In the course of the coffee breeding project carried out at the Instituto Agronômico, outstanding progenies have been derived from mother plants selected by either of the two following methods; a) before selection, individual yield is recorded for a number of years begining with the first crop, and b) selection of mother plants is made on the basis of vegetative vigor and the crop they have at the time of selection, regardless of their previous yields. Yield data and other observations presented in this paper refer to a group of progenies derived from plants selected according to the second method. With open pollinated seeds collected the from best plants found in 1943 in a private coffee plantation at Urupês County, (formerly Mundo Novo) State of São Paulo several progenies were planted at the Experiment Stations of Campinas, Pindorama, Jaú and Mococa, with different types of soils. The original coffee plantation of Urupês had been formed with seeds collected in Mineiros do Tietê County, in a farm nearby the place where the so called Sumatra coffee (Coffea arabica var. typica) was first planted after it was privately imported from the island of Sumatra at the end of the last century. This is probably the reason why the Mundo Novo coffee, as it is called nowadays, was formerly named Sumatra de Mundo Novo. From observations on the morphological type of Mundo Novo coffee plants, it was noted that most plants in all progenies and also in private plantations are morphologically similar to the bourbon variety, while a few plants are similar to the typica variety. Artificial crosses between the highest yielding Mundo Novo plants from progenies grown in Campinas and the murta variety (ttNana) grave indication that these plants are of the bourbon type (ttNaNa) and not of the typica type (TTNaNa), as it is the case of the ordinary Sumatra coffee. It has been assumed, therefore, that the Mundo Novo coffee probably originated from a primitive natural cross between the imported Sumatra coffee and the bourbon variety, already existent in the State of São Paulo when the Sumatra coffee was introduced. The progenies of Mundo Novo coffee here studied present a majority of vigorous and productive plants and a few weak individuals with very low productivity. Total yield per plant and per progeny were analized and discussed, for the period from 1946 to 1951. It is apparent that some progenies behave equally well in all locations and that high yield variability between plants is found in all Experiment Stations. The Mundo Novo coffee does not differ from other varieties in the percentages of the commercial seed types, namely, the normal flat beans, the peaberry type, and those resulting from the simultaneous development of two or more ovules in the same fruit locule. It was found, however, that a large percentage of Mundo Novo plants regularly produce an abnormal quantity of fruits with one or two empty locules. This constitutes a defect of the Mundo Novo coffee, which probably can be overcome by selection. It seems that this defect, which is not correlated with yield, is detenruned by genetic factors. Progenies are found where the number of plants showing unusual amounts of empty fruit locules is very low, other plants being normal is this respect. The most promising Mundo Novo plants from several progenies have already been artificially self-pollinated, and the seeds thus obtained were used to plant foundation seed plots. Within a few years seeds of selected Mundo Novo coffee plants will be available to farmers in the State of São Paulo.