in the drying of a Fluid with Particles in Suspension. 221 



the wave-lines only are strongly marked, they may extend as curving 

 transverse bars ; if the radial streaks are the dominant feature they 

 form radiating curves or feathery or fan structures. A thick mass 

 in its final desiccation may show cracking like the hexagonal flakes 

 of mud at the edge of a pond, or the thick axes of a slowly drying 

 material (like vermilion) may be jointed transversely. Sometimes 

 stellate cracks may form within a mass, as in the example already 

 described (fig. 4). 



4. Possible Applications. 



We have now to consider the possible applications of these simple 

 experiments to structures which occur in Nature. Water, carrying 

 fine suspended material, will often penetrate into rocks along a plane 

 of discontinuity, and drying may then give rise to a " pattern." * 



Istly. Dendritic forms upon joint or other surfaces are generally 

 regarded as the results of crystallisation, and this undoubtedly is 

 often the case, as frost spreads on a window pane. But if the 

 formation takes place under the constraint of a narrow space, it may 

 be caused simply by desiccation in the way described above. It will 

 not be easy in all cases to infer which has been the exact mode of 

 origin ; sometimes the angle at which the branches diverge in the 

 dendritic structure will prove that the formation was governed by 

 crystallising forces ; but it is clear that both conditions often may 

 co-operate.f 



2ndly. If no cavity exists along the plane of weakness, and the 

 coarser axes are adherent and imbedded within the mass, they might 

 appear to represent tubular structures to be, as it were, flattened 

 cylinders. In course of time the molecular and the mineral charac- 

 ter might be changed, and crystallisation even might take place 

 either in the original or the replacing substance. Is it not possible 

 that this may be the explanation of certain puzzling structures 

 sometimes placed as doubtfully organic, such as some of the so-called 

 " fucoid " markings, or the peculiar forms in certain limestones ? J 



3rdly. These last speculations lead us to consider the possibility of 

 similar contractions taking place in space of three dimensions, as we 

 may call it, instead of two in the mass of a rock, instead of along its 

 surface. Thus the principle governing the formation of the land- 

 scape marble, as described by Mr. B. Thompson, presents some 



* In some experiments I used surfaces of rock, and obtained results on a slab of 

 London clay, of slate, &c., only, as might be anticipated, the forms were coarser 

 and less regular than on smooth glass. 



f See note by Professor Bonney appended. 



J This is not intended to refer to the (Krvanella forms described by Mr. 

 Wethered (' Quart. Journ. Geol. Soc.,' 1890 to 1895, vol. 46, p. 270; vol. 47, p. 550 ; 

 vol. 48, p. 377 ; vol. 49, p. 236; vol. 51, p. 196) nor of necessity to Eozoon. 



' Quart. Journ. Qeol. Soc.,' 1894, vol. 50, pp. 393410, 



VOJ, LX1U. B 



