHILIPPI described the leaflets of fernandeziana as spathulate, but 

 his figure (pi. VIII B) shows an almost elliptical leaflet; comp. also my fig. 12 a, /. 

 Unfortunately, none of my specimens are in flower, what makes the identifica- 

 tion with PHlLIPPl's forms uncertain. No, 322 from Pto Ingles has 8 — 14 pairs 

 of narrow elliptical leaflets, densely villous with reddish or silvery hairs, indicating 

 that they were not fully developed. The same form was collected by BERTERO 

 (no. 1 5 16! Herb. Kew); here the leaflets have attained full size and show a 

 tendency to become slightly spathulate. I believe I am right in bringing these 

 to fernandeziana. 



No. 599 from Vaqueria, also found in Q. Juanango, where mature pods were 

 gathered, is a robust form with 8 — 10 pairs of large leaflets (13 — 14 X 5 — 8 mm). 

 I bring this to Reedeana. Seeds were sown in the Gothenburg Garden in 1918, 

 but the plants are still (Sept. 1921) rather weak and differ from the parents in 

 being nearly glabrous, probably a direct result of greenhouse life. Nos. 63 and 

 especially 214 differ from the former by their very slender twigs; the leaflets 

 are broader, perfectly glabrous above and with few hairs below; there are 8—10 

 pairs in no. 63 (size 8 — 10 X 5 mm) and 6—8 pairs in no. 214 (size 6 — 8x 4 — 5 mm). 

 Both are sterile. I distinguish these plants as f. gracilior. With regard to the 

 marked variation in the Masatierran Sophoras, I for the present comprise all 

 forms under the specific name fernandeziana. S. Reedeana is retained as a 

 variety, so that it be not forgotten. A definite arrangement can be made only 

 after an examination of a larger material, with flowers and pods of all the forms. 



Finally, I shall say a few words about the Sophoras of Chile and of New 

 Zealand. S. tetraptera occurs in the latter place in two distinct forms, or 

 species, 5. grandiflora (Salisb.) and microphylla Ait. No specimen from Chile 

 that I have come across matches 5. microphylla, but all are very like S. grandi- 

 flora in most respects. As a rule, the leaflets number 10—20 pairs in both, 

 they may be a trifle broader and more oval in the Chilean plant, but this is 



