Mathematical Modeling for Hatchability of Chicken Eggs

Assessment of the probability for failure to hatch, due to either infertility or embryonic mortality during incubation, is important to evaluate reproductive efficiency in poultry. The main objective of this thesis, therefore, was to develop or improve mathematical models that describe probabilities of fertility (the complement of infertility) and embryonic mortality. Appropriate statistical procedures to estimate model parameters were proposed. Data sets used to validate the developed models were of eggs from inseminated hens, broken open to determine fertility; number of spermatozoa captured in the perivitelline layer (Nsperm); and day of mortality during incubation. Three models were developed to assess fertility. First, Nsperm decreased exponentially with time after insemination. Results indicated that Nsperm increased as dose of spermatozoa inseminated was increased four- to eightfold, if an egg was oviposited within one day after the previous egg, and that a relatively large proportion of variation was variation among individual hens. Second, number of spermatozoa that have to be captured to fertilize an oocyte was distributed as log-normal. Results indicated that as time after insemination increased more spermatozoa might have to be captured to fertilize an oocyte. Third, probability of fertility was described by a normal sigmoidal function of time after insemination. To assess the distribution for time of embryonic mortality a model developed previously was improved. Two of the improvements were that the distribution was modeled by a diphasic Weibull distribution, instead of a diphasic logistic distribution and that model parameters were estimated from noncumulative proportions of mortality, instead of cumulative proportions of mortality. Results showed that time of embryonic mortality was influenced by hatch and by number of days that eggs were stored before incubation. To estimate parameters, the method of maximum likelihood was appropriate for most models. The models developed here to describe probabilities of fertility and of embryonic mortality can help to assess factors that affect reproduction, so as to improve reproductive efficiency. These models might be improved further, by including random effects for individual birds. Alternative statistical procedures to estimate parameters for these mixed models, such as Bayesian analysis, should be studied.

Name thesis:degree_name

Ph.D.

Level thesis:degree_level

Dissertation

Discipline thesis:degree_discipline

Animal Sciences

Grantor

University of Illinois at Urbana-Champaign

Year dc:date

2015

Author dc:creator

• Kuurman, Willem Wiecher

Contributors dc:contributor

• Grossman, Michael

• Biology, Biostatistics

Language dc:language

eng

Identifier

(MiAaPQ)AAI3044149

OAI identifier oai:identifier

oai:www.ideals.illinois.edu:2142/83541