EDIBLE EARTH FROM FIJI. 225 



At Naitasiri, on the bank of the Rewa river, the Indian coolies 

 who work on the sugar plantations recently quarried out some 

 three quarters of a ton of the earth, which by degrees they ate. 

 In their case it was noticed that those who were afflicted with the 

 small intestinal nematode called Anchylostomum duodenale were 

 almost always geophagists ; but it is uncertain whether the dis- 

 comforts arising out of the presence of this worm in the intestine 

 give origin to a craving which the ingestion of the earth seems to 

 partially satisfy, or whether the habit of earth-eating occurs first 

 and is the means of introducing the ova of the parasite. Moist 

 earth is understood to be the principal habitat of these ova during 

 the extra-corporeal stage of their life-history, and therefore the 

 latter supposition has received, perhaps, the most support. 



The material consists of a very soft, pale pink, clayey material, 

 with small white patches of similar substance and occasional lumps 

 of grey to reddish chalcedony. Its hardness is less than 1 on 

 Moh's scale. The matrix is extremely fine-grained for the most 

 part, but here and there are lumps of angular chalcedony up to 1 J" 

 in diameter. Angular and rounded pyramidal quartz crystals are 

 also fairly numerous, and attain a diameter of from J" to jr". 

 There are also present numerous small and very perfect octahedral 

 crystals of magnetite. Mr. W. G. Woolnough, B.Sc, has estimated 

 the amount of quartz and magnetite crystals present, and has 

 found 5 per cent, quartz crystals and 0*5 per cent, magnetite. 



The rock has suffered so much from decomposition that it is 

 difficult to obtain definite evidence as to whether it represents a 

 decomposed volcanic tuff or a decomposed lava. On the whole it 

 is probable that it represents a decomposed tuff of the nature 

 perhaps of a quartz andesite (dacite). It is even possible that 

 the lava may have been sufficiently acid to justify its being classed 

 as a rhyolite. Mr. E. 0. Andrews, b.a., has already described 

 both rhyolites and andesites from Fiji. 



On the whole, in view of the abundance of magnetite crystals, 

 it is more probable that the rock was originally a dacite. It was 

 probably of tuffaceous origin, and though now, as the subjoined 



O— Dec. 6, 1899. 



