KUNGL. SV. VET. AKADEMIENS HANDLINGAR. BAND 57. NIO I. 13 



especially figs. 148 — 150. Kidston (1. c, p. 109) states regarding one of the Scottish spe- 

 cimens referred to Arthrostigma gracile that »at no part does the specimen show any 

 trace of the vascular axis, unless it be found in indications of a ribbon-like band running 

 up the centre of the stem». 



Summary of niorphology. 



The occurrence of a central strand of tracheids proves that Arthrostigma gracile 

 was a vascular plant. The plant is now preserved as band-like impressions, but in the 

 natural state it consisted of cylindrical axes with radially placed leaf- or thorn-like appen- 

 dages. This fact together with the manner of branching and the general habit indicates 

 that Arthrostigma cannot be a rachis of a large divided frond but is of stem-nature. The 

 appendages of these stems can hardly be anything but leaves: though the stems are 

 known from different localities and are of varying thickness, corresponding to thicker 

 stems and narrower branches, there are no traces of any other kind of leaves. Moreover, 

 the appendages in typical specimens show a regular disposition, with distinct parastiches. 

 In other cases, there is an approach to a verticillate arrangement, or the appendages may 

 be very scattered and apparently irregularly placed; they are then generally reduced to 

 short thorn-like projections. The leaves are always remarkably small in respect to the 

 t hick bulk of the stem. The latter had a soft texture with a very narrow central strand 

 of vascular tissue, which consists of scalariform tracheids and appears to form a solid 

 column. 



Distribution and relation to other forms. 



As mentioned above, Arthrostigma gracile was first recorded from Gaspé by Daw- 

 son, who in his first description (1871, p. 41) gives the horizon as Lower Devonian. In 

 the table, p. 85, of the same paper, it is however noted as occurring also in the Middle 

 Devonian. Låter on Dawson (1882 a, p. 104) described additional specimens from the 

 Lower Devonian of the vicinity of Campbellton. The Scottish specimens described by 

 Jack & Etheridge and by Kidston are of about the same age, or from the Lower Old 

 Red Sandstone. 



Among other Lower Devonian plants described under different names there are 

 only some few that can be suspected of being specifically identical with Arthrostigma 

 gracile. As probably identical may be regarded some of the specimens figured by 

 Crépin (1875) under the name Lepidodendron Gaspianum Daws., viz., those in pls. 1 

 and 2. These remains are from beds considered by all Belgian geologists as Lower De- 

 vonian (Crépin, 1. c, p. 7). Another closely related plant is G öppert's Drepanophycus 

 spinaeformis. Tliat the latter should be referred to the same genus seems to be beyond 

 doubt; but the specific identity is doubtful, as none of the specimens from Canada, Scot- 

 land or Norway show the same combination of great size and strongly falcate leaves. 

 Lessonia bohemica and Fucus Novaki Stur (1881) greatly resemble our species, as has 

 already been remarked by Kidston (1. c, p. 110), who does not, however, regard the 

 identity as certain. It should be remarked that figs. 8 and 9 in Stur's pl. 1 show indica- 

 tions of a central strand. The leaves of Stur's specimens, as shown in his figures, are 



