Registro completo |
Provedor de dados: |
Bragantia
|
País: |
Brazil
|
Título: |
Experiências de cavalos para citrus II
|
Autores: |
Brieger,F. G.
Moreira,Sílvio
|
Data: |
1945-01-01
|
Ano: |
1945
|
Resumo: |
The English resumé is given in a different form from the Portuguese "conclusões". In the former we gave mainly the results which may be of general interest, explaining the tables (quadros) and the procedure of statistical analysis, while in the other the properties of the different rootstocks are discussed in detail since they are of immediate local interest. 1. In a previous publication, Moreira (19) explained the layout and the first results of an experiment, under way since 1936 in the Limeira Exp. Sta., on the influence of twelve different types of Citrus rootstocks on three scion varieties : the oranges "Baianinha" and "Pera" and the grape-fruit "Marsh Seedless". We report here three main results obtained during the first six years of the experiment: a) all plants budded on sour and bitter-sweet oranges (C. aurantium), showed definite signs of the new disease "tristeza" ; b) other rootstocks such as citron (C. medica), ponderosa lemon (C. lemon) had such disadvantageous effect that they could be eliminated as suitable stocks ; c) the data of the first four crops permit to determine certain particularities of some rootstocks varieties used in the experiment. The present paper deals with the complete analysis of the data obtained during the years 1939-1942. This work offered various interesting problems from the point of view of statistical method. The results are of scientific interest, showing very strong interactions between rootstocks and scion in Citrus, besides being of immediate practical importance. The widespread effect of the new disease "tristeza" caused on the Citrus culture in the State of São' Paulo where nearly all sweet oranges, both for local consumption and for export, were budded on sour orange rootstocks, requires immediate action and recommendations for the substitution by other suitable rootstocks. 2. The material used consisted of twelve rootstocks and three scion varieties in all their possible 36 combinations. Scion varieties : "Baianinha" and "Pera" orange, Marsh Seedless grape-fruit. Rootstocks : sour and bitter sweet oranges, "laranja caipira" (sweet orange), "laranja lima" (sweet orange), rangpur lime, sweet lime, rough lemon, ponderosa lemon, citron, grapefruit, "cravo" tangerine and trifoliate orange. 3. The experimental layout for each scion consisted of two conjugated randomized "quadras" or super-blocks, each with six rootstocks and four replications. In one super-block the more important rootstocks were united (quadra A) and in the other less important ones (quadra B). In "quadra" A each individual plot contained 9 trees and in "quadra" B, 4 trees only. Thus we have for the analysis of variance the following decomposition of degrees of freedom : Super- Super- Confounding block A block B A and B Between Superblocks (quadras) ............ 1 Between Rootstocks....................... 5 5 10 Between Replications ..................... 3 3 6 Interaction................................ 15 15 30 Residual (between trees of each plot) ...... 192 72 346 At first the analysis was carried out separately for the super-blocks but since a sufficient degree of uniformity was found, especially with regards to the residual error, "quadras" A and B were confounded. The degree cf freedom for the residual error totalises always slightly less than 364, owing to the death of several trees which thus were eliminated. 4. The following measurements were subjected to the analysis : Circumference of the rootstocks, 20 cm from the ground. Circumference of the zone of union, 30 cm from the ground. Circumference of the scion, 40 cm from the ground. Maximum circumference of the tree. Height of the tree. Number and diameter of fruits. 5. At first a homogeneity test was carried out on a number of seedling trees of each roctstock in order to determine whether they were reasonably homogenous (quadro II). Only a few of the trees were found to be slightly more or less variable than the remainder, with 10 values of standard errors between the 5% and 1% limit of probability. Thus trifoliata orange was slightly less and ponderosa lemon slightly more variable than the other 10 rootstocks. The homogeneity test for the 36 combinations is given in "quadro" IV and as a whole showed satisfactory results. Ponderosa lemon and grapefruit gave an increase in variability for the circumference of the rootstocks, in all the combinations. Citron with grapefruit scion gave a very significant increase of variability, and the combination trifoliata-' 'Baianinha" was less variable than the remainder. In the comparison we did not use only the standard error, but also the variance index (Brieger, 4) since the means were of different magnitude. 6. The results of the analysis of variance are given in detail in "quadro" V and are summarised in "quadro" VI. The latter shows that there is a significant increase of variability between replications, indicating a certain degree of heterogenity in the soil. The variability was however only twice that of the residual error between individual trees. There was no interaction and all combinations behaved uniformily in all replications. The difference between combinations of roctstocks and scions were very considerable and caused an increase of variability of 10 times (7 to 13 times) the residual variation. The difference between superblocks was still bigger and amounted to more than 20 times the residual variability. Since the datas analysed included one measurement of the rootstocks and several ones of the scion, there cannot be any doubt about the existence of a very pronounced physiological interaction between the two components. 7. Thus our next object was to stydy the nature and direction of this interaction and to decide which were the combinations most favorable from a practical point of view. Most of these were carried out using the data of 1942. a) The means were calculated and organized in the order of magnitude. Then consecutive differences were calculated and the correspondent value of t computed. Consequently only positive values of t could appear, and a new table of Brieger (5), for unilateral limits, had to be used. The results are given in the respective columns of "quadros" VII to IX. The values of t do not form a uniform and homogeneous series, but there are always jumps, indicated by large and significant values of t. Comparing different measurements for the same combination of rootstocks and scion, a certain uniform twend can be found, and as itisseen in the tables certain rootstocks had a pronounced tendency to be more favorable than others, in 1942. b) Next an order test was carried out, to compare means of the former years, and the results given in "quadros" X to XII. This test has been developed by Brieger (6) in order to permit a statistical analysis in cases where the rank correlation coefficient cannot be applied. The means are arranged in series according to their magnitude and numbers of order or rank are given to them, both in the series of increasing values (direct order) and of dedreasing values (inverted order). Then the probability is calculated of obtaining such numbers of order as observed or lower numbers still. The values thus calculated are given for the direct order in the column called "maiores" and for the inverse order in the column called "menores". A value of P = 0,1 indicates in the direct order that only once in 1000 there may be obtained a series of means as big or still bigger as the values actually observed, and in the inverse order a value for a mean as small or still smaller. 8. Accepting a 1% limit of probability in our case of comparison between the means in four years, we get the following results. "Baianinha". Tendency for bigger production (number of fruits) is found with sweet lime, rangpur lime and rough lemon. The circumference of the rootstock is bigger with rough lemon, sweet orange (caipira) and sweet lime, and that of the scion is bigger with rough lemon, sweet orange (caipira) and rangpur lime. "Pera". The most productive combinations are with rangpur lime, rough lemon, sweet lime and also ponderosa lemon. The biggest rootstocks are found with rough lemon and sweet orange (caipira), and the biggest scions with rough lemon, rangpur and sweet limes. "Marsh Seedless" grapefruit is most productive with rangpur lime, rough lemon and sweet lime. The biggest rootstocks with rough lemon and sweet orange (caipira) ; the biggest scion with rough lemon, sweet orange (caipira) and sweet lime. 9. Next several special characters were studied which seemed to be important. a) In order to determine whether the union between rootstocks and scion was perfect or whether there was incompatibility between the two, the following t-test was carried out (quadro XXIII) : between the circumference of the rootstocks (20 cm from the ground) and the point of union (30 cm), and between the latter and the circumference of the trunk of the scion (40 cm). In the case of pronounced incompatibility the point of union should be bigger than either of the two other values. In the case of compatibility three forms of the trunk may be observed : it may be either conical and in this case both values should be positive and significant; or it may be cylindrical with both values of t insignificant; or, finally; the rootstocks may be cylindrical and the scion conical, with the first value of t insignificant and the other positive and significant. All combinations on citron as rootstocks showed a very pronounced incompatibility-All cases where sour and bitter-sweet oranges were used, showed abnormal behaviour possible due to the effect of the disease "tristeza". All the other rootstocks-scion combination were compatible and gave conical trunks when combined with "Baianinha", while in the case of "Pera" and "Marsh Seedless" the rootstocks were more frequently cylindrical and the scion conical. b) A detailed study mas carried out with regards to the effect of the rootstocks on the form of the trees (quadros XIV-XVI). Since the t-test did not give sufficiently clear results, an index was calculated by dividing the height of the tree by its diameter. This index should be equal to one in round trees, b'gger than one in cylindrical trees and smaller than one in flat topped trees. "Quadro" XV shows clearly that the rootstocks exerted a strong effect on the form of the trees. A complete analysis was carried out for the three more important rootstocks (quadro XVI) : sweet orange (caipira), rough lemon and sweet lime.. The test within - between proved that the effect of the rootstocks on the tree form are significant. Comparing the means, one can conclude that "caipira" which has a cylindrical form itself (index : 1,87), turns the tops of the scions more cylindrical. Rough lemon which is more nearly round (index: 1,12) has practically no effect on the two oranges scion and flattens the top of the grapefruit tree. Sweet lime with an index of 1,20 reduces somewhat the height of the oranges scion and turns the grapefruit slightly more cylindrical. c) A study of linear correlations was carried out in order to determine whether we may facilitate our work by estimating the relative height and form of trees by measuring only the circumference of the trunck of the scion. The respective values, given in "quadro" XVII show a rather complex result, in consequence of the different nature of the physiological interactions. There is a very pronounced variability of individual values as shown by the X²test. Height and c:rcumference of tree tops and, as whole, circumference of the trunck are pratically not correlated. There is some correlation between the latter and the circumference of the tree top. 10. A full description of the effects of the different rootstocks, given in the portuguese test and the conclusions will not be translated. We shall limit us to only our sugestions as to which of the rootstocks may be considered as the most favorable for the varieties studied : For the "BAIANINHA" orange : sweet orange (caipira), rough lemon and sweet orange (laranja lima). For the "PERA" orange : sweet lime, rangpur lime, and the sweet oranges "lima" and "caipira". For the "MARSH SEEDLESS" grapefruit: rough lemon, sweet oranges "caipira" and "lima".
|
Tipo: |
Info:eu-repo/semantics/article
|
Idioma: |
Português
|
Identificador: |
http://www.scielo.br/scielo.php?script=sci_arttext&pid=S0006-87051945001000001
|
Editor: |
Instituto Agronômico de Campinas
|
Relação: |
10.1590/S0006-87051945001000001
|
Formato: |
text/html
|
Fonte: |
Bragantia v.5 n.10 1945
|
Direitos: |
info:eu-repo/semantics/openAccess
|
|