— 303 — 



In Fig. 10 the limits of the annual shoots are indicated by dates; at 

 G is the boundary between the vernal and the summer shoot. 



The vegetative rejuvenating shoots are almost exclusively developed 

 from the upper 1 mm long part of the summer shoot, for which reason 

 the systems of shoots are nearly umbellately branched. 



The sexual dwarf-shoot (Fig. 11) carries: I) at the base 2 bracteolae 

 {a and ß); II) 3 — 4 bracteae (a, b and c); III) a flower beginning with a 

 calyx of 3 leaves (1, 2 and 3). The calyx may be placed in 24 different ways 

 (comp. Figs. 12—15, 17). As to the position of the bracteae I have divided 

 the flowering shoots in 5 groups, of which I have given some diagrams 

 in Figs. 12, 13, 14, 15, 17 and 20. Buchenau has tried to explain the dia- 

 gram Fig. 15 c in the following way: The primary axis of the flowering 

 shoot ends in the bud, K, it carries the leaves cc, ß and a; a supports the 

 flower beginning with the bracteolae b and c. 



This explanation has subsequently been criticised by Eichler, 

 who says that the primary axis cannot be in K, because the leaves of 

 this bud do not in their position follow the position of a, ß and a. Eichler 

 thinks that the flower is supported by the leaf ß; but that also cannot 

 be the case: the leaf a should then stand between the supporting leaf ß 

 and the bracteolae, b and c. If we begin with the more simple positions 

 (Figs. 12, 13, 14) we find the most simple and right explanation: The pri- 

 mary axis of the sexual shoot ends in the flower, and K is a lateral bud. 

 But still the positions of the leaves a, b and c (Fig. 15) would be very strange ; 

 the positions in Fig. 15 must in my opinion be compared with those in 

 Fig. 14. On the basis af Buchenau's explanation Pax has divided the 

 Empetraceae in »2 scharf geschiedene Gruppen«; but this division cannot 

 be maintained, if Buchenau's explanation is not right, because the order 

 of the flowering axis is II (and not III) in Empetrum, Corema and Cera- 

 tiola. On p. 295 another division of the Empetraceae is tried. 



The bud K (Fig. 15) is sometimes developed as a flower having no 

 bracteae. In Fig. 17 I have drawn a diagram of such a flower. Trying to 

 follow the screw-line between the leaves we find it necessary 1) to follow 

 »the long way« between a and ß\ 2) moreover the first leaf in the calyx 

 (1) stands on the back-side of the flower. And it is in these very two 

 characters that the »Zo&eføa-diagram« differs from other diagrams. Un- 

 fortunately no flower with a Lobelia- àmgn&m and only 3 leaves in the calyx 

 is known. But the Rhodoracea, Tripetaleia (Fig. 18) — like Empetrum — 

 has 3 leaves in the coralla, but 5 leaves in the calyx, which has a Lobelia- 

 diagram. Comparing the diagram of Tripetaleia with that of Rhododendron 

 (by Eichler) we find, as to the corolla, that a whorl of 5 leaves is replaced 

 by a whorl of 3 leaves, and the position of the median leaf differs 180°. 

 If in the same manner we imagined a Rhodoracea with only 3 leaves 

 in the calyx, these would have to be placed as the leaves in the calyx of 

 Empetrum and alternating with the leaves S, a and ß (in Fig. 18). In all 

 probability Empetrum has thus a Lobelia-diâgiâm, & character 

 that compels us to seek the nearest relations of the Empetra- 

 ceae among the Rhodoraceae. In 1913 Samuelsson has shown, in 

 an interesting paper, that the development of the flowers in the Empêtra- 



