454 CARNEGIE INSTITUTION OF WASHINGTON. 



ceous dicotyledonous stems are derivatives of woody types of mid to 

 later Mesozoic time, find no support or comfort from either the posi- 

 tive or the negative side of the fossil-plant record. On the contrary, 

 the most generalized of all Mesozoic fossil gymnosperm stems are 

 indisputably the cycadeoids, and these, while including heavier wood 

 types, easily find development far back as highly parenchymatous 

 forms. Moreover, the conifer series from which systematists were but 

 a short time since satisfied to derive the dicotyls are all through the 

 Triassic and down into the Cretaceous characteristically conifers, and 

 as such they change. Undoubtedly some forms closer to Trochoden- 

 dron and Driniys than any yet found must occur, but the point is that 

 the stem record is always a record of extreme parallelism, and so far the 

 herbaceous dicotyl ancestry merely lacks detection. That it was 

 little given to conservation was long since apparent, and the contention 

 may be dismissed that absence from the Jurasso-Cretaceous rocks so 

 far studied is significant. Direction of early dicotyl change, whether 

 from or toward woody types, should eventually be detected from the 

 study of fossils — especially so because of the wide prevalence of types 

 from period to period, even extending to so peculiar an association as 

 that of the curious fern Tempskya and the cycads.^ 



TEMPSKYA. 



That twenty years of search in the Como and Lakota yielded no 

 petrifactions other than cycads and conifers was a great surprise, 

 even a cause of wonder. It seemed that of all the plants which flour- 

 ished with the cycads, not a fruit or a stem aside from the abundant 

 coniferous logs could be found. Had not hundreds of miles of outcrop 

 been gone over repeatedly this might have been attributed to mere lack 

 of observation, while objects so conspicuous as the cycads can not long 

 escape notice. Yet, slowly, a sort of mental picture of the ancient 

 conditions and reasons for this sparsity was formed from sights in the 

 great pure-stand cactus forests of southern Mexico. It was there 

 noted in the dry uplands how, in the short rainy season, muddy tor- 

 rents form in a few minutes in the dry gullies of the cactus country 

 and sweep all surface debris toward the lower lands. Even at night, 

 as lit by incessant flashes of lightning, the old cactus stems could be 

 seen rushing towards the lower lands, where many might be more or 

 less silted in. And this suggests a parallel to the conditions in the Lower 

 Cretaceous desert habitat of the cycads, with its odd and peculiar 

 vegetation. In some such manner the large, compact types, together 

 with the trunks of the fringing forests, were brought into the posi- 

 tions where silting went on, just as one may now see the log rafts 

 forming at the head of Lake Chelan as brought down by the Stehekin 

 River. 



^The distribution of an average type like the Cupressinoxylon from Europe to the Rhaetic (?) 

 of Argentina is, however, no less striking. 



