THE FLORA OF THE COAL FORMATION 439 



growtli of Stigmaria in ponds or lakes. It was indeed precisely tlie 

 reverse of this, the Sigillaria growing in a soil more or less swampy 

 but not submerged, and the formation of coal being at last arrested 

 by submergence. I infer this from the circumstance that remains 

 of Cyprids, Fishes, and other aquatic animals, are rarely found in the 

 underclays and lower parts of the coal-beds, but very frequently in 

 the roofs, while it is not unusual to find mineral charcoal more 

 abundant in the lower layers of the coal. For the formation of a 

 bed of coal, the sinking and subsequent burial of an area previously 

 dry seems to have been required. There are a few cases at the 

 Joggins where Calamites and even Sigillarlce seem to have grown on 

 areas liable to frequent inundation ; but in these cases coal did not 

 accvnnulate. The non-laminated, slicken-sided and bleached condition 

 of most of the underclays indicates soils of considerable permanence. 



In regard to beds destitute of Stigmarian underclays, the very 

 few cases of this kind apply only to shaly coals filled with drifted 

 leaves, or to accumulations of vegetable mud capable of conversion 

 into impure coal. The origin of these beds is the same with that of 

 the carbonaceous shales and bituminous limestones already referred 

 to. It will be observed in the section that in a few cases such beds 

 have become sufficiently dry to constitute underclays, and that con- 

 ditions of this kind have sometimes alternated with those favourable 

 to the formation of true coal. 



There are some beds at the Joggins, holding erect trees in situ, 

 which show that SigiUarice sometimes grew singly or in scattei'ed 

 clumps, either alone or amidst brakes of Calamites. In other in- 

 stances they must have grown close together, and with a dense un- 

 dergrowth of ferns and Cordaites, forming an almost impenetrable 

 mass of vegetation. 



From the structure of SigiUarice I infer that, like Cycads, they 

 accumulated large quantities of starch, to be expended at intervals in 

 more rapid growth, or in the production of abundant fructification. 

 I adhere to the belief expressed in previous papers that Brongniart 

 is correct in regarding the SigiUarice as botanically allied to the 

 Cycadacece, and I have recently more fully satisfied myself on tliis 

 point by comparisons of their tissues with those of Cgcas revohda. 

 It is probable, however, that when bettor known they will be 

 found to have a wider range of structure and affinities than we now 

 suppose. 



There are some reasons for believing that the trees described by 

 Corda under the names of Diploxi/lon, Myelopithijs, and Ileterangium, 

 and also the Anahathra of Witliam, are SigiUarice. Mucli of the tissue 



