26 consecutive patients with early hydrocephalus (median age at shunting 2.5 months) were followed up to the age of 3 years, prospectively. The work-up included: regular evaluation of psychomotor development and CT/MRI prior and 6 months following surgery. The latter with measurement of the area of the ventricles (degree of hydrocephalus) and the area of the hemispheres in cm(2) (area of hemispheres minus area of ventricles = parenchymal area and degree of brain mass) at the level of the slice with the largest area of the ventricles. 16 (62 %) had normal development (= Group A) and 10 (38 %) moderate or severe retardation (= Group B). The mean ventricular surface was 25.4 cm(2) in Group A and 31.2 cm(2) in Group B prior to surgery (p = 0.1) and 8.9 versus 14.1 cm(2) postsurgery (p = 0.2). The percentage decrease postoperatively was 59.9, versus 57.1 % (p = 0.4). The percentage of ventricular surface in relation to the hemispheres prior to surgery was 42.2 versus 51.4 % (p = 0.1) and post-surgery 12.8 versus 22.4 % (p = 0.1). The mean parenchymal surface was 34.9 cm(2) in Group A and 29.0 cm(2) in Group B prior to surgery (p = 0.3) and 60.9 versus 48.1 cm(2) post surgery (p = 0.07). Percentage increase was 87.3 % versus 77.3 % (p = 0.9). The absolute increase was 26.0 versus 19.0 cm(2) postoperatively (p = 0.2). In contrast to these mean values and their distribution with no significant differences in Group A and B there was some correlation between development and percentage of ventricular surface in relation to the hemispheres prior to surgery: more than 40 % is a critical value, because most of the later retarded children belonged to this radiological subgroup and only half of the later normal patients. After surgery, there was a distinct correlation between developmental outcome and parenchymal surface and its absolute increase in cm(2): most of those who had a normal outcome had a parenchymal surface of more than 40 cm(2) in the third trimenon and an increase of 20 cm(2) and more within six months post surgery, whereas this was the case in only 40 - 50 % of the retarded patients. In most young infants with hydrocephalus there is more brain mass than the ventricular enlargement suggests. The prognostic value of quantitative evaluation of neuroimaging should not be overestimated. Nevertheless, some prognostication in this age group is possible prior to surgery by measurement of ventricular area in relation to the hemispheres and after surgery by measurement of the parenchymal area and its increase in cm(2).