1896. NOTES AND COMMENTS. 85 



rounding off of the contents of the mother-cell. In a species of 

 Coscinodiscus he saw a valve with a new diatom within it, and one with a 

 pair of new diatoms, also the same species with cell-contents rounded 

 off into eight and sixteen portions, and further free packets of eight 

 and of sixteen young diatoms, held together b}^ a fine membrane, as 

 they had doubtless escaped from a parent-cell. Cases like the earlier 

 recorded one of Biddnlphia, where one new individual is produced, 

 appear to be merely a rejuvenescence of the mother-cell. Mr. Murray's 

 observations of preliminary divisions of the contents into eight and 

 sixteen are of far greater interest, suggesting a reproduction by free- 

 cell formation, a process hitherto unrecorded in the family. These 

 discoveries show what a rich area of investigation is open to those 

 who have the opportunity of examining the surface flora of the sea a 

 few miles from the coast. The observations in the present instance 

 were made while on a cruise round the northern coast of our island on 

 behalf of the Fishery Board for Scotland. 



Studies on Indian Garnets. 



In a paper on " the Acicular inclusions in Indian Garnets " 

 {Records Geol. Surv. India, vol. xxix, p. 16), Mr. T. H. Holland investi- 

 gates the hair-like bodies in garnets from southern India, and contests 

 Lacroix's conclusion that they consist of rutile. He regards them as 

 an excellent example of schillerisation, and therefore as of secondary 

 origin. The isotropic character of the garnet in which they lie 

 enables their optical properties to be fully studied. Mr. Holland 

 determines that the crystals are monoclinic, with their principal axes 

 parallel to the edge of the octahedron of the including garnet, their 

 orthopinacoids parallel to the face of the rhombic dodecahedron, and 

 their clinopinacoids parallel to that of the cube. It seems possible 

 that the so-called " faces " of the needles, could they be actually seen 

 and not inferred, would be found to be merely superinduced by the 

 mode of solution of the garnet, the material filling up the negative 

 crystal having become continuously crystalline in each case and 

 giving the optical properties recorded. But the constancy of Mr. 

 Holland's results probably gives him good reason for regarding the 

 long axes of the hairs as true crystallographic axes. The blue quartz, 

 moonstone, and hypersthene, in the same rocks are also schillerised, 

 a fact which strongly supports the author's contention that the 

 asterism of the garnets is similarly due to secondary action. 



In another interesting paper, " On the Origin and Growth of Gar- 

 nets and of their Micropegmatitic Intergrowths in Pyroxenic Rocks '' 

 {op. cit., p. 20), Mr. Holland touches wider ground, and leads us to 

 reconsider our position with regard to the structure of some well- 

 known rocks. All over the world there are masses of granular 

 structure, remarkably similar in their general characters, and contain- 

 ing plagioclastic felspar, rhombic and sometimes monoclinic pyroxene. 



