Different substrates were tested, aiming to verify the ones that provided better development conditions for the mandacaru without thorns (Cereus hildemannianus K. Schum.). The work was carried out in the period from January to December 2005 in room temperature at Embrapa Semi-Árido, Petrolina, Pernambuco State, Brazil. A randomized blocks design, in a 5 X 3 factorial with five treatments and four replications, was used. The treatments consisted of five different substrate compositions and three plant types. Bud number, length, weight and dry matter of the flowers, besides weight, length and diameter of the roots were evaluated 360 days after planting. The substrates composed by sand, soil and cattle manure, presented the highest percentages of mandacaru buds growth.

In order to analyze the morphological variability in natural populations, fourteen accesses of Manihot pseudoglaziovii were studied, collected in the Curimataú Paraibano micro-region, in Paraiba State, besides an esculent M. Cranz access (cassava) and a natural hybrid between these two species. Five plants of each access were multiplied through cutting and then cultivated in an experimental area of the PPGZ/CCA/UFPB under standardized conditions, aiming a homogeneous externalization of each genotype. For the morphometric analyses, 20 morphology characters were studied from which analyses of variance and t-test were done, for each character separately. Pearson correlation analyses between pairs of characters were also carried out, once the occurrence of significant correlation justifies the use of multivariate analyses. The cassava access differed regarding the accesses of maniçoba and manipeba, according to the first canonic axis of the multivariate variance analysis. On the other hand, the presumed hybrid between the cassava and maniçoba differed from the others regarding the second canonic axis. The 14 maniçoba accesses presented variation, but they did not differ among themselves, although accesses 1, 2, 8 and 13 revealed themselves distant regarding the other accesses of this species.

This work was conducted in order to evaluate the procedures for analysis of nodule development through non-destructive and destructive methods in the same plant set. A greenhouse experiment was carried out with cowpea grain legume. Plants were weekly measured, from seven days old to physiological maturity, for evaluation of plant height and leaf area, and to obtain digitized images of all the visible part of the rooting systems. Weekly, a plant was also collected from each block for destructive determination of plant and nodule development.Nodules area, length and diameter were measured for those hand-separated from the root system and for those visible on the vessels side. For area, length and diameter of the nodules, initially a escriptive analysis was conducted in each point and next it was determined the average, median, maximum, minimum and the sum for each variable. Both variables, leaf area and plant height, were studied by regression analysis being considered as dependent variables and plant age as independent variable. Independent regressions were conducted for data obtained through the traditional destructive methods. Results have indicated a general concordance for development curves obtained by destructive and non-destructive methods, especially regarding nodule appearing and age in which the maximum value for each variable was found.