Deborah A. Duffield

Date of Award


Document Type


Degree Name

Master of Science (M.S.) in Biology



Physical Description

1 online resource (x, 93 pages)




Captive populations of Asian elephants (Elephas maximus) in North America are not self-sustaining, and increasing reproductive success within captive populations is a high priority. The ability to accurately predict parturition can have a direct impact on elephant welfare. Elephants in captivity often require significant preparation and management throughout the birthing process, and complications during labor and delivery can necessitate immediate intervention, including stillbirth, protracted labor, maternal aggression towards a newborn calf, and dystocia. Being able to predict when parturition will commence can ensure appropriate staff is available and adequate monitoring is performed. Routine endocrine sampling can be used to predict parturition in Asian elephants, with a drop in progesterone (P4) to baseline levels signaling parturition in 2-5 days. However, we determined this method is not without limitations, and it is not used in all institutions that house elephants. As changes in hormones regulate and alter behaviors, we investigated behavioral indicators as an additional management tool for predicting parturition, a time of drastic hormone changes. We conducted a study of five pregnancies in Asian elephants at the Oregon Zoo, U.S.A, and Taronga Zoo, Australia, between 2008 and 2012.

In Chapter 2, I evaluated progesterone (P4) and cortisol levels across three time periods: Baseline; Pre, (the week preceding the drop in P4); and Post, (the period after the P4 drop). Levels of P4 were significantly lower, and levels of cortisol were significantly higher in the days just prior to parturition. I found considerable intra- and inter-individual variation in both endocrine profiles, which can make endocrine assessments difficult to interpret in real time.

In Chapter 3, I investigated whether behaviors in the preparturition period could be predictive of impending parturition in the Asian elephant. ANOVA results indicated a significant difference in the amount of time that elephants spent walking backwards across three time periods (F(2) = 3.723, p = 0.033), with the behavior increasing as parturition approached. These results were supported by a non-parametric Kruskal- Wallis. Using a generalized linear mixed model (GLMM), I found that as P4 levels decrease, walking backwards behavior significantly increases.

In Chapter 4, I evaluated investigative trunk behaviors, or “trunk checks”, directed towards the temporal gland near the ear, mammary glands, vulva and anus of the pregnant dam. Investigative behaviors included both self-directed behaviors and those sent from herd mates towards the pregnant dam. Self-directed behaviors are most likely associated with physical changes in the pregnant dam, such as using the trunk to pull on swollen teats. Other-directed behaviors may stem from chemo-sensory signaling or other types of communication between herd mates, such as detecting changes in progesterone or cortisol. I ran GLMM and found that four trunk-check behaviors varied significantly with P4 and/or cortisol profiles. These were: self-checks of mammary glands increased with decreasing P4 levels; herd-mate-checks of mammary glands increased with decreasing P4 levels; self-checks of vulva increased with decreasing P4 levels and increasing cortisol levels; herd-mate-checks of anus increased with increasing cortisol levels.

In Chapter 5, I evaluated activity budget behaviors in the pregnant elephants. Generalized comparisons were made to published activity budgets of typical captive Asian elephants. I report that activity budgets are within the range of normal activity, though I note a high level of inter-individual variation. In addition, I compared two sampling techniques, including one-zero and instantaneous sampling, that were used for activity budget data collection. I discuss the different results obtained by each sampling technique.

These results are a very promising indication that behaviors, including walking backwards and multiple trunk-check behaviors, are changing over time or with parturition-related hormone profiles. We recommend that keepers, veterinary staff, and other observers that are familiar with the regular behavioral repertoire of a pregnant female should pay close attention to these highlighted behaviors. Keeping track of these behaviors, especially in conjunction with P4 and cortisol tracking, can help staff refine existing windows of expected parturition.

Persistent Identifier

Available for download on Tuesday, December 04, 2018

Included in

Biology Commons