The aim of this study was to develop and examine the validity of using a phenotype scoring system (PTS), a new concept, in evaluating milk yield and constituents in different goat genotypes (Indigenous, British Alpines, Saanen and Toggenburg) raised in small scale production systems. Strategic decisions of small scale African farmers are mostly based on visual appraisal or body condition scoring (BCS) of their animals. BCS has been highly recommended as a means to evaluate both the energy and the health status of animals, especially in beef farming, but this method has been criticized for being too simple and too subjective because its evaluation is often done too late after the damage has already happened. Phenotype scoring (an approach which includes breed, udder size and BCS of the animal) is presented in this study as a better tool to evaluate milk yield in different goat genotypes raised under free range conditions. This has also been a good opportunity firstly to indicate which, among the three dairy breeds of goat under discussion, can adapt best to the African small scale farming system; secondly to review the relevance of some blood metabolites in characterizing milk production in different goat breeds and thirdly to study the milking capacity of the indigenous compared to the dairy goats raised under small scale production systems in South Africa. Thirty-two goats (8 Indigenous, 8 British Alpines, 8 Saanen and 8 Toggenburg) were raised in a free range system at the ARC-Irene experimental farm close to Pretoria. The experiment was a completely randomized experimental design with eight replicates per treatment group. Blood samples were collected by jugular venipuncture into 10 ml heparinised tubes in the morning before feeding on a weekly basis over a period of two months. Blood plasma was immediately aspirated after centrifugation (3000G), kept on ice and brought to the laboratory for the analysis of glucose, cholesterol, urea nitrogen (BUN) and free fatty acid (FFA) concentrations. Immediately after, all does were entirely milked (followed by 1ml oxytocin IM injection and the kids taken away for a period of four hours) before a second milking session took place to measure the daily milk yield of the does. Milk samples were analyzed for lactose, milk proteins, milk fat, milk urea-nitrogen (MUN) and milk somatic cell count (SCC). In addition body condition score (BCS), age and data related to the goat genotype (breed, udder characteristics) were recorded. Results confirmed that milk yield from dairy goats was higher (p<0.001) than the milk yield of indigenous goats during the entire period of study. Milk lactose values recorded in this study (between 3.9 and 4.9%) were the most stable constituent in goat’s milk. Milk protein concentration (between 3.1 and 4.5%) was significantly higher in the indigenous than in dairy breeds, especially in week one and from week four onwards. Milk fat values (between 3.3 and 7.7%) displayed a decline in all breeds; but as from week three, the fat in milk of indigenous does increased and from week five onwards, it remained significantly higher (p<0.001) compared to that of dairy breeds. In conclusion, the superiority of dairy breeds in milk yield was proven while the quality of indigenous goat milk was recognized. Studies on the characterization of milk production in different breeds revealed that the Toggenburg was superior to the other breeds, followed by the British Alpines and the Saanen; but the British Alpines showed a better adaptability to the environment followed by the Toggenburg and the Saanen. The latter could not produce milk without feed supplementation and lost most body condition as compared to the other breeds. Statistical analyses indicated that breed influenced milk yield, milk fat and the protein content of milk (especially in the Saanen and Toggenburg goats). BCS influenced fat content, lactose, milk proteins, MUN and SCC and also milk yield. Udder size influenced milk proteins and milk yield while udder attachment was associated with milk yield only. These results show that PTS, because it takes into account BCS, breed and udder size, is a better tool for predicting milk yield of goats herded in small scale farming systems. Africans interested in dairy goat farming should adopt PTS as a means to evaluate milk yield especially since milk is sold per volume and not by quality in Africa. Finally, milk from the indigenous goats is superior in terms of lactose, fat and protein content. The latter quality attributes can be used as selection criteria since the milk industry pays premiums for the fat and protein content of milk.