'* MOLECULAR coalescence/' ETC. 229 



still more tiny granules, possibly demonstrable as dumb-bells 

 by a higher magnifying power. 



2. Solutions of pure oxalic acid were placed in the tubes. 



(a) Oxalic acid was used in large excess of the chloride of 

 calcium. The plug Avas examined at the end of two days, 

 experience having now shown, in the case at least of oxalate 

 of calcium, such an interval to be sufficient. 



Three several lines of forms were here observed (PI. XVI^ 

 figs. 2, 3). 



(1) Beginning at the oxalic end, small spicules of no very 

 definite form were followed through many stages of aggrega- 

 tion to, first, the perfect and, next, the feathered octohedron. 



(2) Beginning also at the oxalic end, small rounded but 

 irregular masses were shaped into thick-waisted, clear- 

 coloured, homogeneous dumb-bells, which became thinner 

 waisted and more characteristic at the point where the per- 

 fect octohedra were found ; underwent sudden enlargement 

 and radial crystallisation "where the octohedron became 

 feathered, and finally assumed the " Avheatsheaf '^ condition. 



(3) At the point where the foregoing series were com- 

 pleted, tiny crystalline bodies appeared, soon seen as they 

 grew to be octohedra much split up into planes parallel with 

 the basial plane (Avhich was oblong), and partly divided by a 

 superficial incision perpendicular to the long axis of the 

 oblong plane (PI. XVI, fig. 4). The full development of this 

 series is seen at PI. XVI, fig. 4, in the shape of very large 

 ovoidmuch facettedand laminated crystals, not unlike the large 

 tablets of PI. XV, fig. 1. We seemed to have here before us 

 the changes leading from the octohedron to the tablet. 

 Here also the " Avhcatsheaves " had become compressed 

 into spheres Avith radiating texture and with rough spiked 

 surface. 



Just at this ]3oint the density of the deposit was so great 

 as to draw a perfectly opaque white line across the middle of 

 the plug, with a comparatively clear stratum beloAV. In most 

 of the plugs such a line existed, and appeared to indicate the 

 point at Avhich the two solutions balanced one another. I 

 shall call it the line of greatest deposit. In this line only 

 crystalline forms occuried, viz., small rhombohedra, large 

 ovoid tablets, and large perfect octohedra. Below the line 

 all the forms rapidly decreased in size ; the ovoid crystals 

 thinned away, by casting off their outer lamincc, till they were 

 reduced to delicate hexagonal plates or lozenges, very trans- 

 parent and symmetrical ; a further simplification of ou.tline 

 by growth of certain sides at the expense of others produced 

 the rhombic form ; and finally; at the calcic extremity, the 



