Thumbnail Image

Theses

Browse

1981 - 19845
Reset filters
Subject: Agricultural Statistics and Informatics × End date: 1984 × Type: Thesis × Thesis Type: M.Sc ×
Reset filters

Search Results

Now showing 1 - 5 of 5
  • Thumbnail Image
    ThesisItemOpen Access
    Study of Genetic Diversity in Desert and Culinary Types of Banana Varieties
    (Department of Statistics,College of Veterinary,Mannuthy, 1981) Mercey, K A; KAU; George, K C
    The data taken from the Bana Research Station, Kannara for 30 culinary varieties for 13 morphological characters and 56 dessert varieties for 12 morphological characters were the base material for this study. Genetic divergences in the varieties were studied using D2-statistic and canonical analysis. The varieties were grouped into clusters by using Tooher’s minimum generalized distance concept. The same clustering pattern was obtained through canonical analysis. In the case of culinary varieties 12 clusters were formed consisting 11 varieties in the first cluster 5 each in second and third and the others were single variety clusters. Whereas the dessert varieties were grouped into 7 clusters 13 varieties in the first cluster 17 in the second cluster, 20 in the third, 2 each in the fourth and fifth and the last two were single variety clusters. The intra and inter cluster distance were diagrammatically represented in a two dimensional space. The scatter diagram showing the mean values of the canonical variates in order to have an idea of the appropriateness of the clustering pattern were also formed in both the type of varieties. In the case of culinary varieties the character bunch weight and in the case of dessert varieties the character finger length was contributing maximum towards divergence. The character girth was contributing minimum towards divergence in both the cases.
  • Thumbnail Image
    ThesisItemOpen Access
    Designs balanced for residual effects
    (Department of Statistics, College of Veterinary and Animal Sciences, Mannuthy, 1984) Sathianandan, T V; KAU; George, K C
    The usual problem in long term experiments is that due to residual effects of treatments. The effect of a treatment that persists for a period after the application of the treatment is referred to as residual effect of that treatment. In the present study an attempt is made to construct designs which will balance for first order residual effects to suit the above mentioned situations. By definition a design is said to be balanced if every treatment follows every other treatment equally frequently. We have established three different methods of construction of such type of designs. The first method of construction is by using cyclic latin squares as in the line of Amble (1977) and we have shown that such an arrangement is balanced for first order residual effects. The second method of construction is based on the set of (v-1) orthogonal latin squares of order v in the case of v treatments. A third method of construction of designs balanced for first order residual effects is also given. This is based on the procedure given by Nair (1967) for the construction of designs balanced for pairs of residual effects. A general intuitive method of analysis is also given.
  • Thumbnail Image
    ThesisItemOpen Access
    Confounded asymetrical factorial designs construction and analysis
    (Department of Statistics, College of Veterinary and Animal Sciences, Mannuthy, 1984) Santy George; KAU; George, K C
    Experiments in agriculture require several types of designs. Situations in which treatments are combinations of factors with asymetrical factors are many. When the total number of treatment combinations is large, confounding is practised In order to get more precise estimates. Confounding is inextricable mixing up interaction effects with block effects. In the present study four different methods of construction of asymmetrical designs are attempted. In the first method polynomials In Galois field ore used for construction. These polynomials ore devised on the basis of two lemmas and following the line of Kichen end Srivastava (1959).
  • Thumbnail Image
    ThesisItemOpen Access
    Study of half-sib correlations and parent-off spring correlations under half-sib mating system
    (Department of Statistics, College of Veterinary and Animal Sciences, Mannuthy, 1983) Shailaja, R; KAU; George, K C
    A purely theoretical investigation entitled ’ a study of half-sib correlations and parent-off spring correlationsi under half-sib mating system was carried out with the following objectives. i) to derive the joint distribution - correlation table as well as the correlation of half-sib pairs under different generations of half-sib mating, in the case of autosomal genes, assuming single locus with two alleles. ii) to derive the joint distribution - correlation table as well as the correlation of parent-off spring pairs under different generations of half-sib mating, in the case of autosomal genes, assuming single locus with two alleles. iii ) to derive the joint distribution - correlation table and the correlation between both the parents and several offspring as well as the correlation between one parent and several offspring under different generations of half-sib mating, in the case of autosomal genes, assuming .single locus with two alleles. iv) to derive the joint distribution - correlation table as well as the correlation of the different ’ half-sib pairs, viz ., brother-sister, brother-brother and sister-sister and also the correlations of different parent-off spring pairs, viz ., mother-daughter, mother-son and father-daughter, in the sex-linked gene case. As the generation matrix of half-sib pairs is too tedious to derive, the Joint distribution of half-sib pairs, both in the autosomal gene case and sex-linked gene case, was derived from fir s t principles and the correlation was calculated from the correlation table, assuming additive genic effects and by using the product-moment correlation coefficient formula. The correlations were worked out for the first ten generations of half-sib mating. The Joint distribution of parent-offspring pairs under half-sib mating system, both in the autosomal gene case and sex-linked gene case was derived by using the generation matrix approach. Here also, the correlations were worked out for the fir s t ten generations of half-sib mating. A comparative study of the half-sib correlations and parent-offspring correlations, conducted both numerically and graphically, revealed that even though the half-sib correlations are of lower order than parent-off spring correlations, the rate of Increase in the former is much higher than the rate of increase in the latter.
  • Thumbnail Image
    ThesisItemOpen Access
    Comparative study of selection indices for the improvement of poultry
    (Department of Statistics, College of Veterinary and Animal Sciences, Mannuthy, 1983) Narayanikutty, N; KAU; George, K C
    Data collected on White Leghorn strain N were entitled for the construction of selection indices. The characters under this study were 20 week and 40 week body weights, egg weight and egg production. From the study of general in d ices constructed for selection among males and females with and without the use of record on dam, the index with the use of record on dam was found to be more efficient than the one without the use of record on dam. Simultaneous selection index considering all the four characters was constructed. 3?be efficiency of this index was tested by finding the percent gain in efficiency due to this index over that due to straight selection . This comparison showed that the index selection was more efficient than straight selection