This investigation is concerned with primary variations in the silver content of gold which occurs in hydrothermal deposits, particularly those of hypothermal character which are found in Basement rocks in Southern Rhodesia. The nature of the gold produced by a number of different mines has been studied by reference to production data, and microscope techniques as well as gold and silver assays have been used to determine and to explain the variations in gold fineness. The literature does not contain a great deal of information which is relevant to this topic, but an attempt has been made here to summarize the more important contributions by different writers. From this it emerges that the interpretations given by different investigators are in conflict and that paradoxes may arise when efforts are made to explain observed variations in fineness in terms of certain generalizations which have become entrenched in the literaure. In particular, it is shown that falling temperature alone cannot account for the occurrence of silver-rich gold in certain deposits. The Gwanda district of Southern Rhodesia has been selected as a typical gold belt, and the variation in fineness in 150 producers is described. The deposits are hypothermal in character, and the average fineness of the gold is high but variable, but in a small proportion the fineness is low. It is shown that the nature of the host rock and the distance of a deposit from the granite contact appear to have no influence on the fineness of the gold and that there is no zonal arrangement of fineness values. There is a suggestion that diversity of mineral species in any particular area may be accompanied by rather wide fluctuations in the gold fineness. The variations of fineness in eight typical Southern Rhodesian deposits are studied in detail, by analysis of production data, by assaying specimens of the ore and by the examination of polished specimens of gold-bearing ore. Briefer reference is made to two other deposits in the territory, and to deposits in other countries which appear to bear out the conclusions reached in this section. It emerges that there are two factors which can commonly be correlated with variations in fineness. The first of these is the grade of the ore: high-grade ore generally contains purer gold than low-grade ore. Secondly, the textural evidence indicates that gold which separates relatively early in the paragenesis contains more silver than that which is deposited in the final stages of metallization. A general survey which draws on the literature as well as on the writer's examinations of deposits in the territory indicates that, in general, gold which is associated with late-stage minerals such as tellurides, antimony, bismuth and bismuthinite is silver-poor. Gold associated with galena may be either silver-rich or silver-poor, whereas gold which is of the same age as chalcopyrite or sphalerite is very frequently rich in silver. The difficulty which is encountered in establishing the age of gold which is intimately associated with pyrite and arsenopyrite renders uncertain the correlation between fineness and age of gold in these latter cases. There are, however, indications that gold which is truly contemporaneous with either pyrite or arsenopyrite is silver-rich. In the discussion, the objections to the common practice of singling out temperature as the most potent factor controlling gold fineness are listed. Chief amongst these objections is the fact that gold does not in all deposits increase in fineness with increasing depth: examples are quoted where fineness was found to decrease as deeper levels of the ore body were exploited. It is shown that there is no consistent relationship between the size of gold grains and their silver content. It is the writer's conclusion that in hydrothermal deposits in this territory the high fineness of the gold is due to increasing solubility of silver in the ore fluids in the late states, and that where hydrothermal deposits are characterized by gold with low average fineness, an unusually large proportion of the gold has been deposited early in the paragenotic sequence. In the majority of hypothermal deposits, however, the bulk of tho gold separates late in the sequence and the fineness is accordingly high. It is believed that the relationship which exists between fineness and tenor in many deposits is due to protracted crystallization of gold in those portions of the ore body which remained permeable to the latest stages. These portions of the ore body, which represent either valuable ore shoots or ore shoots in miniature, are likely to contain gold of variable character, but the average silver content will be low because a large proportion of the gold is "late" gold. The factors which might cause epithermal gold to have a lower fineness than mesothermal or hypothermal gold are briefly discussed. Some possible applications of this study are indicated in the final chapter. It is claimed that records of gold fineness might constitute a valuable addition to mill records. Tentative suggestions are made regarding a method whereby the approaching exhaustion of a deposit might in some cases be predicted. With regard to the origin of the gold in the Witwatersrand sediments, it is pointed out that the modified placer hypothesis is not fully equipped to explain certain of the variations in the composition of the gold.