THE CIIE>riCAL EXAMINATION OF THE MATEKTALR FROM FUNAFUTI. 387 



from top to bottom, and the results of this investigation are also added in a separate 

 report.* 



There is one class of processes, with which geologists are very familiar, as taking 

 place very slowly but very certainly in great rock masses that are permeated and 

 traversed by weak solvents, the action vaguely referred to as " segregation." 



Nothing is more certain than the conclusion that rocks which contain minute 

 quantities of silica, iron disulphide, or the iron, calcium or magnesium carbonates, 

 diffused through their mass, may have those materials collected into nodules and 

 sometimes into masses of considerable size, by this " segregative " action. In many, 

 perhaps in all cases, the centres of this segregative action are determined by the 

 presence, in the first instance, of a somewhat higher percentage of the particular 

 substance at certain spots. Thus, in the chalk, which consists of a mass of calcareous 

 organisms, plant and animal, through which were diffused, at first, a number of 

 skeletons of the organisms that secrete colloid silica — sponge spicules, radiolarian 

 skeletons, and diatomaceous frustules — the colloid silica passes into solution, and 

 appears to be attracted to those portions of the mass in which an abnormally high 

 portion of silica happens to be present, as in the remains of siliceous sponges. In 

 this way the nodular masses of flint are formed, the nodules growing irregularly till 

 all the free silica has been absorbed into them. This silica replaces the calcium 

 carbonate at the spots where it is collected, pseudomorphs in colloid silica being 

 formed of the Glohigerina and other calcareous skeletons which make up the chalk, 

 and the interspaces filled up with the same colloid silica. Ultimately the colloid 

 silica tends to pass into the more stable crystalline form (quartz) and we thus obtain 

 that form of chalcedony — composed of mixtures in varying proportions of quartz and 

 opal (colloid silica) — which we call flint. 



Precisely similar changes take place in sandstones, clays, limestones, &c., in which 

 are diffused small quantities of iron disulphide, and the various carbonates or 

 other mineral substances. A concentration of the material begins, usually, around 

 some fossil, and this goes on till the whole organism is mineralised and is afterwards 

 continued till a nodule, completely enveloping it, is formed. 



We have seen that the leaching-out process in organic structures containing the 

 two carbonates, consequent upon the greater solubility of the calcium carbonate, as 

 compared with the magnesium carbonate, is continually going on in the waters of the 

 ocean. In this way a material is formed in which the percentage of magnesium 

 carbonate may rise up to 16 per cent, or possibly more. Now this mass in a coral 

 reef is everywhere permeated and acted uj)on by sea-water, containing a very notable 

 proportion of magnesium, principally in the condition of chlorides and sulphates. 

 May not these materials, enriched by the magnesium carbonate, exercise an attractive 

 action on the magnesium salts of the ocean waters, giving rise to double decomposition 

 and the gradual replacement of a part of the calcium in the carbonates by magnesium ? 



* See Sectioji XIV (p. 392). 

 3 D 2 



