148 CARL SKOTTSBERG 



adult plant nearly always are peltate. Inflorescences collected Dec. 1916 had 

 lost their petals and stamens. The stigmas were white in these specimens, not 

 crimson as stated for the plants examined before. 

 Area of distribution: Endemic in Masatierra. 



81. G. bracteata Steud. - JOHOW, Estud. 99; SKOTTSBERG, Stud. 10, 

 Fig. 1. Syn. G. insularis Phil, non Schindl. — Fig. 16 b. 



Masatierra: In the central parts of the island, much more rare than the 

 former and belonging to the region of dwarf trees and shrubs along the ridges. 

 Seldom found below 500 m. — El Rabanal (JOHOW); Pico Central (JOHOW); 

 Portezuelo de Villagra, common in the vicinity of the pass, c. 550—600 m 

 (unr.-fr. 3 /i2 16, no. 1; also observed by JOHOW); C. Salsipuedes, scattered in 

 the thickets on the crest, c. 600 m; highest part of O. Villagra, c. 500 m, with 

 an unusually large stem but otherwise typical. — Forma foliis plerumque pel- 

 tatis: C. Centinela, 700 — 800 m, common (fr. ll U 17, no. 610). A remarkable 

 form, found in considerable quantities on the high ridge east of Co Yunque. 

 The leaves generally are peltate; this is, as will be shown below, not uncommon 

 in intermediate forms, explained as hybrids, but no. 610 is a typical bracteata 

 as to all other vegetative characters. Still, it may not be quite pure. 



Area of distribution: Endemic in Masatierra. 



[G. insularis Phil, (see SKOTTSBERG, Stud. 14). Young plants of G. brac- 

 teata, collected by the writer, are identical with PillLlPPl's type of insularis in 

 Herb. Santiago, leg. Germain 1854! 



G. pyramidalis Schindl. (see SKOTTSBERG, Stud. 14). 1 have tried in vain 

 to get hold of the type; I am convinced that it is not a separate species. The 

 type consists of a piece of an old spike of BERTERO no. 1463, without leaves 

 or ligules. It differs in the very small bracts, considerably smaller than in the 

 top region even, of the other species. I don't know if the bracts in the type 

 are intact; I have seen old spikes where they have disappeared, leaving a 

 stump standing. In Herb. Kew I saw such a piece of BERTERO 1463 (this 

 number embraces all the forms of Gunnera from Masatierra); it belongs to G. 

 bracteata.] 



G. bracteata Steud. X peltata Phil. -- Plate 13, 2; text fig. 16 c- i. 



Near Portezuelo JOHOW found a Gunnera with rugose leaves and entire 

 scales, regarded by him as a bastard between the two species. I have found 

 numerous intermediate forms in several places; as hardly one is like the other, 

 I believe they are hybrids and also that the F i-generation is fertile so that 

 we get an F 2 and following generations. As the two species differ in many 

 characters and as there are so many possibilities to combine these, the theoretical 

 result will be a long series of forms uniting the two extremes, and this is 

 exactly what we find. 



The main differences between peltata and bracteata may be summarized 

 as follows. 



G. peltata: Leaves orbicular, reniform of generally peltate, rugose, petiole 

 and lower surface of the blade with scattered tubercles on the larger veins, 



