Improving knowledge for the captive rearing practice of South Island kiwi (Apteryx haastii, A. mantelli ‘Haast’ and A. rowi)

Abstract

Kiwi species, once abundant throughout New Zealand, are now confined to remnant forest patches, national parks and security sites on offshore islands. On average 94% of young kiwi are killed before they reach 100 days old largely due to predation by stoats (Mustela erminea). In order to combat the decline of kiwi, Operation Nest Egg (ONE) began in the mid 1990's; this program involved the removal of wild kiwi eggs to be reared in captivity and then released to the wild as older juveniles with an improved chance of survival. Kiwi species involved in ONE currently include Apteryx mantelli (North Island brown (NIBK)), A. rowi (Okarito Rowi (Rowi)), A. haastii (great spotted kiwi (GSK)) and A. mantelli ‘Haast’ (Haast Tokoeka). Preliminary re-sults and experiences from practitioners, particularly those involved with GSK, suggest that the success in captive rearing seen for NIBK may not be applicable to all kiwi species. Accord-ingly, knowledge gaps existing around rates of failure in hatching and survival of kiwi chicks under captive rearing have led to this research project. In order to fill these knowledge gaps, a comparative approach was taken which involved the collation of all existing captive rearing data for GSK, Rowi and Haast Tokoeka. Direct observation of GSK nesting and nestling behav-iour in a wild situation was also undertaken to make observations that could help refine cap-tive rearing practice. A database was created for all eggs that were received and chicks that hatched through ONE over the past seven years at the two main South Island captive rearing facilities. Generalised Linear Modelling was employed to examine the effects of various factors on the likelihood of eggs hatching, the need for medical attention, and the likelihood of chicks surviving in captivity post-hatch. These models showed that facility and year were important explanatory variables for hatch probability, as well as species and the egg age when uplifted. When data was re-moved from the dataset to look specifically at the species and years that both facilities were operating with/in, both facility and species interacted to have a significant effect on hatcha-bility with the West Coast Wildlife Centre (WCWC) exhibiting higher hatchability than Willow-bank Wildlife Reserve (WWR) and Haast Tokoeka higher hatchability than Rowi. Statistical analyses particularly highlighted aspects of Haast Tokoeka eggs that were different including the fact they appear to be heavier, have lower moisture-loss during the course of incubation, and superior hatchability when compared to GSK and Rowi. Kiwi species and egg age were key determinants for chicks’ survival post-hatch, along with clutch number (i.e. the first or second egg from a given pair). This showed that pre-hatch con-ditions determine chick survivorship post-hatch. Thirty percent of all Haast Tokoeka chicks reared at WWR from 2006-2013 required medical intervention compared to 22% at the WCWC. In contrast, only 9% of Rowi required medical intervention at WWR whereas 20% did at the WCWC. Likelihood of medical intervention was found to increase with increasing num-ber of days keepers assist fed individuals. When compared, the three kiwi species differ significantly in both their pre-hatch and post-hatch conditions. Haast Tokoeka showed the greatest difference to Rowi, exhibiting very high hatchability, but low chick survivorship and a high rate of medical intervention whereas Rowi had moderate hatchability and incredibly high chick survivorship. GSK showed the highest level of problems at hatching. Developing a captive management protocol specific to ONE practice for GSK, Haast Tokoeka and Rowi is pertinent based on the variability in hatching and chick-rearing success seen between the three species, two facilities, and years. Data was deficient on nesting and nestling behaviour of wild GSK and no direct observation of individuals inside the nest throughout the incubation and parental-care stages had been made prior to this research. In this thesis I presented information collected with an infra-red camera inside a nest burrow. Egg turning was found to occur once every three hours on average over the night period (more frequently than what is practiced in captive rearing). The male per-formed nest cleaning duties and remained in the nest for much longer periods than the chick; he continued this pattern of nest attendance for at least 26 days post-hatch.