46 CARL SKOTTSBERG, A BOTANICAL SURVEY OF THE FALKLAND ISLANDS. 



surroiinded by the calyx lobes, and the upper half more fleshy (fig. 9 a, b). In fig. 10 

 the berry is seen carrying the dry corolla on the top, as if it were epigynous; this 

 is of course not the case. Bentham and Hooker remark on G. antipoda Först. : 

 »capsule sometimes a bacca in a not-altered calyx, sometimes dry in a calyx flesh}' 

 or dry.» On samples of this plant in Herb. Upsal. I liave observed a fleshy calyx 

 conbined with a berry, just as a G. ynicrophylla. It seems as if the limits between 

 the two genera are very vague indeed; really, a G. antipoda with a berry and a dry 

 calyx is a Pernettya. Further, I have noticed, that in Pernettya pimiila the lower 

 part of the calyx always becomes slightly enlarged (see Pl. II, fig. 11). I do not 

 venture to unite the two genera, which would require a full revision of their species, 

 but I thought it better to publish my observations, as they may prove of some in- 

 terest for botanists dealing with the Ericacese in general. 



In the heath, spread över the islands, but always scarce: E. F., Port Louis! 

 Port Stanley, opposite the town! W. F., Hornby Mountains! Mount Adam! (with 

 Astelia and on top). — Chile, Valdivia — Fuegia, Staten I (only in the rainy zone). 



Pernettya Gaud. 



19. Pernettya puinila (L. fil.) HooK. (Arbutus pumila et empetrifolia L. fil., 

 P. empetrifolia Gaud.) 



Both this and the former are gynodioecious! 



The most common dwarf shrub next to Empetrum. Sometimes viviparous; for 

 particulars see authors paper »Uber Viviparie bei Pernettya^ in Svensk Botanisk Tid- 

 skrift 1912. — Andes of S. Chile and Patagonia, Fuegia, Staten I. 



Primulaceae. 

 Primula L. 



6. Primula magcllaiiica Lehm. (P. farinosa L. v. magellanica HoOK. fil.) 

 In Engler's Pflanzenreich, Pax says that this is perhaps better regarded as a 

 subspecies than as a variety. I have preferred to keep it as a species. It differs 

 considerably in habit, in size and in the colour of the flowers from the Swedish 

 form, the only one of which living material lias been obtainable by me. The corolla 

 tube is of a dull yellowish green, the limb pure white; only sometimes its lower surface 

 shows a faint tinge of lilac. Anthers bright yellow as well as the ring round the base 

 of the limb; style dull green, stigma yellow. 



Reiche states (Fl. de Chile V, p. 93) that he has not been able to discover any 

 dimorphism in the flowers, and I have come to exactly the same result af ter exami- 

 ning a very large number of living flowers from the Falklands; so, as far as we know 

 now, the S. American form is isostylous (Pl. II, Fig. 12). The anthers are fixed on 

 about the same level as the stigma, V2 mm. below to V2 mm. above it. Selfpollination is 



