974 



P'O* .... 39.63 ~ 21.26 18.37 



CaO .... 16.76 -^ J6.76 — 



MgO .... 11.68 11.68 



H'O . . . . 31.93 -^ 13.47 18.46 



51.49 48.51 



The percentages of the composition of tnagnesiumphosphate will 

 then be 



P'O* .... 37.88 

 iMgO .... 24.07 

 H*0 .... 38.05 



100.00 

 which fairly correspond with the formula 



HMgPO^ + 37, H'O 

 which requires 



P'O^ .... 39.58 

 i\lgO .... 22.50 

 H'O .... 37.92 



100.00 



In criticizing the above calculation we should bear in mind that 

 a complete correspondence wilh the results of the analysis cannot 

 be expected on account of our ignorance of the influence of small 

 quantities of SiO', APO' and Fe'O'. In |)art at least they originate 

 from the enclosed fragments of woodtissue, which apart from that 

 also contains P'0% CaO and MgO. 



The magnesiumphosphate in the above calculation is unknown 

 in nature'); but there is anotliei' reason to assume that the second 

 substance is not a magnesiumphospliate, but a calcium-magnesium- 

 phosphate, viz. the fact that the original calciumphosphate does not 

 nearly make up half the concretions. It follows then that calcium- 

 oxide must also be present in (he fibrous aggregates. 



The answer to the second question viz. to what the considerable 

 difference in tiie chemical composition of the concretions is to be 

 attributed, is given in the ash-analyses tabulated below; N' VI (of 

 the sapwood) and VII (of the heartwood) we owe to R. Romanis ') 

 and those of VIII to G. Thoms '). 



1) We only have Bobierrite MgS(P0-t)2 -f- 5 H-Q wilh 40.21 P»0\ 34 2^ MgO. 

 25,51 H20 "and Newberyite HMgP0* + 3H«0 with 40,71 P^os, 23,14 MgO, 

 36,15 H^O. 



2) J. S. Gamble. A Manual of Indian Timbers, London 1902, p. 532. 



3) Beitrag zur Kenntniss des Teakholzes, Die Landwirtschaftl. Versuchsstationen 

 23, Berlin 1879, p. 419. 



