I. This paper deals with an experiment carried out to evaluate the effect of sugar cane upper end on the composition of the stalks and juice of sugar cane harvest as a raw material for the sugar industry. The variety studied was CB 41-76. The data were collected from plant cane at intervals of a two weeks, always from the same field, from a small central area of 3.000 square meters approximately, 60 stalks were cut in each occasion, randomly chosen from the whole area. They were afterwards separated into three groups of 20 stalks one for each of the treatments, namely: a) Complete stalk, with no leaves or sheaths. b) Stalks harvested by the technique of REYNOSO, that is, as usually done in practice. c) Stalks with the tops completely cut out, that is, cut by the techinique of REYNOSO and then with 3 other top internodes eliminated. The treatments caused significant differences on the weight of cane and coefficient of purity of juice, but the percentual differences between the average treatments a and c is 13% and 2%, respectively. II. Treatment differences for cane pol, cane fibre, brix, juice pol, reducing sugars, juice ashes, glucose coefficient, saline coefficient and available sucrose (pol) per cent were not significant. III. Time of harvest was an important factor affecting the composition of the cane and the juice. This paper shows that there is no sound basis for the heavy fines applied some sugar mills to planters who do not cut low enough the tops of the cane stalks.