BookPDF Available

Improving Australia's Crocodile Industry Productivity -Understanding runtism and survival

Authors:
  • Centre for Crocodile Research, Noonamah, Northern Territory, Australia
  • Berrimah Veterinary Laboratory

Abstract and Figures

Results/key findings From the analyses presented herein, runtism constitutes 49% of deaths followed by deaths for no visible ailments (23%) and disease (12%). There were significant collection area and genetic effects, as well as time of hatch and number of hatchling effects. With the exception of runtism (0.71 SE 0.08), heritability was estimated to be 0.76 (SE 0.09) for all other causes of death using a Pair model due to confounding of the data. Additional data collection will rectify this situation allowing clutch to be included. The heritability estimates from the Animal model varied from 0.28 (SE 0.02) for deaths for no visible reason to 0.60 (SE 0.04) for runting. Crocodile breeding values estimated from these data show considerable variation which will allow producers to start selecting superior, and replacing inferior, animals from the higher risk mortality categories (runtism, no visible ailments and disease-related) to quickly ensure the economic impact of these causes of death are minimised. Many of the findings in the histopathology study were expected due to the emaciated state of the runts that characterises the condition. However, the major findings were the presence of marked lymphoid atrophy, suggesting immunosuppression, and vacuolated adrenocortical cells due to chronic stress.
Content may be subject to copyright.
RIRDC
Innovation for rural Australia
Improving Australias Crocodile
Industry Productivity
— Understanding runtism and survival—
RIRDC Publication No. 09/135
Improving Australia’s
Crocodile Industry Productivity
— Understanding runtism and survival —
Sally Isberg, Cathy Shilton and Peter Thomson
September 2009
RIRDC Publication No 09/135
RIRDC Project No PRJ-000550
© 2009 Rural Industries Research and Development Corporation.
All rights reserved.
ISBN 1 74151 934 9
ISSN 1440-6845
Improving Australia’s Crocodile Industry Productivity— Understanding runtism and survival—
Publication No. 09/135
Project No. PRJ-000550
The information contained in this publication is intended for general use to assist public knowledge and
discussion and to help improve the development of sustainable regions. You must not rely on any information
contained in this publication without taking specialist advice relevant to your particular circumstances.
While reasonable care has been taken in preparing this publication to ensure that information is true and correct,
the Commonwealth of Australia gives no assurance as to the accuracy of any information in this publication.
The Commonwealth of Australia, the Rural Industries Research and Development Corporation (RIRDC), the
authors or contributors expressly disclaim, to the maximum extent permitted by law, all responsibility and liability
to any person, arising directly or indirectly from any act or omission, or for any consequences of any such act or
omission, made in reliance on the contents of this publication, whether or not caused by any negligence on the
part of the Commonwealth of Australia, RIRDC, the authors or contributors.
The Commonwealth of Australia does not necessarily endorse the views in this publication.
This publication is copyright. Apart from any use as permitted under the Copyright Act 1968, all other rights are
reserved. However, wide dissemination is encouraged. Requests and inquiries concerning reproduction and
rights should be addressed to the RIRDC Publications Manager on phone 02 6271 4165.
Researcher Contact Details
Dr Sally Isberg
Porosus Pty Ltd,
PO Box 86,
Palmerston NT 0831
Phone: 08 8988 5554
Fax: 08 8988 2001
Email: sally@crocfarmsnt.com
In submitting this report, the researcher has agreed to RIRDC publishing this material in its edited form.
RIRDC Contact Details
Rural Industries Research and Development Corporation
Level 2, 15 National Circuit
BARTON ACT 2600
PO Box 4776
KINGSTON ACT 2604
Phone: 02 6271 4100
Fax: 02 6271 4199
Email: rirdc@rirdc.gov.au.
Web: http://www.rirdc.gov.au
Electronically published by RIRDC in September 2009
Print-on-demand by Union Offset Printing, Canberra at www.rirdc.gov.au
or phone 1300 634 313
ii
Foreword
The Australian crocodile industry relies on the production of saltwater crocodile skins for the
international skin trade. For the industry to continue to develop and ensure its environmental and
economic sustainability, it not only needs to ensure a reliable supply of hatchlings (wild or captive
egg harvests) but also to ensure maximum survival rates are achieved to meet the end product usage.
The emerging status of this industry means improvements in animal husbandry and a better
understanding of the underlying dynamics of production inefficiencies will ensure the industry meets
this goal. In addition, understanding the dynamics of the underlying causes of production inefficiency,
such as mortality rates, aids in defining research priorities.
Runting causes crocodile mortalities and results reported indicate that wild collection area effects, and
captive breeding genetic effects are highly significant. The initial histopathology results presented
herein indicate that immunosuppression and chronic stress are the most likely cause of runting.
Recommendations are given to continue addressing this area of large economic loss.
This project was funded from RIRDC Core Funds which are provided by the Australian Government.
This report, an addition to RIRDC’s diverse range of over 1900 research publications, forms part of
our New Animal Products R&D program, which aims to accelerate the development of viable new
animal industries.
Most of RIRDC’s publications are available for viewing, downloading or purchasing online at
www.rirdc.gov.au. Purchases can also be made by phoning 1300 634 313.
Peter O’Brien
Managing Director
Rural Industries Research and Development Corporation
iii
Acknowledgments
The outcomes reported in this project arose from a collaborative effort between RIRDC, Darwin
Crocodile Farm (Porosus Pty Ltd) and the University of Sydney. Without the enthusiastic support of
the Porosus Pty Ltd Board of Directors, this project would not have eventuated and the authors wish
to sincerely acknowledge their contribution to this project.
Sincere thanks must also go to the management and staff at Darwin Crocodile Farm for their
assistance with data collection over the last three years. At times, there have been interesting debates
regarding which category of death to put some animals in. The conversations have, to this effect, been
thought-provoking and challenging.
The authors would like to thank the generous donation of in kind work by Berrimah Veterinary
Laboratories and, in particular, Anton Janmaat for originally agreeing to laboratory participation in
the study. The following laboratory staff performed technical aspects of the pathology portion of the
study: Lynne Chambers (haematology, data entry), Sue Aumann (histology processing and serum
biochemistry), Steve Davis (corticosterone assay), Suresh Benedict (bacteriology) and Lois Small
(parasitology).
iv
Abbreviations
Abbreviations used in this study are defined below (in alphabetical order), although they are also
described within the text.
Age Age of animal (days)
ALP Alkaline phosphatase
ALT Alanine amino-transferase
AST Aspartate amino-transferase
BVL Berrimah Veterinary Laboratory
Bwt Bodyweight (g)
CBV Crocodile breeding value
CK Creatine kinase
CI Confidence interval
EDTA Ethylenediaminetetraacetic acid
GGT Gamma glutamyl-transferase
HDays Number of days between hatching date and 1st of January in that particular year
NoHatch Number of live hatchlings in a particular clutch
NVA No visible ailments
PCV Packed-cell volume
SD Standard deviation
SE Standard error
TL Total length (mm)
v
Contents
Foreword.................................................................................................................................................. iii
Acknowledgments ................................................................................................................................... iv
Abbreviations ............................................................................................................................................v
Executive Summary..................................................................................................................................x
1. Introduction...........................................................................................................................................1
Objectives.............................................................................................................................................1
2. Methods and Materials.........................................................................................................................2
2.1 Survival analysis.............................................................................................................................2
2.1.1 Crocodile resources...............................................................................................................2
2.1.2 Collection of mortality data ..................................................................................................2
2.1.3 Statistical analyses ................................................................................................................5
2.2 Histopathology ...............................................................................................................................6
2.2.1 Haematology.........................................................................................................................7
2.2.2 Serum biochemistry ..............................................................................................................7
2.2.3 Corticosterone assay .............................................................................................................7
2.2.4 Bacteriology..........................................................................................................................8
2.2.5 Parasitology ..........................................................................................................................8
2.2.6 Histology...............................................................................................................................8
3. Survival Analysis Results ...................................................................................................................10
3.1 Porosus resource...........................................................................................................................10
3.1.1 Survival analysis results......................................................................................................10
3.1.2 Kaplan-Meier survival functions ........................................................................................15
3.2 Pedigree resource..........................................................................................................................19
3.2.1 Pair model survival analysis results....................................................................................19
3.2.2 Correlation between crocodile breeding values (CBVs).....................................................24
3.2.3 Pair model Kaplan-Meier survival functions......................................................................25
3.2.4 Animal model survival analysis results ..............................................................................28
4. Histopathology Results.......................................................................................................................29
4.1 General findings ...........................................................................................................................29
4.2 Haematology ................................................................................................................................29
4.3 Biochemistry ................................................................................................................................29
4.4 Corticosterone assay.....................................................................................................................30
4.5 Bacteriology .................................................................................................................................30
4.6 Parasitology..................................................................................................................................30
4.7 Histology ......................................................................................................................................31
vi
4.7.1 Lymphoid tissue..................................................................................................................31
4.7.2 Adrenal gland......................................................................................................................31
4.7.3 Bone....................................................................................................................................31
4.7.4 Liver, gastrointestinal tract and pancreas............................................................................32
4.7.5 Yolk sac ..............................................................................................................................32
4.7.6 Other tissues........................................................................................................................33
5. Discussion.............................................................................................................................................34
5.1 Survival analysis...........................................................................................................................34
5.1.1 Collection area and Pair effects ..........................................................................................34
5.1.2 Heritability estimates ..........................................................................................................35
5.1.3 Hatch days (HDays)............................................................................................................36
5.1.4 Number of hatchlings (NoHatch)........................................................................................36
5.2 Histopathology .............................................................................................................................36
6. Implications .........................................................................................................................................39
7. Further research .................................................................................................................................40
8. References............................................................................................................................................41
vii
Tables
Table 1. Summary statistics of nests from different collection areas in the Porosus resource available for
analysis....................................................................................................................................................3
Table 2. Descriptive statistics of the runts and “normals” in each phase of the study..........................................6
Table 3. Summary statistics of the Porosus resource data used in the survival analyses from Darwin
Crocodile Farm.....................................................................................................................................10
Table 4. Total number of deaths in the Porosus resource including a breakdown into the six defined
categories. .............................................................................................................................................10
Table 5. Significance summary for explanatory variates used in the Porosus resource analyses for the
different causes of mortality .................................................................................................................11
Table 6. Estimates (±SE) of year effects for each mortality cause, their hazard ratios and the antilog of the
95% confidence interval (CI) using the Cox proportional hazards model ...........................................12
Table 7. Probability of a crocodile surviving to day 365 and day 1077 for each cause of mortality..................16
Table 8. Summary statistics of Pedigree resource data used in the survival analyses from Darwin Crocodile
Farm......................................................................................................................................................19
Table 9. Total number of deaths in the Pedigree resource and a breakdown into the six defined mortality
categories. .............................................................................................................................................19
Table 10. Significance summary for explanatory variates used in the Pair model analyses for the different
causes of mortality 20
Table 11. Correlation coefficients between crocodiles breeding values (CBVs) for the different causes of
mortality................................................................................................................................................ 25
Table 12. Probability of a crocodile surviving to day 365 and day 1002 for each cause of mortality..................25
Table 13. Significance summary for explanatory variates used in the Animal model analyses for the
different causes of mortality using the Pedigree resource ....................................................................28
Table 14. Crocodile breeding values (CBVs) were offset by a weighted economic value using the
percentage of death in each category (%) multiplied by AU$52.37.....................................................34
viii
Figures
Figure 1. Examples of congenital defects. a) cleft palate; b) weak yolk scar suture, rupturing exposes
internal organs; c) undershot jaw; d) crocodile born with no eyes; e) born with no tail. ..................4
Figure 2. Compared to the same aged “normal” crocodile (below), the runt crocodile (top) appears
emaciated with wasting particularly obvious in the neck and tail areas............................................ 4
Figure 3. Probability of mortality and standard errors for each collection area (Areacode) for overall
survival, congenital defects and runtism .........................................................................................13
Figure 4. Probability of mortality and standard errors for each collection area (Areacode) for the disease-
related, stress-related, no visible ailments (NVA) and management categories of mortality.......... 14
Figure 5. Kaplan-Meier estimated survival functions for crocodiles between hatch and one year of age
(365 days) for each cause of mortality ............................................................................................17
Figure 6. Kaplan-Meier estimated survival functions for crocodiles between hatch and day 1077
(2.95 years) for each cause of mortality. .........................................................................................18
Figure 7. A) Log hazard pair estimates (±SE) of overall juvenile survival in the Pedigree resource. B)
Pair CBVs (±SE) at 365 days using the Cox’s proportional hazards model. ..................................21
Figure 8. Pair CBVs (±SE) at 365 days for the runtism cause of death.......................................................... 23
Figure 9. Pair CBVs (±SE) at 365 days for disease-related deaths.................................................................23
Figure 10. Pair CBVs (±SE) at 365 days for stress-related deaths. ..................................................................24
Figure 11. Pair CBVs (±SE) at 365 days for deaths occurring from no visible ailments (NVA)..................... 24
Figure 12. Kaplan-Meier estimated survival functions for crocodiles in the Pedigree resource between
hatch and one year of age (365 days) for each cause of mortality using the Pair model.................26
Figure 13. Kaplan-Meier estimated survival functions for crocodiles in the Pedigree resource between
hatch and day 1002 (2.75 years) for each cause of mortality using the Pair model. .......................27
Figure 14. Dollar deviation of crocodile breeding values as expressed as a dollar ($) deviation from the
herd average for the runting and NVA causes of death................................................................... 35
ix
Executive Summary
What the report is about and who is the report targeted at?
This project assessed the incidence of different causes of juvenile saltwater crocodile deaths on an
Australian crocodile farm. In addition, a pilot histopathology study was conducted to determine if
there are any primary causes for runting in captive saltwater crocodiles. This information is targeted at
Australian crocodile producers to enhance their production efficiency by reducing juvenile
mortalities, particularly from runting.
Background
Industry standard mortality rates have been accepted to be 10-15% in the first year and 5% thereafter
on Australian crocodile farms. This obviously has a large economic impact on the industry and
decreases the overall level of production efficiency. Previous investigations into captive crocodile
mortality have grouped deaths into one overall encompassing category. However, it was of interest to
know the incidence and trends associated with specific causes of deaths on farms so that management
regimes could be adjusted accordingly. Furthermore, the previous heritability estimate for overall
juvenile survival was 0.15 (SE 0.04). It was of interest to estimate the heritability, and subsequent
breeding values, for the different causes of mortality for incorporation into CrocPLAN.
Anecdotal evidence suggests that runting results in the highest incidence of mortality on crocodile
farms. Runting refers to extremely poor growth in young animals compared to similarly aged
conspecifics. Runts are eventually lost to the industry due to culling or early natural death in a
profoundly wasted state. Little research has been conducted into the reasons why this syndrome
occurs.
Aims/objectives
There were two main objectives of this project. Firstly, to conduct a categorical risk analysis of all
mortality data collected over three years at Darwin Crocodile Farm. This was done using two datasets
i) combined data from both wild and captive-bred crocodiles, and ii) captive-bred crocodiles of known
parentage only. This will allow mortality to be investigated within a non-genetic and genetic
framework, respectively. Estimated breeding values of crocodiles from the latter will be incorporated
into CrocPLAN as separate breeding values rather than the collective “overall” survival. Secondly, a
histopathology study was conducted to examine the issue of runtism and attempt to observe any
differences between “normal” and runt crocodiles. Similar studies have been conducted in other
crocodilians but not for saltwater crocodiles.
Methods used
Both the survival analysis and histopathology study were conducted using animals and data collected
at Darwin Crocodile Farm. Animal ethics approval was obtained from the University of Sydney
(N00/9-2005/3/4204).
Mortality data were collected using clutch and individual scute-cut identification which allowed
clutch of origin, hatch date and parentage/wild nest area to be determined. In addition, the cause of
death was recorded as one of the following categories: congenital defects, runting, disease-related,
stress-related, no visible ailments (NVA: unknown) and management. Three years of data were
collected for analysis using various Cox’s proportional hazards models and adjusted for various
environmental, geographical and genetic effects.
x
The histopathology component of the study was conducted at Berrimah Veterinary Laboratories
(Berrimah, Northern Territory). The study was split into two phases: 2005 and 2007. Each phase
consisted of ten “runts” and ten “normal” crocodiles of similar age. The animals were sacrificed and
subjected to a thorough examination including full post-mortems, general bacterial culture, faecal
parasitology, standard diagnostic haematology and serum biochemistry, histological evaluation of an
extensive range of tissues, and, in the 2007 group, serum corticosterone.
Results/key findings
From the analyses presented herein, runtism constitutes 49% of deaths followed by deaths for no
visible ailments (23%) and disease (12%). There were significant collection area and genetic effects,
as well as time of hatch and number of hatchling effects. With the exception of runtism (0.71 SE
0.08), heritability was estimated to be 0.76 (SE 0.09) for all other causes of death using a Pair model
due to confounding of the data. Additional data collection will rectify this situation allowing clutch to
be included. The heritability estimates from the Animal model varied from 0.28 (SE 0.02) for deaths
for no visible reason to 0.60 (SE 0.04) for runting. Crocodile breeding values estimated from these
data show considerable variation which will allow producers to start selecting superior, and replacing
inferior, animals from the higher risk mortality categories (runtism, no visible ailments and disease-
related) to quickly ensure the economic impact of these causes of death are minimised.
Many of the findings in the histopathology study were expected due to the emaciated state of the runts
that characterises the condition. However, the major findings were the presence of marked lymphoid
atrophy, suggesting immunosuppression, and vacuolated adrenocortical cells due to chronic stress.
Implications for relevant stakeholders and recommendations
Runtism should be set as the number one mortality research priority. Areas of particular research
interest to reduce the incidence of this syndrome, as well as the other categories, are a thorough
investigation of the crocodilian immune system, exploration of potential viral infection(s) and chronic
stress, although other areas should also be explored including alternative pen designs, ethology and
endocrinology.
There are significant geographical effects in the incidence of each cause of death. As a result,
producers should adjust their management policies appropriately for each area. The difference in
crocodile breeding values for the different causes of death will allow producers to select against
higher risk categories when considering the implementation of their genetic improvement programs.
xi
1. Introduction
Juvenile crocodile deaths still remain an area of large economic loss for Australian crocodile
producers. Webb (1989) commented that producers should aim for 95% survival in the first year after
hatch, but in reality, survival rates are typically between 85-90% (Isberg et al. 2004). After the first
year, the risk of mortality decreases and a 95% survival rate is the aim from one year old to slaughter
at about 3.5 years on average (Webb, 1989). There is a large variation in this trait and Isberg et al.
(2004) revealed that the probability of a crocodile surviving to day 400 is only 56%, which is
extremely low. Since every animal is potentially worth in excess of AU$500 at harvest, the economic
loss from mortality and space inefficiency is immense.
Isberg et al. (2004) reported the heritability for the breeding objective, juvenile survival, to be 0.15
(SE 0.04). However, this heritability estimate was based only on whether an animal lived or died and
not why the animal died. Crocodile deaths, as with deaths in any species, can occur for a variety of
reasons and is often ignored when addressing the overall issue of survival (Southey et al. 2004).
Therefore, gains obtained from implementing recommendations based on overall survival may not be
as great compared to those recommendations that consider different causes of death. For this to occur,
the highest risk factors need to be identified.
Runts constitute a large proportion of juvenile deaths on Australian crocodile farms (Hibberd et al.
1996; personal observation) and have been an ensuing problem for producers. Runtism is described as
a condition of hatchling crocodiles whereby they fail to grow in comparison to the rest of their cohort
and generally appear emaciated (anorexic; Huchzermeyer 2003). Buenviaje et al. (1994) suggested
that runting was a failure to adapt to a particular rearing or management environment, whilst Peucker
and Mayer (1995) proposed that the condition is inherited. Mayer (1998) reported that injecting runts
with vitamins or changing the type of food (live worms, tinned cat food) could be potential cures,
although both strategies required further study. In contrast, Anderson et al. (1990) and Kanui et al.
(1993) conducted trials to investigate the effect of human and bovine growth hormone, respectively,
with varying success.
Bacterial hepatitis and septicaemia, caused predominantly by gram negative bacteria, were described
by many authors as a leading cause of death in hatchling saltwater crocodiles (Ladds and Sims 1990,
Buenviaje et al. 1994, Hibberd et al. 1996, Ladds et al. 1996). These were described as opportunistic
infections that predominated in the winter months. However, as the industry has developed
management regimes (provision for heating systems, etc), deaths from bacterial septicaemias have
become less frequent (Buenviaje et al. 1994). A similar situation has also been reported for parasitic
and mycotic infections.
Objectives
There were two major objectives of this study. They were:
a) To evaluate the specific risk factors associated with juvenile mortality, in particular runtism and
disease susceptibility, within both a genetic and non-genetic framework.
b) To conduct a histopathology study to examine the issue of runtism and attempt to observe any
differences between “normal” and “runt” crocodiles.
1
2. Methods and Materials
Darwin Crocodile Farm (Porosus Pty Ltd) has provided data and animal resources essential for the
research reported herein. The initial part of this section, describes the data collection process and
statistical methodology for the survival analysis component of this study, whilst the latter part
describes the sampling strategy and methodology for the histopathology. Animal ethics approval was
obtained from the University of Sydney (N00/9-2005/3/4204).
2.1 Survival analysis
2.1.1 Crocodile resources
Data were collected from all animals hatched at Darwin Crocodile Farm (Noonamah, Northern
Territory, Australia) between 2005 and 2007. The eggs were sourced from either the captive breeding
population or during wild egg collections under the approved Northern Territory management plan
(Department of Natural Resources, Environment, and the Arts). The eggs were incubated on-farm
under standard industry conditions described in Isberg et al. (2004). Upon hatching, the crocodiles
were identified according to their clutch number using the scute marking system described in
Richardson et al. (2002) and Isberg et al. (2004). The crocodiles were then placed into raising pens
and fed standard industry diets according to their size and age class similar to those described by
Isberg et al. (2004).
Two data-sets were created from the data allowing different analyses to be performed. The first
dataset, herein referred to as the Porosus resource, included all progeny records from both wild and
captive nests hatched at Darwin Crocodile Farm between 2005 and 2007 (n = 36,346). For the
purposes of the analysis, all captive nests were allocated into one collective group, whilst the wild
eggs were allocated into twelve separate collection areas according to the landowner (for example,
private cattle station, indigenous community group, etc.; summary statistics are given in Table 1). The
second dataset, herein referred to as the Pedigree resource, includes only the progeny records from 67
known-parent breeding pairs at Darwin Crocodile Farm (Pedigree resource; n = 2,721). This dataset
was used to estimate genetic parameters.
2.1.2 Collection of mortality data
Mortality data were collected in a similar manner described in Isberg et al. (2006) during routine
feeding and cleaning procedures. The dead animal’s clutch of origin was determined from the scute
cuts and used to retrospectively determine the date of hatch (used to calculate age at death) and the
origin of the clutch (captive breeding pen or wild egg collection area). In addition, it was decided to
allocate the cause of each death to one of six categories (described below) so the main risk factors
associated with juvenile crocodile deaths on Australian crocodile farms could be identified.
The categories used to allocate crocodile deaths were congenital defects, runtism, disease-related,
stress-related, no visible ailments and management-related. Further descriptions of these are given
below.
Congenital defects- Deaths related to this category were generally seen immediately upon hatch and
include defects such as unabsorbed yolk sacs, weak yolk scar sutures, jaw deformities (cleft palate,
under-shot or over-shot jaws), spinal deformities, tail deformities (no tail, partial tail missing, “curly”
tails), and any other gross deformity. These deaths generally occur within the first month of life.
Figure 1 shows some examples of defects that would be classified as congenital.
2
Table 1. Summary statistics of nests from different collection areas in the Porosus resource
available for analysis. n is the total number of live hatchlings from each area in the particular
year, NoHatch is the number of live hatchlings from each clutch put to farm.
2005 2006 2007
Area n Av. NoHatch
± SD n Av. NoHatch
± SD n Av. NoHatch
± SD
1 2750 32.76 ± 11.07 2323 29.57 ± 10.20 2729 27.37 ± 10.16
2 0 - 1508 35.23 ± 11.68 787 32.27 ± 11.61
3 2868 36.77 ± 11.70 2661 37.84 ± 11.14 3012 38.35 ± 11.03
4 263 28.67 ± 9.61 180 41.08 ± 11.68 469 36.21 ± 11.84
5 12 12 ± 0 110 35.76 ± 16.89 18 11.78 ± 4.28
6 69 26.86 ± 7.36 0 - 0 -
7 1533 38.67 ± 10.41 1265 41.25 ± 11.45 250 39.58 ± 10.45
8 0 - 84 42.02 ± 1.00 170 38.49 ± 9.65
9 65 32.32 ± 9.05 465 37.17 ± 14.95 1111 38.77 ± 12.26
10 2313 35.87 ± 11.02 1621 34.51 ± 10.47 1641 35.22 ± 12.40
11 556 35.43 ± 8.62 1199 35.46 ± 10.92 1237 35.26 ± 11.33
12 1192 38.85 ± 14.14 969 38.42 ± 12.65 231 35.38 ± 9.18
13 0 - 0 - 685 43.51 ± 15.00
Runtism was defined by an emaciated, non-thriving animal in comparison to others of a similar age
(Huchzermeyer, 2003), shown in Figure 2.
Disease-related was determined following a pathological investigation at Berrimah Veterinary
Laboratory (BVL; Northern Territory Department of Primary Industries, Fisheries and Mines). These
were considered independent to the stress-related deaths described below. When appropriate,
antibiotics were administered as determined by antibiotic sensitivity testing at BVL.
Stress-related is defined when deaths occurred within a short time period after a management-induced
stress event. These include minimising size variation within pens (grading), moving animals between
pens, hot water services failing or pens left without water. In the majority of these cases, animals sent
to BVL returned positive septicaemia pathology results and antibiotic treatment followed as
appropriate.
No visible ailments (NVA) - Deaths were allocated to this category when neither a disease outbreak
nor stress incident was noted. These deaths usually occur randomly in pens with no distinct trend in
mortalities.
Management is any other event that does not fit into the above categories. An example is an injury
event.
When a crocodile died, it was denoted a one (1) in the appropriate category. However, in each year
cohort, there were crocodiles that were still in the production system when this study period
concluded (31st December, 2007). These animals were included in the study as censored records
(coded as 0) to account for the study period ending before mortality could be observed (Southey et al.
2001). In addition, to maintain data integrity, there were some observations that were omitted as their
scute cuts corresponded to nests that had zero hatchlings, either their hatch date or their death date
were not recorded so their age at death could not be calculated or their scute cuts were not recorded.
3
Figure 1. Examples of congenital defects. a) cleft palate; b) weak yolk scar suture, rupturing
exposes internal organs; c) undershot jaw; d) crocodile born with no eyes; e) born with no tail.
e)
d)
c)
b)
a)
Figure 2. Compared to the same aged “normal” crocodile (below), the runt crocodile (top)
appears emaciated with wasting particularly obvious in the neck and tail areas.
4
2.1.3 Statistical analyses
The survival time data were analysed using a Cox’s proportional hazards model in Survival Kit V3.12
(Ducrocq and Sölkner 1994; 1998) to identify risk factors. In addition, the data were analysed using a
competing risk approach whereby different hazards of mortality could be assigned for the six different
causes of mortality (Southey et al. 2004) described in Section 2.1.2. All categories were assumed to
be independent. Furthermore, different models were used to analyse the two different datasets
(Porosus resource and Pedigree resource) and a 5% significance level was chosen to evaluate
explanatory variables by backward elimination.
2.1.3.1 Porosus model
The model used for the Porosus data was specified as
ln[hijk(t)] = ln[h0(t)] + (βHDHDaysjk + βNoNoHatchjk + Yeark + Areacodej)
where hijk(t) is the hazard function for the ith individual from the jth areacode in the kth year at time t,
h0(t) is the unspecified baseline hazard function, HDaysjk is the number of days between hatching date
and the 1st of January in that particular year for an individual from the jth areacode in the kth year; βHD
is the regression coefficient for HDays; NoHatchjk is the number of live hatchlings in a particular
clutch from the jth areacode in the kth year; βNo is the regression coefficient for NoHatch; Yeark is the
fixed effect of the kth year (k = 2005, 2006, 2007); and Areacodej is the fixed effect of the ith area of
collection (i = 1,….,13).
2.1.3.2 Pedigree model
Two models were used to analyse the Pedigree resource to obtain estimates of variance components
for heritability and breeding value estimation as follows.
1. Pair model
ln[hijk(t)] = ln[h0(t)] + (βHDHDaysjk + βNoNoHatchjk + Yeark + Pairj + Clutchjk)
2. Animal model
ln[hijk(t)] = ln[h0(t)] + (βHDHDaysjk + βNoNoHatchjk + Yeark + Animali)
where hijk(t) is the hazard function for the ith individual from the jth pair in the kth year at time t, h0(t) is
the unspecified baseline hazard function, HDaysjk is the number of days between hatching date and
the 1st of January in that particular year for an individual from the jth pair in the kth year; βHD is the
regression coefficient for HDays; NoHatchjk is the number of live hatchlings in a particular clutch
from the jth pair in the kth year; βNo is the regression coefficient for NoHatch; Yeark is the fixed effect
of the kth year (k = 2005, 2006, 2007); Pairj is the random effect of pair (assumed N(0,σ2Pair)); Clutchjk
is the common environment (random) effect of a clutch produced by the jth pair in the kth year
(assumed N(0,σ2Clutch)); and Animali is the random effect of the ith individual (assumed N(0,σ2Animal)).
5
The Pair model log-survival heritability estimates were calculated as
6
2
Clutch
2
Pair
2
Pair
2
logt 2
h
π
σσ
σ
++
=
using the estimates of the variance component, and (Isberg et al. 2004), whilst the
animal model estimates were calculated as
2
Pair
σ
2
Clutch
σ
6
2
Animal
2
Animal
2
logt 2
h
π
σ
σ
+
=
using the variance component estimate, .
2
Animal
σ
2.2 Histopathology
Forty crocodiles were sacrificed from Darwin Crocodile Farm for the histopathology component of
this study. The study was split into two phases. Phase 1 was conducted in November 2005 and Phase
2 in July 2007. Phase 1 of the study was a pilot phase which involved a wide range of standard
veterinary pathological procedures. Phase 2 was designed to target data collection to specific
parameters that were identified in Phase 1 as possible differences in runts compared to “normal”
crocodiles.
In both Phases, ten runt crocodiles and ten normal crocodiles were randomly sampled over the period
of one week (Table 2). Crocodiles were fasted for 72 hours prior to sampling to remove any effect of
recent feeding on blood or tissue parameters. Crocodiles were randomly selected from several pens.
The inclusion of ten normal crocodiles from the same cohort in each phase of the project was to have
a control group reared under the exact same conditions to allow direct comparison with runts and
facilitate interpretation of results. This was necessary since baseline or “normal” clinical pathological
parameters and the histological appearance of tissues are not well documented in crocodiles of this
species and age group.
Table 2. Descriptive statistics of the runts and “normals” in each phase of the study.
Year Status Av. Age (SD) Av. TL (SD) Av. Bwt (SD)
Runt 197 (36.44) 35.41 (1.41) 74.4 (24.06)
2005 Normal 225 (15.15) 56.02 (4.73) 454.9 (127.44)
Runt 124.7 (11.61) 34.35 (1.47) 64.3 (6.48)
2007 Normal 118.4 (2.46) 48.35 (3.77) 286.3 (76.86)
During Phase 1, crocodiles were transported from the farm to the laboratory where blood sampling,
euthanasia and post-mortems were conducted over the ensuing four hours. For Phase 2, blood
sampling and euthanasia occurred at the farm, immediately after removing the crocodile from the pen,
with subsequent post-mortem sampling occurring over the next four hours after transport to the
laboratory. Total length and body weight were recorded for each crocodile (Table 2).
All crocodiles were blood sampled from the cervical sinus using the technique described by Lloyd and
Morris (1999). The initial 0.5 ml of blood was reserved for haematological study and the remainder of
the blood sample used for biochemical analyses. Crocodiles were humanely euthanised immediately
following blood sampling with a lethal intravenous dose of pentobarbitone sodium into the ventral tail
vein.
6
In both phases, a full gross necropsy was performed on each crocodile, taking note of any grossly
evident abnormalities or differences between organs and tissues of runts and normals that may signify
a problem with a particular tissue or organ system. In Phase 1, samples were aseptically obtained for
bacterial culture from each crocodile and faeces was collected from the colon for parasitological
study. In both phases, the carcass was fixed in 10% neutral buffered formalin for histological
processing.
2.2.1 Haematology
Routine diagnostic veterinary haematology was used to evaluate the red and white blood cell
components of the circulating blood. These components can provide information on general health
status, bone marrow function and whether the animal may be suffering from infection.
Haematological investigations were performed on all crocodiles in both Phases of the study using
blood anticoagulated with ethylene diamine tetra-acetic acid (EDTA). Packed-cell volume (PCV),
which measures the ratio of red blood cells to total blood volume, was determined by centrifugation of
blood in microhaematocrit tubes. The differential white blood cell count was made from a direct
smear of the blood using a fast Wright’s-Giemsa type stain (Diff Quik, Lab Aids Pty. Ltd., Narrabeen,
NSW, Australia). The total white blood cell count was determined by diluting 25 µL of blood in a
1:32 ratio with phloxine B stain solution, counting the number of eosinophils and heterophils in nine
large squares of a standard haemocytometer and using the differential count to calculate the total
number of white blood cells. The percentages of lymphocytes and monocytes from the differential
count and the total white blood cell count were then used to calculate the absolute numbers of
lymphocytes and monocytes in the blood.
2.2.2 Serum biochemistry
A wide range of serum biochemical analyses were performed on all crocodiles in both Phases.
Analyses included evaluation of electrolyte status (sodium, chloride and potassium), which provides
information primarily on nutrition, hydration status and organ (primarily kidney) function. Function
and evidence for necrosis of the liver was investigated by measuring tissue enzymes (alanine amino-
transferase (ALT), aspartate amino-transferase (AST), alkaline phosphatase (ALP) and gamma
glutamyl-transferase (GGT)). Creatine kinase (CK) was measured as an indicator of muscle damage
and to facilitate interpretation of liver enzymes, some of which may also be produced in muscle.
Serum minerals (calcium and phosphorus) were measured since they are required at certain
concentrations for proper bone formation, in addition to providing pertinent information on kidney
function in reptiles. Serum proteins were evaluated to assess the ability of the liver to produce protein
(albumin), nutritional status (albumin) and immune status (globulins). Total serum iron was measured
to investigate possible reasons for anaemia. Uric acid level, and to a lesser degree creatinine and urea,
may reflect kidney function in reptiles, and glucose provides an indication of nutritional status, liver
and pancreatic function. All analyses with the exception of globulin were performed on an automated
biochemistry analyser (Konelab 20i, Thermos Electron Corporation, Vantaa, Finland). Serum was
harvested from blood that was allowed to clot in plain blood tubes left at room temperature for two to
four hours. Globulin concentrations were calculated by subtracting albumin concentration from total
protein concentration.
2.2.3 Corticosterone assay
In Phase 2, serum corticosterone, the major stress hormone in reptiles, was measured using a
corticosterone kit test according to the manufacturer’s directions (Corticosterone HS EIA, IDS Ltd.,
Boldon, U.K.). In addition to the 20 crocodiles in the main Phase 2 sample, an additional seven
normal crocodiles and three runts were sampled from the same pens and during the same period, in
order to increase sample size for this assay. Where the serum corticosterone exceeded the upper limit
of the assay range, the result used was the highest detectable value of the kit (20 ng/ml). In order to
reflect the ongoing background stress level exhibited by the crocodiles and minimise elevation of
corticosterone due to prolonged handling and transport following removal from the pen, the assay was
7
performed using serum harvested from blood collected at the farm immediately following removal of
the crocodile from its usual pen.
2.2.4 Bacteriology
General bacterial culture of two filtering organs (liver and spleen) was conducted in Phase 1 to
investigate the possibility that runt crocodiles were more likely to have bacterial infection compared
to normal crocodiles. The liver and spleen were aseptically removed at the beginning of the post-
mortem for culture in all crocodiles. Additionally, sterile swabs for culture were used to sample the
contents of enlarged yolk sacs noted in two crocodiles (one runt and one normal). Techniques for the
culture of tissues involved aseptically homogenising the tissue, then applying the material to a sterile
swab. Swabs were used to inoculate tryptic soy agar with sheep blood and MacConkey agar with
crystal violet (Oxoid Australia Pty Ltd., Thebarton, South Australia). Agar plates were examined for
colonies after 24 and 48 hours incubation at 35ºC. Gram negative bacteria were speciated using
Microbact biochemical strips (Oxoid Ltd., Hants, UK).
2.2.5 Parasitology
To investigate the possibility that intestinal parasites are associated with runting, faecal flotations
were performed on all crocodiles in Phase 1. The technique involved collection of all faeces in the
colon of the crocodile, emulsification in a zinc sulphate solution, centrifugation and microscopic
examination of both the surface layer and sediment. This technique will reveal significant numbers of
nematode, trematode, cestode or pentastomid eggs, as well as coccidial oocysts.
2.2.6 Histology
Histological examination allowed a detailed examination of the microscopic architecture and cellular
morphology of tissues, and is much more sensitive than gross examination (i.e. with the unaided eye)
in the detection of abnormalities. In Phase 1, a complete range of tissues, encompassing all organ
systems, was examined histologically (see below). In Phase 2, selected tissues were examined that
were noted in Phase 1 being distinct between runt and normal crocodiles. For preparation of
histological slides, tissues that had been fixed in 10% neutral buffered formalin were trimmed and
placed in cassettes for routine histological processing. Tissues were sectioned at 5 µm and stained
with haematoxylin and eosin, a routine stain for histological examination of tissues.
Organs/tissues examined included heart with large vessels at the heart base, lung, trachea, kidney,
liver, oesophagus, stomach, duodenum, jejunum, large intestine, yolk sac remnant, pancreas, fat body,
spleen, tonsil, thymus, thyroid, adrenal and pituitary glands, brain, spinal cord, eye, skin, skeletal
muscle, bone, joint and bone marrow. The specific tissue orientation, region of an organ and size of
section were made as uniform as possible to maximise the ability to compare the tissue among
individuals and between runt and normal crocodiles. For example, sagittal sections of the heart,
incorporating the apex of the ventricle, atrium and great vessels at the base of the heart, and a
complete sagittal section of the brain. The pancreas was sectioned in the mid-region where lobes are
intermingled with duodenal loops, and intestinal segments were taken at approximately the same level
along the intestine and villus height was compared to width at the base of villi as a means of assessing
possible villus atrophy in runts. For assessment of the growth plate, bone marrow and a synovial joint,
a sagittal section of the distal femur to the proximal tibia was used. For the thymus, a routine
transverse section was made incorporating the tissues running down the ventral neck (trachea,
oesophagus, blood vessels, thymus and surrounding fibrous connective tissue) at the level of the
bifurcation of the trachea, which is the usual location of the bulk of the thymic tissue. Where thymic
lobes appeared reduced or absent in the initial section, additional sections in the same vicinity were
made to ensure an accurate histological picture of the thymus was being obtained. In most cases, all
organs listed above were examined in all crocodiles in Phase 1, although in some instances, small
8
organs, such as pituitary glands, were missed in individuals. Crocodiles were sexed by histological
examination of the gonads.
Since histological examination of tissues is a subjective procedure, an attempt to control bias was
made by not labelling slides as to whether the crocodile was a runt or normal.
9
3. Survival Analysis Results
3.1 Porosus resource
A total of 36,346 crocodiles hatched between 2005 and 2007 and were used in the overall survival
analysis. These animals were allocated into 13 areas (Area 1 is all captive nests and 2-13 are wild nest
areas) depending on their clutch of origin. There was a total of 5,043 crocodile deaths until the last
day of the study (31st December, 2007) leaving 31,303 (63.13%) animals still in the production
system. Dummy records were created for these animals and right censored with an average censoring
time of 649 days. Less than one per cent of total records needed to be omitted for data integrity
purposes. A summary of these data is shown in Table 3, whilst a summary of the number of deaths
within each category are given in Table 4.
Table 3. Summary statistics of the Porosus resource data used in the survival analyses from
Darwin Crocodile Farm. Right censored records were animals that were assumed to still be in
the production system at the termination of the study period (31st December, 2007). Some
observations were omitted due to data integrity concerns.
Year Total
hatchlings Total no.
deaths No. right
censored records Av. censor age
(± st. dev.) No. omitted
observations
2005 11,621 1,847 9,774 1020 (24) 134
2006 12,385 1,556 10,829 641 (28) 170
2007 12,340 1,640 10,700 287 (23) 53
TOTAL 36,346 5,043 31,303 357
Table 4. Total number of deaths in the Porosus resource including a breakdown into the six
defined categories.
Reason for death
Year Total no.
deaths Congenital
defects Runt
Disease-
related Stress-
related No visible
ailments Managemen
t
2005 1,847 25 855 158 184 572 53
2006 1,556 31 408 392 80 386 259
2007 1,640 106 1,191 34 91 194 24
TOTAL 5,043 162 2,454 584 355 1,152 336
3.1.1 Survival analysis results
A summary of the results for each cause of mortality from the Porosus resource is given below. A
summary of the significant explanatory variables is provided in Table 5. With the exception of
disease-related deaths, the regression coefficients for HDays were all positive indicating that for each
day later a clutch hatches, the risk of mortality also increases. The regression coefficients for number
of hatchlings from each clutch (NoHatch) for each cause of mortality was negative indicating that the
greater number of hatchlings to hatch from a clutch, the lower the risk of mortality. Year (Tables 5
and 6) and Areacode (Table 5) were significant for each cause of mortality. Figure 3 shows the
different Areacode hazards of mortality for overall survival, congenital defects and runtism, whilst
Figure 4 shows the hazards for the disease-related, stress-related, no visible ailments and management
death categories. For comparative purposes, Area 1 (captive nests) was used as the baseline area.
10
Table 5. Significance summary for explanatory variates used in the Porosus resource analyses
for the different causes of mortality. Regression coefficients (SE) on the log-hazard scale are
given for the significant (p=0.000) HDays and NoHatch terms. A 9 indicates if the term was
significant for Year or Areacode. 8 indicates the term was non-significant.
HDays NoHatch Year Areacode
Overall 4.40×10-3 (5.57×10-4) -2.23×10-2 (1.20×10-3) 9 9
Congenital 4.48×10-4 (3.14×10-3) -5.22×10-2 (6.27×10-3) 9 9
Runt 1.09×10-2 (7.85×10-4) -2.32×10-2 (1.72×10-3) 9 9
Disease-related -1.17×10-2 (1.77×10-3) -2.06×10-2 (3.54×10-3) 9 9
Stress-related 8 -1.46×10-2 (4.60×10-3) 9 9
NVA 8 -2.65×10-2 (2.50×10-3) 9 9
Management 8 8 9 9
3.1.1.1 Overall Porosus survival analysis
This analysis was conducted using the data without distinguishing the cause of crocodile mortality.
Therefore, there were 5,043 crocodile deaths and 31,303 censored records available for analysis. All
explanatory variables were significant (HDays, NoHatch, Year and Areacode; Table 5).
The antilog of the estimate for HDays (Table 5) is 1.004405 indicating that for each day later a clutch
hatches, the hazard of mortality increases by 0.44%. This is further exemplified when calculated on a
weekly basis whereby the hazard of mortality is increased by 3.12% for every week later a clutch
hatches. Furthermore, the number of hatchlings that result from a clutch (NoHatch) is also significant
in predicting the hazard of mortality. For each additional hatchling, the risk of mortality decreases by
2.25% (antilog of -0.223).
Of the three years of data analysed for overall survival, 2005 had the greatest number of deaths
observed (n = 1947) and was used as a base to compare the other years. Table 6 shows the regression
coefficients (±SE), hazard ratios (exponentiated coefficients) and the antilog of the 95% confidence
interval for the year effects expressed as a deviation from the base year (in this case 2005). Compared
to 2005, the hatchlings in 2006 had a 12% lower hazard of mortality, whilst 2007 had a 29% greater
hazard.
Figure 3 shows the relative risk ratio of each collection area (Areacode) as a deviation of Area 1
(captive nests at Darwin Crocodile Farm). Areas 2 (24%), 10 (13%) and 12 (26%) had significantly
higher risks of mortality, whilst Areas 3 (30%) and 7 (18%) had significantly lower risks of overall
mortality. Areas, 4-6, 8-9, 11 and 13 were not significantly different from Area 1 (p>0.05) mainly due
to their low number of uncensored records.
3.1.1.2 Congenital defects
There were 162 deaths in the congenital defects category with an average failure time of 27.77 days
(min-max: 1 - 515days). All explanatory variates were significant (Table 5). For every day later a
clutch hatches, the hazard of mortality increases by 0.04% (0.31% per week), whilst for every
additional hatchling produced from a clutch, the risk of mortality decreases by 5.08%.
For this analysis, 2007 was used as the base year for comparison as it had the greatest number of
uncensored deaths (Table 6). The reason for the increased number of congenital deaths in 2007 was
most likely a management decision to give more animals the opportunity to survive rather than any
actual year effect. In comparison, 2005 and 2006 had significantly lower hazards of 72% and 65%,
respectively.
11
Only Areas 9 and 13 were significantly different to Area 1, with increased hazards of 46% and 92%,
respectively. All other areas were non-significantly different (p>0.05) due to the low number of
observed failures in this category. Area 6 had no deaths in this category (Figure 3).
Table 6. Estimates (±SE) of year effects for each mortality cause, their hazard ratios and the
antilog of the 95% confidence interval (CI) using the Cox proportional hazards model. The
hazard ratio is the antilog of the estimate and represents the risk of mortality. All are
expressed as ratios relative to the year with the greatest number of observed deaths.
95% CI
Year Estimate ± SE Hazard Ratio Lower Upper
Overall survival
2005 - 1.00 - -
2006 -0.13 ± 0.04 0.88 0.81 0.95
2007 0.26 ± 0.04 1.29 1.20 1.40
Congenital defects
2005 -1.26 ± 0.24 0.28 0.18 0.45
2006 -1.04 ± 0.22 0.35 0.23 0.54
2007 - 1.00 - -
Runtism
2005 -0.67 ± 0.05 0.51 0.46 0.57
2006 -1.49 ± 0.06 0.23 0.20 0.25
2007 - 1.00 - -
Disease-related
2005 -0.92 ± 0.10 0.40 0.33 0.48
2006 - 1.00 - -
2007 -2.45 ± 0.18 0.09 0.06 0.12
Stress-related
2005 - 1.00 - -
2006 -0.64 ± 0.14 0.53 0.40 0.70
2007 -0.00 ± 0.15 1.00 0.75 1.34
No visible ailments
2005 - 1.00 - -
2006 -0.21 ± 0.07 0.81 0.70 0.94
2007 -0.72 ± 0.09 0.48 0.40 0.58
Management
2005 -1.73 ± 0.17 0.18 0.13 0.25
2006 - 1.00 - -
2007 -1.00 ± 0.37 0.37 0.58 0.23
12
0.00
0.50
1.00
1.50
2.00
2.50
3.00
3.50
4.00
12345678910111213
Areacode
Risk ratio
Overall
Congenital
Runtism
b
b
c
c
ca
a a
a
b
c a
a c
Figure 3. Probability of mortality and standard errors for each collection area (Areacode) for
overall survival, congenital defects and runtism. Data points labelled with the same letters are
significantly different from Area 1 (captive nests) for their respective cause of mortality.
3.1.1.3 Runtism analysis
There were 2,454 deaths from runtism over the study period with an average age of death of 192 days
(min-max: 5 - 1035 days). All explanatory variables were significant (Table 5). The hazard of
mortality increases significantly by 7.91% for every week later a clutch hatches (1.09% per day).
However, this hazard is offset by 2.29% for every additional hatchling produced from a clutch.
Using 2007 as a base year, crocodiles that hatched in both 2005 and 2006 had significantly lower
hazards of mortality (49% and 77%, respectively; Table 6). The reason for the large discrepancy
between 2007 and the other years was largely due to management. This lead to a disease outbreak and
early antibiotic administration which resulted in the majority of these animals never thriving and
eventually they became runts. Although the initial animals were categorised in the stress-related cause
of mortality, once the course of antibiotics was finished and no further bacteria were being detected
(as reported by pathological studies at BVL), the animals that continued not to thrive were considered
as runts.
Areas 2 (59%), 10 (37%) and 12 (67%) had significantly larger hazards of mortality compared to Area
1, whilst Area 3 was significantly lower (36%; Figure 3). All other areas were non-significantly
different to Area 1.
3.1.1.4 Disease-related analysis
584 disease-related deaths were recorded with an average age of death of 129.7 days (min-max: 2 –
596 days). Interestingly, the hazard of mortality for an animal dying due to a disease-related cause
decreased by 7.84% for each week later the clutch hatched (1.16% per day), and decreased further
(2.03%) for each additional hatchling that was produced from the clutch (Table 5).
13
For this cause of mortality, 2006 was used as the base year for comparison and both 2005 (60%) and
2007 (91%) had lower hazards of mortality due to disease-related causes (Table 6).
All areas were non-significantly different to Area 1 (p>0.05) with the exception of Area 3 which had a
25% lower hazard of disease-related mortality. Area 6 had no observed disease-related mortalities
(Figure 4).
3.1.1.5 Stress-related analysis
355 stress-related mortalities were recorded during the study period (average age 292.77 days; min-
max: 2 – 1,023 days). The time of hatch was non-significant, whilst for each additional hatchling
produced in a clutch, the hazard of mortality decreased by 1.45% (Table 5).
0.00
0.25
0.50
0.75
1.00
1.25
1.50
1.75
2.00
2.25
2.50
12345678910111213
Areacode
Risk ratio
Disease-related
Stress-related
NVA
Management
c
d
b
a c
a
b
c d
d
c
d
d
Figure 4. Probability of mortality and standard errors for each collection area (Areacode) for
the disease-related, stress-related, no visible ailments (NVA) and management categories of
mortality. Data points labelled with the same letters are significantly different from Area 1
(captive nests) for their respective cause of mortality.
14
Although 2005 was used as a base year, 2007 was non-significantly different (p>0.05). However,
2006 had a 47% lower hazard of mortality compared to the other two years (Table 6). The greater
hazard in 2007 was related to a management difficulty mentioned above in 3.1.1.3 where the initial
cause of deaths were categorised as stress-related.
Only Area 3 had a significantly different hazard compared to Area 1 (42% lower), whilst Areas 6 and
8 had no mortalities recorded in this category (Figure 4).
3.1.1.6 No visible ailments
There were 1,150 deaths in this category with an average failure time of 235.53 days (min-max: 1 -
1035days). HDays was non-significant (Table 5), whilst an increase in NoHatch decreased the hazard
of mortality by 2.62%.
For this analysis, 2005 was again used as the base year for comparison (Table 6) as it had the greatest
number of observed deaths. In comparison, 2005 and 2007 had significantly lower hazards of 19%
and 52%, respectively.
Areas 3 (30%), 7 (23%) and 9 (35%) had significantly lower hazards compared to Area 1, whilst all
other areas were non-significantly different to Area 1 (Figure 4).
3.1.1.7 Management
Management deaths occur occasionally on the farm due to unforeseen events. There were 336 deaths
recorded in the Management category during the study period (average age 357.92 days; min-max: 1
– 957 days). As expected, neither HDays nor NoHatch were significant (Table 5) as these events
occur randomly. Crocodiles that hatched in 2005 (82%) and 2007 (63%) had significantly lower
hazards of mortality than animals that hatched in 2006 (Table 6). This was mainly a function of the
newly constructed yearling pens where piling-up of animals became a problem due to delayed
refilling of water into these pens. When the husbandry structure was optimised, these mortalities
decreased. Accordingly, Management was the appropriate category for these animal deaths to be
categorised under.
Interestingly, there were significant area effects (Figure 4). Areas 7, 10, 11 and 12 had significantly
lower hazards of mortality compared to Area 1 (59%, 34%, 33% and 52%, respectively).
3.1.2 Kaplan-Meier survival functions
For each mortality category, a Kaplan-Meier estimate of the baseline survival function from hatch to
365 (Figure 5) and 1,035 (Figure 6) days of age was produced. The plot shows the probability of a
crocodile surviving to any given day, and shows that the first year is definitely the period of highest
mortality. The probabilities of crocodiles surviving to day 365 and day 1,077 for each cause of
mortality are given in Table 7. A brief description of each survival function is given below.
Congenital – 75% of congenital defect deaths occur by day 67 with the majority being animals with
unabsorbed yolk sacs. The curve declined steeply to day 67 (99.72%) and then gradually becomes
horizontal. There was only one death after day 365 (day 515) which was an animal with a spinal
defect.
Runtism – Deaths from runtism constitute the majority of deaths that occur on the farm. Few deaths
classified as runtism occur before day 71. However, between days 71 and 250, the probability of
survival decreases rapidly (90.69% at day 250). After day 250, the decline continues albeit at a slower
rate.
15
16
Disease-related - The probability of a hatchling crocodile surviving a disease outbreak to three months
is 99%. However, between three and five months, around the same time the residual yolk sac is fully
absorbed, the probability of a hatchling surviving a disease-related outbreak is reduced by 1.5%
(97.5% survival probability at day 160). From this time, the probability of survival plateaus to the end
of the first year (96.8%) and the remainder of its production life (96.7%).
Stress-related – The trend for this cause of mortality was similar to that described above for disease-
related mortalities, although it occurs later after hatching. The first 100 days is relatively free of
stress-related deaths as no movements, etc occur during this time (probability of survival 99.78%).
However, after day 100, management regimes such as grading and moving begin and the probability
of survival decreases by 1.23% to day 424 (98.55%). The mortalities then plateau until day 616
(98.46%) and then drops again by 0.45% until day 829 (98.01%).
No visible ailments – Deaths in this mortality category are consistent during the first 267 days with
the probability of survival decreasing by 4%. This is proceeded by a continual decline at 94.29% to
day 1077.
Management – The probability of mortality due to management is negligible over the first 289 days
(0.66%). However, by day 428, the probability had increased by 2.11% before plateauing until day
502 and then further increasing to day 822. The first drop in survival was mainly influenced by the
slow re-filling of water into some pens, as previously mentioned in Section 3.1.1.7.
Table 7. Probability of a crocodile surviving to day 365 and day 1077 for each cause of
mortality.
Probability of survival
to day:
Cause of mortality 365 1077
Overall survival 88.15% 83.39%
Congenital defects 99.40% 99.39%
Runtism 89.74% 86.05%
Disease-related 96.77% 96.71%
Stress-related 98.64% 97.85%
NVA 96.13% 94.29%
Management 98.50% 96.24%
0.88
0.89
0.9
0.91
0.92
0.93
0.94
0.95
0.96
0.97
0.98
0.99
1
0 50 100 150 200 250 300 350
Survival time (days)
Estimated survivor function
Overall
Congenital defects
Runtism
Disease-related
Stress -related
NVA
Management
17
Figure 5. Kaplan-Meier estimated survival functions for crocodiles between hatch and one year of age (365 days) for each cause of mortality. The
y-axis has been scaled from 0.88 to 1.00 to provide a clearer view of the trend for each curve.
0.82
0.84
0.86
0.88
0.9
0.92
0.94
0.96
0.98
1
0 100 200 300 500 600 700 800 900 1000 1100
Survival time (days)
Estimated survivor function
400
Overall
Congenital defects
Runtism
Disease-related
Stress-related
NVA
Management
Figure 6. Kaplan-Meier estimated survival functions for crocodiles between hatch and day 1077 (2.95 years) for each cause of mortality. The
vertical line indicates one year of age (365 days). The y-axis has been scaled from 0.82 to 1.00 to provide a clearer view of the trend for each curve.
365 da
y
s
18
3.2 Pedigree resource
Progeny records were collected from 67 “pair” families (53 sire families; 67 dam families). Parents
were all known-breeding pairs that were originally wild caught and assumed to be unrelated. There
was a total of 2,721 hatchlings of which 407 mortalities were recorded during the study period.
Summary statistics are shown in Table 8. Those animals that were still in the production system at the
end of the trial period were right censored (n = 2,314; 85.04%) with an average censoring time of 629
days. A summary of the number of deaths within each category are shown in Table 9. Computational
problems in fitting the survival model occurred when analysing deaths due to the management
category, consequently this cause of death was omitted from further analyses.
Table 8. Summary statistics of Pedigree resource data used in the survival analyses from
Darwin Crocodile Farm. Right censored records were animals that were assumed to still be in
the production system at the termination of the study period (31st December, 2007).
Year Total
hatchlings Total no.
deaths No. right
censored records Av. censor age
(± st. dev.)
2005 909 143 766 1007 (289)
2006 815 104 711 628 (151)
2007 997 160 837 252 (71)
TOTAL 2721 407 2314
Table 9. Total number of deaths in the Pedigree resource and a breakdown into the six defined
mortality categories.
Reason for death
Year Total no.
deaths Congenital
defects Runt
Disease-
related Stress-
related No visible
ailments Managemen
t
2005 143 4 63 12 23 39 2
2006 104 1 30 16 20 29 8
2007 160 2 108 9 19 22 0
TOTAL 407 7 201 37 62 93 10
3.2.1 Pair model survival analysis results
A summary of the results from the Pedigree resource using the pair model are given below, whilst a
summary of the significant explanatory variables are given in Table 10. Hdays was significant for all
causes of death with the exception of congenital defects and no visible ailments (NVA), whilst
NoHatch was non-significant for all causes except for disease- and stress-related deaths. Year was
significant for the overall survival and runtism analyses. The random effect of Pair was significant in
all of the analyses. Clutch was modelled as an interaction term between Pair and Year, and thus was
only evaluated for the overall survival and runtism causes of death where year was also significant.
However, Clutch was only significant for the runtism cause of death. Table 10 also shows the
heritability estimate for each cause of mortality. The heritability estimates for all causes of death were
0.76 (SE 0.09) with the exception of runtism which is 0.71 (SE 0.08).
3.2.1.1 Overall Pedigree survival model
The 407 deaths occurred at an average age of 192.23 days (min – max: 2 – 1,002 days). The antilog
estimate for HDays (Table 10) indicated that for each day later a crocodile hatches the hazard of
mortality increases by 0.59% per day or 4.23% per week. 2007 was used as the base year and in
comparison, 2005 and 2006 had reduced hazards of mortalities of 72% and 62%, respectively.
19
Pair 1 had the lowest log hazard estimate of -1.59 (antilog estimate (e-1.59) = 0.204), whilst Pair 51 had
the highest estimate of 1.75 (antilog estimate = 5.78; Figure 7A). This means that a juvenile from a
clutch produced by Pair 1 has the highest chance of surviving to slaughter whilst juveniles produced
by Pair 51 have the lowest chance, compared to all other breeding pairs. More specifically, if we
denote S0(t) as the baseline survival function, that is the probability that an individual survives to age
t, averaged across the population, then the survival function for offspring of Pair 1 will be [S0(t)]0.204
(increased survival) whereas those from Pair 51 will have a survival function of [S0(t)]5.78 (reduced
survival). So in general, the survival function for offspring of a particular pair will be [S0(t)]R, where
R is the hazard ratio for a particular pair, being the antilog of the BLUP estimate on the log hazard
scale. The baseline survival function, S0(t), is routinely available in survival analysis output (Ducrocq
and Sölkner 1994; 1998), and has been shown in Section 3.2.2 (Figures 12 and 13).
Since the hazard of mortality changes with time, it was decided that the most appropriate time to
approximate breeding values was at day 365 (or one year). Juvenile survival CBVs are expressed as a
percentage difference in survival to 365 days, relative to the population average, and have been
calculated as
{}
1001(365)][S1001
(365)S
(365)]S[
CBV 1R
0
0
R
0
ii
i×=×
= ,
with approximate standard errors
{
}
100SE(BLUP)(365)][S(365)lnSR)SE(CBV i
R
00ii ××=
where Ri and SE(BLUPi) are the hazard ratios (exponentiated hazard BLUP estimates) and standard
error of the BLUP estimates, respectively (Isberg et al. 2004).
Table 10. Significance summary for explanatory variates used in the Pair model analyses for
the different causes of mortality. Regression coefficients (SE) on the log-scale are given for
the significant (p<0.000) HDays and NoHatch terms, whilst 9 indicates if the term was
significant for Year. The respective variance component is given for each significant (p=0.000)
Pair and Clutch term. 8 indicates if the term was non-significant. Heritability (SE), h2, are also
given.
HDays NoHatch Year Pair Clutch h2 (SE)
Overall 5.91×10-3 (2.93×10-3) 8 9 1.00000 8 0.76 (0.09)
Congenital 8 8 8 1.00000 8 0.76 (0.09)
Runt 2.36×10-2 (7.66×10-3) 8 9 2.00126 2.00008 0.71 (0.08)
Disease-related -2.96×10-2 (9.71×10-3) -4.73×10-2 (2.04×10-2) 8 1.00007 8 0.76 (0.09)
Stress-related -2.76×10-2 (8.10×10-3) -4.42×10-2 (1.61×10-2) 8 1.00001 8 0.76 (0.09)
NVA 8 8 8 1.00001 8 0.76 (0.09)
20
The probability at 365 days (S0(365)) used to calculate the crocodile breeding values (CBVs) for
overall survival was 0.85. From the CBVs shown in Figure 7 for overall juvenile survival in the
Pedigree resource, offspring from Pair 1 have a 11% reduced risk of mortality compared to the herd
average, whilst offspring from Pair 51 have a 46% greater risk of dying.
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
2.5
Log Hazard Estimate for 365 day Survival
-60
-50
-40
-30
-20
-10
0
10
20
1
26
47
65
24
37
27
62
66
16
57
2
6
53
7
46
28
19
23
64
49
3
33
25
14
60
20
5
41
36
42
50
63
67
54
44
39
11
29
15
55
21
59
12
9
40
10
22
18
48
34
56
8
31
43
32
58
4
45
30
38
17
61
13
35
52
51
Pair
365 day Survival CBV
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
2
2.5
Log Hazard Estimate for 365 day Survival
-60
-50
-40
-30
-20
-10
0
10
20
1
26
47
65
24
37
27
62
66
16
57
2
6
53
7
46
28
19
23
64
49
3
33
25
14
60
20
5
41
36
42
50
63
67
54
44
39
11
29
15
55
21
59
12
9
40
10
22
18
48
34
56
8
31
43
32
58
4
45
30
38
17
61
13
35
52
51
Pair
365 day Survival CBV
A
B
Figure 7. A) Log hazard pair estimates (±SE) of overall juvenile survival in the Pedigree
resource. B) Pair CBVs (±SE) at 365 days using the Cox’s proportional hazards model.
21
3.2.1.2 Congenital defects
There were very few congenital deaths (n = 7) with an average failure time of 105.29 days (or 3.5
months). All deaths from congenital defects had occurred by day 357. There were no significant
explanatory variables. Despite such few observations of congenital deaths in this dataset, the random
effect for Pair was significant (p = 0.02) giving the heritability estimate of 0.76 (SE 0.09). The
probability of a crocodile surviving to 365 days from a congenital defect is 1.00. However, the
standard errors for the CBVs calculated at day 365 were all greater than the CBV estimate and are,
therefore, not presented.
3.2.1.3 Runtism
The majority of deaths were again due to runtism (n = 201; Table 9) at an average age of 173.99 days
(min – max: 5 – 853 days). For every day later a clutch hatches, the risk of deaths increases by 2.39%.
The risk of an animal dying that hatched in either 2005 or 2006 is 33% or 24% respectively lower
than an animal that hatched in 2007.
The 365 day probability of survival was estimated to be 0.93. Figure 8 shows that offspring from Pair
26 had a 4.89% lower risk of dying compared to the herd average, whilst Pair 51 had a 18.22% greater
risk.
3.2.1.4 Disease-related deaths
As with the Porosus resource analysis of disease-related deaths, the hazard of mortality due to disease-
related illness decreases the later a clutch hatches (by 2.92% per day) and the greater the number of
hatchlings (4.62% per additional hatchling). There were 37 disease-related deaths in the Pedigree
resource with an average age of death of 150.87 days (min- max: 9 – 619 days).
The CBVs for disease-related deaths is shown in Figure 9 using the probability of survival to day 365
of 99%. Offspring from Pair 20 have the lowest mortality compared to the farm average (31%), whilst
Pair 38’s offspring are at the greatest risk of dying (203%).
-40
-35
-30
-25
-20
-15
-10
-5
0
5
10
26
37
1
3
65
6
38
47
39
55
16
24
67
28
7
64
57
45
66
50
34
33
62
12
25
29
53
46
36
48
2
49
42
8
27
11
19
31
41
54
21
60
14
30
10
9
40
59
18
44
20
23
58
56
43
15
63
5
22
32
4
35
61
13
17
52
51
Pair
365 day Runt CB
V
22
Figure 8. Pair CBVs (±SE) at 365 days for the runtism cause of death.
-3.5
-3
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
20
34
15
5
2
1
3
24
64
23
53
17
45
46
18
54
31
29
61
21
26
13
55
33
14
22
43
66
25
65
56
27
7
51
19
32
60
44
58
62
63
59
16
57
47
52
41
48
40
8
49
11
67
37
39
50
12
6
36
9
42
10
28
4
35
30
38
Pair
365 day Disease CB
Figure 9. Pair CBVs (±SE) at 365 days for disease-related deaths.
3.2.1.5 Stress-related deaths
The hazard of mortality decreases by 2.72% for every additional day later a clutch hatches and
reduces a further 4.32% for each additional hatchling produced from a clutch. A total of 62 stress-
related deaths occurred during the study period between 4 and 929 days old (average 221.34 days).
Pair 30 offspring had a 0.75% greater chance of surviving a stress event compared to the herd average
(Figure 10). Pair 45 offspring had a 270% greater chance of dying from such an incident. These CBVs
were calculated using 99% as the 365 day survival rate from the survival function.
3.2.1.6 No visible ailments
NVA deaths were the second most common deaths in the Pedigree resource (n = 93) with an average
age of 203.24 days (min – max: 2 – 1,002 days). As expected, there were no significant explanatory
variables for NVA. However, Pair was significant and the CBVs are shown in Figure 11. Pair 2
offspring had an 83% greater chance of survival, whilst Pair 55 had a 410% chance of dying for no
visible reason compared to the herd average. These were calculated using the 365 day survival rate of
0.97.
23
-5
-4.5
-4
-3.5
-3
-2.5
-2
-1.5
-1
-0.5
0
0.5
1
1.5
30
1
23
24
4
31
64
53
33
46
54
61
50
37
27
26
66
25
42
55
32
19
15
43
16
51
65
56
63
9
10
60
40
58
44
62
59
47
3
11
52
57
41
49
67
6
28
29
5
17
2
14
36
22
13
7
38
12
35
48
8
39
20
34
18
21
45
Pair
365 day Stress CB
V
Figure 10. Pair CBVs (±SE) at 365 days for stress-related deaths.
-7
-6
-5
-4
-3
-2
-1
0
1
2
39
20
1
66
26
47
6
13
27
28
23
60
63
49
24
62
65
17
58
19
37
9
50
59
41
56
18
53
67
46
57
31
44
10
48
22
7
5
14
34
52
45
16
43
32
33
25
4
36
11
3
21
15
42
30
51
64
35
12
38
54
40
29
61
8
55
Pair
365 day NVA CB
V
Figure 11. Pair CBVs (±SE) at 365 days for deaths occurring from no visible ailments (NVA).
3.2.2 Correlation between crocodile breeding values (CBVs)
To understand the interaction between the breeding values for each cause of mortality for the different
pairs, a correlation matrix was created (Table 11). There were significant correlations (p<0.05)
between overall survival CBVs and both runtism CBVs (0.80) and NVA CBVs (0.34). However, this
was most likely due to the number of mortalities recorded in each of these categories that constituted
72% of overall deaths in the Pedigree resource. It is interesting to note that there is a significant, albeit
weak, correlation between the congenital defects and NVA CBVs (0.25)
24
25
Table 11. Correlation coefficients between crocodiles breeding values (CBVs) for the different
causes of mortality. An asterisk (*) indicates the correlation was significantly (p<0.05) different
to zero.
CBV
Overall
survival Congenital
defects Runtism
Disease-
related Stress-
related
Congenital defects 0.12
Runtism 0.80* 0.04
Disease-related 0.24 0.21 0.01
Stress-related 0.22 -0.06 -0.02 0.04
CBV
NVA 0.34* 0.25* 0.09 0.16 0.06
3.2.3 Pair model Kaplan-Meier survival functions
For each cause of death, a Kaplan-Meier estimate of the baseline survival function from hatch to 365
(Figure 12) and 1002 (Figure 13) days of age was produced. Table 12 shows the probability of a
crocodile surviving to day 365 and day 1,002, respectively.
Table 12. Probability of a crocodile surviving to day 365 and day 1002 for each cause of
mortality.
Probability of survival
to day:
Cause of mortality 365 1,002
Overall survival 84.77% 79.64%
Congenital defects 99.70% 99.70%
Runtism 93.11% 91.03%
Disease-related 99.29% 99.25%
Stress-related 98.69% 98.07%
NVA 97.30% 96.16%
0.84
0.86
0.88
0.9
0.92
0.94
0.96
0.98
1
0 50 100 150 200 250 300 350
Survival time (days)
Estimated survivor function
Overall
Congenital defects
Runtism
Disease-related
Stress-related
NVA
26
Figure 12. Kaplan-Meier estimated survival functions for crocodiles in the Pedigree resource between hatch and one year of age (365 days) for
each cause of mortality using the Pair model. The y-axis has been scaled from 0.84 to 1.00 to provide a clearer view of the trend for each curve.
0.78
0.8
0.82
0.84
0.86
0.88
0.9
0.92
0.94
0.96
0.98
1
0 100 200 300 400 500 600 700 800 900 1000
Survival time (days)
Estimated survivor function
Overall
Congenital defects
Runtis m
Disease-related
Stress-related
NVA
27
Figure 13. Kaplan-Meier estimated survival functions for crocodiles in the Pedigree resource between hatch and day 1002 (2.75 years) for each
cause of mortality using the Pair model. The y-axis has been scaled from 0.78 to 1.00 to provide a clearer view of the trend for each curve.
3.2.4 Animal model survival analysis results
Using the pair model (Section 3.2.1) above, and with the exception of runtism (h2 0.71 SE 0.08), the
heritability estimates for all of the other causes of mortality were the same (0.76 SE 0.09). As a way
of further investigating the heritability for the different causes, an animal model was used. In the pair
model, the offspring were analysed as full-siblings to produce the pair (and clutch) variance
component. However, using the animal model, the relationship between each individual is considered
(mother, father, full-sibling, half-sibling, etc) when estimating the variance components.
A summary of the results from the Pedigree resource using the animal model are given in Table 13.
For each day later a clutch hatches, the risk of mortality is increased by 0.97% for both the disease-
related and stress-related causes of mortality. For the stress-related cause of mortality, this was further
confounded by each additional hatchling produced from clutch (0.96%). Year was significant for both
the overall survival (p = 0.03) and runtism (p = 0.00) analyses.
Animal was significant for all models. In comparison to the pair model, the animal model heritability
estimates varied for all causes of mortality. The lowest heritability estimate was 28% (SE 0.02) for
deaths with no visible ailments, whilst the highest estimate was for runtism (60% SE 0.04).
Interestingly, the variance components estimated for the overall survival and congenital defects causes
were the same (1.040) which resulted in the same heritability estimate (39% SE 0.03).
Table 13. Significance summary for explanatory variates used in the Animal model analyses
for the different causes of mortality using the Pedigree resource. Regression coefficients (SE)
on the log-scale are given for the significant (p=0.000) HDays and NoHatch terms, whilst 9
indicates if the term was significant for Year. 8 indicates if the term was non-significant. The
respective variance component is given for each Animal term which was used to calculate the
heritability (SE), h2, estimates.
HDays NoHatch Year Animal h2 (SE)
Overall 8 8 9 1.03984 0.39 (0.03)
Congenital 8 8 8 1.03984 0.39 (0.03)
Runt 8 8 9 2.43362 0.60 (0.04)
Disease-related -2.95×10-2 (9.86×10-3) 8 8 2.12290 0.56 (0.04)
Stress-related -2.88×10-2 (7.25×10-3) -4.12×10-2 (1.53×10-2) 8 1.00000 0.38 (0.03)
NVA 8 8 8 0.65468 0.28 (0.02)
28
4. Histopathology Results
4.1 General findings
In Phase 1, the mean total body length in runt crocodiles was 35.41 cm compared to 56.02 cm in
normal crocodiles (Table 2), whilst in Phase 2, runt crocodiles were an average 34.35 cm total body
length compared to 48.35 cm in normal crocodiles. There were similar differences in body weight
between the two phases and runt versus normal groups (Table 2). Overall, the Phase 2 crocodiles were
a few months younger than Phase 1 animals although there were no significant differences within the
two phases (p>0.05). Fifteen runts and 16 normal crocodiles were male. With the exception of two
normal crocodiles where sex could not be ascertained, five runt and two normal crocodiles were
female.
Gross necropsy did not reveal abnormalities in runt compared to normal crocodiles apart from wasted
body condition, enlarged gall bladders and pancreatic atrophy. The latter two findings were most
notable in Phase 1, and likely reflect decreased food intake in runts. In both groups, scattered
pinpoint shallow skin erosions (< 1 mm) were common, primarily evident over the ventral thorax and
abdominal regions. No attempt was made to quantify the number of lesions between runt and normal
crocodiles, since the overall impression was that both groups were approximately equally affected,
and the lesions were minor.
In order to investigate the possibility of metabolic bone disease in runt crocodiles, bone strength was
subjectively assessed in both runt and normal crocodiles by ability to manually bend or break the long
bones or jaw. In both runt and normal crocodiles, bone strength was assessed as similar and within
normal range (bones did not bend appreciably before breaking). There was no gross evidence of bone
malformation, such as bent or twisted limbs, irregular jaw contour, loose teeth or curvature of the
vertebral column in runt crocodiles.
The yolk sac was specifically examined grossly in crocodiles in Phase 1 to investigate the possibility
that infection or abnormal resorption of the yolk sac was related to runting. Yolk sac remnants were
either not appreciable or visible as 1-2 mm diameter white foci on the serosa of the jejunum in 18
crocodiles in Phase 1, with no consistent difference in appearance between runt and normal
crocodiles. Two crocodiles, one runt and one normal, had 1 cm diameter yolk sac remnants containing
caseous material.
4.2 Haematology
The haematological examination of whole blood revealed that runt crocodiles were moderately to
markedly anaemic compared to normal crocodiles (lower packed red blood cell volume in runts with
mean 13% compared to 20% in normal crocodiles). Numbers of white blood cells were comparable
between normal and runt crocodiles, although some interesting changes were noted in the few normal
crocodiles from which bacteria were isolated (either increased numbers of various white blood cells or
evidence of toxicity in heterophils, which are standard responses to infection).
4.3 Biochemistry
The only difference in electrolytes (sodium, potassium and chloride) in runt compared to normal
crocodiles was elevated sodium in runt crocodiles (runt mean 140.2 mmol/L; normal mean 144.2
mmol/L). Serum albumin was markedly lower in runt crocodiles (mean 10.9 g/L) compared to normal
crocodiles (mean 17.6 g/L). Globulins were slightly lower in runt (mean 21.4 g/L) compared to
normal crocodiles (mean 24.0 g/L) as was serum glucose (runt mean 5.8 mmol/L; normal mean 7.9
mmol/L). Both the lower albumin and glucose in runt crocodiles was likely due to decreased food
intake and body wasting. Serum iron was not different between normal and runt crocodiles. Among
29
the parameters that may suggest problems with renal function (uric acid or urea), there were no
differences between the two groups. Runt crocodiles had higher concentrations of creatinine
compared to normal crocodiles. However, this parameter is poorly understood in reptiles and is
generally considered of questionable significance in assessment of renal function in reptilians (Divers
2000a). The lack of elevation in uric acid runt crocodiles, and lack of abnormalities in any other
parameters that would be suggestive of impaired renal function, suggest that creatinine may not be a
valuable parameter in the assessment of renal function in crocodiles.
Among the enzymes, those associated with the liver showed few differences. ALT and AST were
generally similar between groups, ALP was slightly elevated in a few runt crocodiles and GGT was
slightly higher in runt crocodiles (mean 5.4 U/L compared to 4.4 U/L in normal crocodiles), perhaps
reflecting decreased bile flow resulting from inappetence in runt crocodiles (Divers 2000b). Creatine
kinase (CK), an enzyme that is released from muscle when it undergoes marked exertion or acute
damage, was higher on average in normal crocodiles (2557 U/L) compared to runt crocodiles (296
U/L). This is likely due to the greater strength of normal crocodiles and ability to struggle during
restraint for blood sampling. AST was also higher in normal (mean 61.9 U/L) compared to runt
(mean 56.0 U/L) crocodiles. This enzyme is released from both the liver and muscle in reptiles
(Divers 2000b), and likely also reflects mild acute muscle damage in normal crocodiles as a result of
manual restraint during blood sampling.
Calcium and phosphorus were slightly lower on average in runt crocodiles in both phases of the study
(runt crocodile calcium 2.4 mmol/L, phosphorus 1.2 mmol/L; normal crocodile calcium 2.7 mmol/L,
phosphorus 1.6 mol/L).
4.4 Corticosterone assay
The corticosterone assay was only conducted during Phase 2 of the study. Of the thirty animals tested,
seven (two normals and five runts) exceeded the detection limits of the standard curve generated with
the kit (20 ng/ml). For analysis purposes, the corticosterone levels of these animals were set to 20.
The mean serum corticosterone for normal crocodiles was 10.13 ng/ml compared to 16.18 ng/ml for
runt crocodiles.
4.5 Bacteriology
Bacterial culture of the liver and spleen, performed in Phase 1 of the study, did not result in isolation
of bacteria in the majority of crocodiles. In one runt crocodile, Morganella morganii was isolated
from the liver, and Corynebacterium sp. was isolated from the spleen and liver in a second runt
crocodile. Streptococcus agalactiae was isolated from the liver in three normal crocodiles, and also
from the spleen of one of these and a swollen limb of another. S. agalactiae previously caused low
numbers of cases of necrotising fasciitis on the farm from which the study animals originated (Bishop
et al. 2007). The yolk sac content of one runt crocodile and one normal crocodile with enlarged yolk
sac were cultured yielding isolates of Salmonella sp. and Edwardsiella tarda, respectively.
4.6 Parasitology
Faecal flotations of all crocodiles in Phase 1 did not reveal any helminth eggs or coccidial oocysts.
30
4.7 Histology
4.7.1 Lymphoid tissue
The lymphoid tissue of the thymus, tonsil, spleen and bone marrow were assessed in all crocodiles
since there appeared to be substantial differences in lymphoid populations between runt and normal
crocodiles.
In normal crocodiles, the thymus was composed of multiple large lobes with densely populated
cortical tissue distinct from the less densely populated central medulla. In the majority of runt
crocodiles, there was a marked reduction in the amount of thymic tissue. Rather than being an
aggregate of large, distinct lobes, the thymus appeared as multiple small widely separated and poorly
delineated sparse lymphoid aggregates, with minimal to no distinction between cortex and medulla. In
several runt crocodiles, thymus tissue was so limited that several sections of the neck tissues in the
known region of the thymus were necessary in order to find a few small lobes. A few runt crocodiles
had thymic lobes intermediate in appearance between normal crocodiles and the majority of the runts.
Forming the tonsil, all normal crocodiles had abundant lymphoid tissue present in large nodules
subepithelially in prominent mucosal folds in the caudal dorsal oropharynx. Runt crocodiles had
structurally similar tissue in this region, but subjectively, mucosal folds appeared generally less
pronounced and contained smaller and fewer lymphoid nodules.
The splenic lymphoid population was not as obviously depleted as the thymus and tonsillar tissues in
runts compared to normal crocodiles. Runt crocodiles had moderately-sized discrete lymphoid
nodules surrounding arterioles, as did normal crocodiles. Further quantitative assessment of splenic
lymphoid tissue would be necessary to discern any definitely significant differences in splenic
lymphoid tissue. Golden-brown globular pigments were apparent within the splenic macrophages of
most crocodiles. This pigment stained strongly with Perl’s stain for iron, indicating iron storage. This
pigment was assessed as being present in low to moderate amounts in normal crocodiles and in
moderate to abundant amounts in runt crocodiles.
In other tissues, abundance of the lymphocyte population was variable. In Phase 1, all ten of the
normal crocodiles had a few scattered interstitial lymphoid aggregates in the lung, which were not
noted in any of the runts. However, histologically there were no obvious major differences in quantity
of lymphoid tissue associated with the gastrointestinal tract (submucosa of the oesophagus and lamina
propria of the intestine).
4.7.2 Adrenal gland
The adrenocortical cells, which produce the stress hormone corticosterone, were generally more
vacuolated in runt compared to normal crocodiles in both phases.
4.7.3 Bone
Growth plates and bone marrow samples were examined from all crocodiles. The growth plates of
normal crocodiles were characterised by distinct zones of resting, proliferating and hypertrophying
chondrocytes forming a primary spongiosa of wide spicules and trabeculae of progressively
mineralised cartilage in the proximal metaphysis. Formation and remodelling of cartilage spicules in
the primary spongiosa and bone trabeculae in the metaphysis appeared very active, with surfaces
covered in rows of either contiguous osteoblasts or groups of osteoclasts. In the growth plate of runt
crocodiles, the components noted above were apparent, but zones of cartilage were less distinct, and
proliferation and hypertrophying zones appeared relatively thinner compared to the resting cartilage
closer to the joint than in normal crocodiles. Mineralising cartilage spicules were very scanty, as were
osteoblasts and osteoclasts, indicating decreased bone formation and remodelling. The general shape
and contours of the bone appeared comparable between normal and runt crocodiles, as did the
31
approximate thickness of the diaphyseal cortices. There were no abnormalities of the femorotibial
joint noted and the joint appeared similar between runt and normal crocodiles.
Histological estimates of bone marrow cellularity were generally higher for normal compared to runt
crocodiles. Normal crocodiles had bone marrow that was 40-60% cellular, with most having an
estimated 50% cellularity. Runt crocodile cellularity ranged from 15-50%, with most estimates being
20-30%. There were no consistent differences in ratio of myeloid to erythroid cell lines between
normal and runt crocodiles. The myeloid to erythroid ratio was frequently estimated at 1:1, but varied
from 4:1 to 1:2. The bone marrow of normal crocodiles generally contained moderate numbers of
lymphocytes, while lymphocytes were rare in the bone marrow of runt crocodiles.
4.7.4 Liver, gastrointestinal tract and pancreas
The liver in runt and normal crocodiles from Phases 1 and 2 was generally similar, with both having a
mild to moderate degree of fine vacuolation to the cytoplasm, suggestive of some degree of
physiological glycogen storage. Two runt crocodiles had larger cytoplasmic vacuoles, more typical of
lipid and possibly indicating a degenerative fatty change or abnormality of lipid metabolism. In four
of the runt crocodiles, there was a mild to moderate degree of portal fibrosis and cholangiolar
proliferation. In the two runts with moderate portal fibrosis, there was heterophilic and granulomatous
inflammation centred on bile ductules. In one of these crocodiles, Morganella morganii was isolated
from the liver, indicating that the lesion at least in one crocodile was likely due to low grade chronic
cholangiohepatitis (infection of the bile ducts).
In Phase 1, an attempt was made to semi-quantitatively compare the functional regions of the
gastrointestinal tract by comparing the relative thicknesses of the various areas of the gastric mucosa
and height of intestinal villi. In the stomach, the relative abundance of gastric glands versus mucous
cells was similar between the two groups, as was the thickness of the mucosa. In the duodenum and
jejunum, the length of villi compared to the width at the base was roughly similar between normal and
runt crocodiles. This provides one indication that the poor growth of runt crocodiles was not due to
decreased digestive capacity or absorptive area of the intestine.
In Phase 1, three runt crocodiles had occasional sporulated coccidial oocysts within enterocytes. To
investigate this further, in Phase 2, standard sections of the duodenum, jejunum and colon were
examined carefully for coccidia. This revealed only one sporulated coccidia oocyst in the jejunum of a
normal crocodile. The low numbers of coccidia in a minority of runt crocodiles in Phase 1, and the
absence of coccidia in runt crocodiles in Phase 2 indicates that coccidiosis is not associated with
runting in the present study. Coccidia were not noted in faecal flotations, either because too few were
present, or the spores were released free from the oocysts in the faeces, making them too small to
detect by the conventional floatation test used.
The pancreatic acinar cells in runt crocodiles frequently appeared to contain less zymogen (atrophied)
compared to those of the normal crocodiles. Pancreatic atrophy is generally associated with decreased
food intake in animals.
4.7.5 Yolk sac
Yolk sac histology varied slightly among all crocodiles, with no notable consistent differences
between the two groups. The yolk sac remnants were composed of variable amounts of coagulated
eosinophilic, partially mineralised material (yolk remnant) containing cholesterol clefts, surrounded
by multinucleated giants cells, foamy macrophages, aggregates of lymphocytes and fibrous tissue,
joined to the adjacent jejunum by fibrous tissue covered in serosa. There was no difference in size of
the yolk sac remnant between the normal and runt crocodiles, and none of the yolk sacs, including the
two relatively large yolk sac remnants from which bacteria were cultured, appeared excessively
inflamed.
32
4.7.6 Other tissues
The normal crocodiles had marked vacuolation of adipocytes within the fat body, indicative of
abundant fat storage. The adipocytes of the fat body of the runt crocodiles were smaller (contained
fewer lipids) than those of normal crocodiles, confirming the gross impression of decreased fat
storage in the runts. Skeletal myofibres generally appeared smaller in runt compared to normal
crocodiles, out of proportion to their smaller body size. This likely reflects muscle atrophy, a common
finding in animals in a poor nutritional state.
There were no notable histological differences in the skin of runt compared to normal crocodiles.
Both had low numbers of poxvirus lesions. Histologically, these lesions appeared as discrete foci of
marked epidermal hypertrophy with eosinophilic cytoplasmic inclusions commencing in the parabasal
layers and progressing superficially to involve the entire cytoplasm, with compression of the nucleus
into a thin peripheral rim. At the surface, the keratocytes sloughed containing large purple inclusions.
These lesions are the typical “molluscum contagiosum” type of poxvirus lesions. In one “normal”
crocodile, a typical pox lesion was accompanied by moderate mixed inflammation in the adjacent
dermis, with abundant bacterial cocci within macrophages, typical of a mild lesion caused by
Streptococcus agalactiae in other crocodiles submitted from this farm (Bishop et al. 2007). S.
agalactiae was isolated from the liver and the swollen left arm of this crocodile.
Pituitary and thyroid glands appeared histologically similar between runt and normal crocodiles. In
the thyroid gland, follicles exhibited moderate variation in follicle diameter. All follicles contained
colloid and there were signs of activity in the follicular epithelium (cuboidal to columnar with small
vacuoles in colloid adjacent to apical epithelium) in both groups. There were no notable differences in
brain, spinal cord, eye or kidney between runt and normal crocodiles.
33
5. Discussion
5.1 Survival analysis
The main reason for conducting this series of survival analyses was to identify the major risk factors
associated with crocodile deaths on Australian crocodile farms using Darwin Crocodile Farm as a
model. Furthermore, the identification of major risk factors should then be used to set priorities for
further research to reduce crocodile deaths. From the analyses, it can be shown that runtism (48.66%)
constitutes the highest risk factor of deaths at Darwin Crocodile Farm proceeded by deaths for no
discernible reason (NVA; 22.84%). Disease-related deaths (11.58%) are expectantly downwardly
biased due to early detection and treatment with efficacious antibiotics, whilst stress-related
mortalities (7.04%) are highly dependent on staff experience, handling efficiency and maintaining
optimal husbandry regimes.
5.1.1 Collection area and Pair effects
For all causes of mortality, there were strong geographical effects (egg collection areas; Figures 3 and
4) and, for the Pedigree resource data, highly significant genetic (pair; Figures 7-11) effects.
Identifying areas or pairs where survival rates are lower than other areas will allow appropriate
management decision strategies to be implemented. For example, with the exception of congenital
defects, which were non-significantly different to captive nests (Area 1), Area 3 had a consistently
lower hazard of mortality across all causes of death (lower hazard of mortality range between 25%
and 42%). Similarly, Area 7 had a 23% lower hazard for NVA deaths. In contrast, Areas 2, 10 and 12
had higher risks of dying from runting compared to Area 1. In fact, 7.15%, 9.33% and 9.07% of total
hatchlings from each of these respective areas die from runting alone.
As shown in Figures 7-11, Pairs can be ranked according to their CBV for each cause of death. The
overall survival CBV (Figure 7) ranking of a pair is the combined effect of all the specific causes of
mortality. However, since runting is of greater economic impact, compared to the other categories
described, it should be given a higher weighting for selective breeding. In traditional breeding
programs, breeding objectives are weighted using an economic estimate of the impact of a one unit
deviation. In CrocPLAN (Isberg et al. 2004), the economic weighting for overall juvenile survival was
estimated to be AU$52.37 for each additional hatchling surviving to slaughter age/size. Since
economic weightings are neither available for each separate cause of death nor feasible since the
economic cost of an animal dying is the same regardless of the cause, it was decided to use the 2004
CrocPLAN estimate of AU$52.37 further offset by the percentage of deaths that occur within each
category. Table 13 shows the percentages that were used to offset the economic weighing for each
category. Figure 14 shows the economic impact of deaths from runtism and NVA, expressed as
deviations from the farm mean.
Table 14. Crocodile breeding values (CBVs) were offset by a weighted economic value using
the percentage of death in each category (%) multiplied by AU$52.37. *The missing 1.72% are
from management deaths that are not genetically determined and were not included in the
analyses.
Percentage of
total deaths (%) Economic weighting used
(% × AU$52.37)
Congenital defects 1.72% $0.90
Runting 49.39% $25.86
Disease-related 9.09% $4.76
Stress-related 15.23% $7.98
NVA 22.85% $11.97
TOTAL* 98.28% $51.47
34
-$1,400
-$1,200
-$1,000
-$800
-$600
-$400
-$200
$0
$200
$400
1
26
47
65
24
37
27
62
66
16
57
2
6
53
7
46
28
19
23
64
49
3
33
25
14
60
20
5
41
36
42
50
63
67
54
44
39
11
29
15
55
21
59
12
9
40
10
22
18
48
34
56
8
31
43
32
58
4
45
30
38
17
61
13
35
52
51
Pair
Deviation ($)
Runt $ CBV
NVA $ CBV
Figure 14. Dollar deviation of crocodile breeding values as expressed as a dollar ($) deviation
from the herd average for the runting and NVA causes of death.
Based on these findings, management decisions can be made to select against the inferior captive
breeding pairs, or policies can be re-evaluated in regards to poor performing collection areas for the
high risk mortality categories.
One assumption that was made was that all categories were independent. Correlations between CBVs
for the different mortality categories were determined to investigate if this assumption was correct
(Table 11). Runtism (r = 0.80) and NVA (r = 0.34) had significant correlations with overall survival.
However, this was expected as these two categories had the highest number of deaths. Interestingly,
NVA was significantly correlated with congenital defects (0.25) and the biological significance of this
correlation is worthy of further investigation. All other correlations were non-significantly different
from overall survival (range = -0.06 to 0.21) indicating their effective independence. This
independence does potentially present issues when selecting for mortality as few animals would be
favourable for several causes (Southey et al. 2004). However, as shown above, using the percentage
of deaths as a weighting factor and an indicator of economic significance can be a way of overcoming
this potential problem. In addition, identification of animals with a higher risk of mortality from one
cause, knowing that it is independent from other causes, will assist in molecular genetic studies to
detect genes associated with these different causes, such as quantitative trait loci (QTL; RIRDC
Project No US139A) and candidate gene studies.
5.1.2 Heritability estimates
The estimates of heritability produced in this study (pair model = 0.71-0.76; animal model = 0.28-
0.56) are higher than that previously estimated for crocodile survival (0.15; Isberg et al. 2006) and for
any other domestic livestock species reported so far (sheep 0.08-0.33 (Riggio et al. 2008), 0.11
(Welsh et al. 2006); dairy 0.05-0.07 (Chirinos et al. 2007); rabbits 0.0-0.12 (Eady et al. 2007)).
Interestingly, Ricklefs and Cadena (2008) conducted a study on non-domesticated (zoo) mammals and
birds and reported similar large heritability estimates (0.18-1.68). Crocodiles should also be classified
as non-domesticated as adult breeding crocodiles on crocodile farms are generally from the wild.
Furthermore, unlike at Janamba Croc Farm where some replacement of non-performing animals had
occurred prior to the study conducted by Isberg et al. (2004), no replacement or selection had
occurred at Darwin Crocodile Farm prior to this study. Thus, these heritability estimates probably
reflect that of the wild C. porosus population in the Northern Territory.
35
Of interest was the similarity between the pair model heritability estimates. With the exception of
runting, all other causes of death and overall survival had a heritability estimate of 0.76 (SE 0.09).
Despite having three years of data, clutch was confounded (non-significant) for all categories except
runtism indicating that there were not enough observations in all these categories. Clutch is known to
be a highly significant effect in all aspects of crocodile production (Isberg et al. 2004) so the
continued collection of data in this manner will allow its presumably important effect to be more
accurately estimated. In contrast, the animal model showed differences in the heritability estimates
(range = 0.28 (SE 0.02; NVA) to 0.60 (SE 0.04; runting)).
5.1.3 Hatch days (HDays)
Anecdotal evidence and industry perception has suggested that clutches which hatch later in the year,
when ambient weather conditions are cooler and less humid, have a lower chance of survival despite
the provision of heated water and enclosed pens to increase humidity for “optimal” growth. This
variable was modelled using HDays and was shown to be a significant effect in the majority of
analysis regardless of which dataset (Porosus or Pedigree) or model (Pair or Animal) was being
evaluated.
Where HDays was significant for overall survival, congenital defects and runting, for every day later a
clutch hatched, the hazard of mortality increased (ranging between 0.04% and 2.39%). However, for
disease- and stress-related mortalities, with the exception of the Animal model estimates, HDay
estimates reported that for every day later a clutch hatches, the hazard of mortality decreases between
1.16% and 2.92%. The contrasting results for disease- and stress-related deaths is most probably due
to the early movement of earlier hatchlings from “starter” hatchlings pens to allow room for later
hatching crocodiles (disease-related) and the overcrowding (due to size/growth) later in the year
before animals are moved from hatchling to yearling pens (stress-related).
5.1.4 Number of hatchlings (NoHatch)
It has been suspected that the lower the hatchability of a clutch, the lower the chance of survival of the
resultant hatchlings. There are multiple reasons why embryos can be compromised prior to collection,
particularly in relation to wild nests, including environmental (flooding, overheating, etc), genetic and
population dynamics (female age, nutritional status, etc). The Porosus resource analyses showed the
highly significant effect of NoHatch in all mortality categories, whilst NoHatch was significant in
only the disease- and stress-related mortalities in the Pedigree resource. On farm, nests are able to be
collected within 24 hours of laying, thus reducing the risk of mortality in the embryological phase of
development compared to those nests from the wild which are collected at all stages of development
and environmental exposure.
5.2 Histopathology
The main purpose of the pathology portion of this broad investigation into runting in crocodiles was
to investigate the possibility of infectious disease as a primary cause of the markedly poor growth. A
broad range of testing conventionally used for detection of disease in veterinary pathology, including
haematology, serum biochemistry, faecal parasitology, gross post-mortem, bacteriology and
histopathology failed to reveal overt involvement of a primary infectious disease.
A secondary purpose of the pathology portion of the study was to look for evidence of any
abnormalities that might suggest other underlying conditions present in runt crocodiles. A major
finding in this regard was the marked lymphoid atrophy, most obvious in the thymus and tonsils, in
runt compared to normal crocodiles. A well known cause of lymphoid atrophy in many species,
including crocodilians, is chronic stress resulting in chronically elevated corticosteroid hormones that
suppress many aspects of the immune system (Capen 2007, Lance et. al. 2000, Schmidt et. al. 2003,
Schobitz et. al. 1994, Valli 2007). The presence of chronic stress in runt crocodiles was suggested by
the vacuolated appearance of the adrenocortical cells in the adrenal gland, which in other species is a
36
sign of increased activity of the gland (Capen 2007). To investigate this further (Phase 2), serum
corticosterone was quantified and found to be higher in runt compared to normal crocodiles. Chronic
stress can be caused by a wide variety of factors, including environmental factors and chronic
intercurrent disease, although there was no overt evidence of the latter in the present study (Lance et.
al. 2000, Webster et. al. 1998). It is also possible that maternal stress could influence the functioning
of the foetal or hatchling hypothalamic-pituitary-adrenal axis, resulting in a chronic stress response in
the offspring (Edwards and Burnham 2001, Steyermark and Spotila 2000).
Another less likely explanation for the lymphoid atrophy is that runt crocodiles have a primary
congenital immunodeficiency problem. However, in other species, such conditions are usually rare
hereditary diseases with various manifestations peculiar to specific breeds and affected animals. It
seems unlikely that animals from multiple clutches from a variety of parents would be so frequently
affected by a similar primary immunodeficiency. Finally, the possibility of lymphoid atrophy due to
the effect of a virus that specifically targets lymphoid tissue, has not been ruled-out. There are no
viruses documented to have this effect in crocodiles, although there are precedents in other species.
For example, survival of in utero or neonatal infection with pestiviruses, parvoviruses or retroviruses
in mammals and a wide variety of viruses in birds may leave a young animal with an abnormally
atrophic thymus (Schmidt et al. 2003, Valli 2007). This possibility seems remote, since viruses do not
usually target only one tissue, and there was no other evidence of acute lymphoid necrosis or viral
pathology in any other tissues of the runt crocodiles. However, insidious involvement of a virus
would be worthy of investigation in the future.
An interesting finding in the present study was that despite the marked lymphoid atrophy apparent
when tissues were examined histologically, this was not reflected in circulating white blood cell
counts. In mammals and birds, a typical chronic stress response is elevation of circulating neutrophils
and depression of circulating lymphocytes (Campbell 2004, Duncan et al. 1994). However, in this
study, runt and normal crocodiles had comparable levels of circulating heterophils (the reptilian
counterpart to neutrophils) and lymphocytes. This lack of response of circulating white blood cells to
stress in crocodiles has been noted by other researchers (Turton et al. 1997), indicating that
measurement of circulating white blood cells may be a poor indicator of lymphoid atrophy and
immunosuppression in crocodiles. Slightly decreased serum globulins in runt crocodiles in this study
may be one manifestation of lymphoid atrophy specifically of B-lymphocytes that are responsible for
immunoglobulin (antibody) production for humoral immunity.
Both serum calcium and phosphorus were mildly lower in runt compared to normal crocodiles.
However, since a large proportion of serum calcium in bound to albumin in circulation, a lowering of
serum albumin will result in a clinically irrelevant lowering of total serum calcium (Mader 2000). The
mildly lower total serum calcium in runt crocodiles compared to normal crocodiles may therefore be
at least partially a reflection of their markedly lower serum albumin, rather than evidence of a
problem with calcium metabolism. Serum calcium is rarely affected by dietary calcium level directly.
However, lower dietary intake of phosphorus, as a result of generally lower food intake, in runt
crocodiles, could directly result in a lowering of serum phosphorus in runts compared to normal
crocodiles (Duncan et al. 1994). Finally, since both serum calcium and serum phosphorus may be
elevated in young animals due to bone formation and remodelling (Duncan et al. 1994), the lower
levels of these minerals in runt crocodiles may in part be due to their relatively poor rate of growth.
The physiological significance of the mildly lower levels of these minerals in the serum of runt
crocodiles is questionable. There was no gross evidence of a disease of bone formation. Also, the
histological appearance of the growth plate was more consistent with inactivity due to poor growth
resulting from starvation and/or chronic stress rather than one of the specific metabolic bone diseases
such as rickets due to substantial imbalances in calcium, phosphorus or vitamin D metabolism
(Hochberg 2002, Thompson 2007). Chronic elevations of corticosteroids inhibit both growth hormone
secretion and action in mammals, and experimentally results in markedly poor growth in alligators
(Elsey et al. 1990, Hochberg 2002, Lance et al. 2000).
37
Many of the other findings in runt crocodiles in the present study, including absence of fat, low blood
albumin, slightly elevated GGT and histological findings of pancreatic atrophy and inactive growth
plates are likely signs of wasting or inanition. The moderate anaemia in runt crocodiles also seems
likely to be a non-specific finding related to inanition. Iron deficiency is a common cause of anaemia
in young animals. However, serum iron was comparable between runt and normal crocodiles and,
histologically, runt crocodiles appeared to have more iron stores in the spleen than normal crocodiles.
Bone marrow histological examination was pursued to look for pathology of bone marrow, where red
blood cells are produced, that could explain the decrease of red blood cells in circulation. However,
apart from an overall decrease in cellularity of the marrow in runt crocodiles, the erythroid cell line
was unremarkable, with both runt and normal crocodiles having abundant red blood cell precursors in
the marrow.
This pathological investigation into possible causes of runting in crocodiles has been relatively
extensive. There is so little published material on this subject that comparison to existing data is brief.
McInerney (1994) looked at serum liver enzymes and reported a few minor differences although
statistical significance were not presented. Foggin (1987, 1992) has perhaps dealt with the subject of
pathological changes in runts most extensively, in a few paragraphs in general chapters on disease.
Some of the findings reported were also found in the present study (for example, anaemia and low
total serum protein in runts and general atrophy involving the fat body, pancreas and liver) whilst
others were not (ascites and intestinal villus atrophy in runts). Foggin (1987, 1992) also speculated
that runt crocodiles were affected relatively more from miscellaneous infectious diseases, possibly
due to an immune deficiency. Although there was no evidence of increased infectious disease in runt
crocodiles in the present study, there was evidence of immune deficiency in the form of marked
lymphoid atrophy. However, the runt crocodiles that were sampled in this study were not obviously
sick (apart from being relatively small) when sampled. The overall decreased survival of runt
crocodiles seems likely to be partially a result of their eventually succumbing to opportunistic
infections as they become increasingly compromised by inanition and stress.
38
6. Implications
Runtism, and ways to overcome this syndrome, should be set as the number one mortality research
priority. The histopathology study has shown that there are no known infectious diseases causing the
onset of this syndrome, although immunosuppression, evidenced by marked lymphoid atrophy,
suggests that there could be some insidious unknown cause, such as a virus. In addition, ways to
decrease chronic stress (vacuolated adrenocortical cells) in these animals should be further
investigated. However, the significant collection area and pair effects (including the large heritability
estimate) also suggest an underlying genetic predisposition to this syndrome.
The major disease diagnosed in crocodiles in the “disease-related” category of mortality was
septicaemia caused by ubiquitous gram-negative bacteria. The predominant bacteria found on the
farm were Providencia rettgeri, Salmonella sp., Edwardsiella tarda, Morganella morganii,
Psuedomonas sp., Streptococcus sp., Escherichia coli and Proteus vulgaris, although others were also
found sporadically (e.g. Klebsiella sp., Citrobacter sp., Enterobacter sp.). The majority of these
disease outbreaks, if diagnosed early, are contained quickly with sulphafurazole or tetracycline
antibiotics. Selection of breeding animals for disease-resistance is achievable using the models
presented in this study.
Deaths caused from no visible ailments (that is, no notable management disturbance, disease outbreak,
etc.) is an area of considerable concern and should be seen as the second major research priority.
These animals are always in good condition and present a negative bacterial culture when submitted
to the lab. The sporadic use of probiotics (started in 2006 and used more regularly in 2007) has
reduced the incidence of NVA deaths but the underlying cause is yet to be realised. Current thinking
indicates that viral disease(s) may be responsible and further investigation into viral isolation and
characterisation needs to be undertaken.
The difference in CBVs for the different causes of deaths will allow producers to have a more
proactive approach to implementing their genetic improvement programs. The higher risk mortality
categories (Runtism, NVA and disease-related) can be incorporated into their CrocPLAN programs,
whilst lower risk causes can be omitted until their economic impact, and thus priority status, are
changed.
39
7. Further research
The survival analyses have set the research priorities for improving on-farm mortality rates as 1)
runtism, 2) no visible ailments, 3) disease-related, 4) stress-related and 5) congenital deaths.
Histopathology revealed runtism to be associated with marked lymphoid atrophy. Substantiation of
immunosuppression and possible primary causes should be further investigated. One obvious area
where limited research has been conducted is virus isolation and characterisation in relation to
crocodile deaths. Crocodile pox virus has been described in a few crocodilians (Chordopoxvirinae;
Afonso et al 2006) as well as West Nile virus (Flaviviridae; Jacobsen et al. 2005) and Adenovirus
(Huchzermeyer 2003). Unlike bacteria, to culture viruses, live cell lines are required. Although cell
lines have been developed for the American alligator (Huchzermeyer 2003), a cell line for saltwater
crocodiles (Crocodylus porosus) is yet to be established. However, RIRDC have already
commissioned a study (RIRDC Project Number PRJ-002461), to begin in late 2008 in collaboration
with BVL, Porosus Pty Ltd and the University of Sydney, to establish an embryonic saltwater
crocodile cell line to identify potential viral diseases (BVL) and to identify endogenous retroviruses
(University of Sydney). The identification of infectious viral agents may also aid in determining the
cause of deaths from no visible ailments.
In addition to determining if viruses are infectious agents causing the large number of undiagnosed
deaths, a further understanding of the crocodilian immune system is required. Initial work has been
conducted on the serum complement activity in Crocodylus porosus and freshwater crocodiles
(Crocodylus johnstoni; Merchant and Britton 2006) and other innate immune functions in other
crocodilians (Merchant et al. 2006). However, none of these studies have so far factored age, clutch or
area effects into the analysis to look at inter- and intra-clutch variation of the innate immune system in
captive environments. Furthermore, negligible research has been done to investigate the role and
function of the acquired immune system.
Vacuolated adrenocortical cells indicated chronic stress in the runt crocodiles, a histological finding
that was later verified by the elevated corticosterone levels in runt compared to normal crocodiles.
What causes the chronic stress in these animals compared to the vast majority? Obviously in
commercial situations, animals are accommodated to minimise the stress in the majority of animals.
These runts, although they constitute the minority of overall animals (5% in total cohort), still have a
large economic impact worthy of further investigation. Other than immunology discussed above,
areas of potential investigation include manipulation of the captive environment (e.g. pen design
(RIRDC Project Number WMI-4A), diet, etc), ethology and endocrinology. Limited studies have
already started to investigate some of these (Anderson et al. 1990, Kanui et al. 1993, Peucker and
Mayer 1995).
The confounding clutch effect with limited years warrants continued investigation. Although the rank
of each pair effect is unlikely to change, the Pair model heritability estimates are upwardly biased due
to the limited years of data collected. Data will continue to be collected in the manner described
herein as it has provided a useful management tool for comparing performance over years and
forecasting events.
The significant effect of NoHatch in the survival analyses reported herein illustrates the post-hatching
effect of poor embryo survival. An investigation to model embryo survival using similar models
described in this study would be advantageous to identify critical stages in embryo development. Data
are already being collected at Darwin Crocodile Farm for this purpose.
40
8. References
Afonso CL, Tulman ER, Delhon G, Lu Z, Viljoen GJ, Wallace DB, Kutish GF, Rock DL. Genome of
crocodilepox virus. Journal of Virology 2006:80;4978-4991.
Anderson O, Kimwele C, Aulie A, Kanui T. Effects of recombinant human growth hormone in
juvenile Nile crocodiles (Crocodylus niloticus). Comparative Biochemistry and Physiology
1990:97A;607-609.
Bishop EL, Shilton C, Benedict S, Kong F, Gilbert GL, Gal D, Godoy D, Spratt BG, Currie BJ.
Necrotizing fasciitis in captive juvenile Crocodylus porosus caused by Streptocossus agalactiae:
an outbreak and review of the animal and human literature. Epidemiology and Infection
2007:135;1248-1255.
Buenviaje GN, Ladds PW, Melville L, Manolis SC. Disease-husbandry associations in farmed
crocodiles in Queensland and the Northern Territory. Australian Veterinary Journal 1994;71:165-
173.
Campbell T. Hematology of birds. In: Thrall MA editor. Veterinary Haematology and Clinical
Chemistry. Lippincott Williams & Wilkins, Baltimore, 2004:225-258.
Capen CC. Endocrine glands. In: Maxie MG editor. Jubb, Kennedy, and Palmer’s Pathology of
Domestic Animals, Volume 5. Elsevier, Philadelphia, 2007:325-428.
Chirinos Z, Carabano MJ, Hernandez D. Genetic evaluation of length of productive life in the Spanish
Holstein-Friesian population. Model validation and genetic parameters estimation. Livestock
Science 2007:106;120-131.
Department of Natural Resources, Environment, and the Arts. Management plan for Croocdylus
porosus in the Northern Territory 2005-2010. pp. 25.
http://www.nt.gov.au/nreta/wildlife/programs/approved.html
Divers SJ. Reptilian renal and reproductive disease diagnosis. In: Fudge AM editor. Laboratory
medicine, avian and exotic pets, W.B. Saunders, Philadelphia, 2000a:217-222.
Divers SJ. Reptilian liver and gastrointestinal testing. In: Fudge AM editor. Laboratory medicine,
avian and exotic pets, W.B. Saunders, Philadelphia, 2000b:205-209.
Ducrocq, V. and J. Sölkner. ''The Survival Kit'', a FORTRAN package for the analysis of survival
data. Proceedings of the 5th World Congress on Genetics Applied to Livestock Production,
Ontario, Canada. 1994:XXII;51-52.
Ducrocq, V. and J. Sölkner. ''The Survival Kit -- V3.0'', a package for large analyses of survival data.
Proceedings of the 6th World Congress on Genetics Applied to Livestock Production, Armidale,
Australia. 1998:XXVII;447-448.
Duncan JR, Prasse KW, Mahaffey EA. Veterinary laboratory medicine, clinical pathology, 3rd
edition. Iowa State University Press, Ames, 1994.
Eady SJ, Garreau H, Gilmour AR. Heritability of resistance to bacterial infection in meat rabbits.
Livestock Science 2007:112;90-98.
Edwards HE, Burnhan WM. The impact of corticosteroids on the developing animal. Pediatric
research 2001:50:433-440.
41
Elsey RM, Joanen T, McNease L, Lance V. Growth rate and plasma corticosterone levels in juvenile
alligators maintained at different stocking densities. Journal of Experimental Zoology
1990:255;30-36.
Foggin CM. Diseases and disease control on crocodile farms in Zimbabwe. In: Webb GJW, Manolis
SC, Whitehead PJ editors. Wildlife management: crocodiles and alligators. Surrey Beatty & Sons,
Chipping Norton, 1987:351-362.
Foggin CM. Diseases of farmed crocodiles. In: Smith GA, Marais J editors. Conservation and
Utilization of the Nile Crocodile in Southern Africa. Handbook on Crocodile Farming, The
Crocodile Study Group of Southern Africa, Pretoria. 1992:107-140.
Hibberd EMA, Pierce RJ, Hill BD, Kelly MA. Diseases of juvenile estuarine crocodiles, Crocodylus
porosus. Proceedings of the 13th Working Meeting of the IUCN Crocodile Specialist Group,
Argentina, 13-17 May 1996:303-312.
Hochberg Z. Mechanisms of steroid impairment of growth. Hormone research 2002:58(suppl 1);33-
38.
Huchzermeyer FW. Crocodiles: biology, husbandry and Diseases. CABI Publishing, UK,
2003:pp337.
Isberg SR, Thomson PC, Nicholas FW, Barker SG, Moran C. A genetic improvement program for
farmed saltwater crocodiles. Rural Industries Research and Development Corporation, Canberra.
2004:http://www.rirdc.gov.au/reports/NAP/04-147.pdf.
Isberg, S.R., Thomson, P.C., Nicholas, F.W., Barker, S.G. and Moran, C. Quantitative analysis of
production traits in saltwater crocodiles (Crocodylus porosus): III. Juvenile survival. Journal of
Animal Breeding and Genetics 2006:123; 44-47.
Jacobsen ER, Ginn PE, Troutman JM, Farina L, Stark L, Klenk K, Burkhalter KL, Komar N. West
nile virus infection in farmed American alligators (Alligator mississippiensis) in Florida. Journal
of Wildlife Diseases 2005:41;96-106.
Kanui T, Kimwele C, Aulie A. Influence of recombinant bovine growth hormone on growth and feed
intake in juvenile Nile crocodiles (Crocodylus niloticus). Comparative Biochemistry and
Physiology 1993:106A; 381-384.
Ladds PW, Bradley J, Hirst RG. Providencia rettgeri meningitis in hatchling saltwater crocodiles
(Crocodylus porosus). Australian Veterinary Journal 1996:74;397-398.
Ladds PW, Sims LD. Diseases of young captive crocodiles in Papua New Guinea. Australian
Veterinary Journal 1990:67;323-330.
Lance VA, Morici LA, Elsey RM. Physiology and endocrinology of stress in crocodilians. In: Grigg
GC, Seebacher F, Franklin CE editors. Crocodilian Biology and Evolution, Surrey Beatty & Sons,
Chipping Norton, 2000:327-340.
Lloyd M, Morris PJ. Phlebotomy techniques in Crocodilians. Bulletin of the Association of Reptilian
and Amphibian Veterinarians. 1999:9;12-14.
Mader, D.R. 2000. Reptilian Metabolic Disorders. In: Laboratory medicine, Avian and Exotic Pets,
Fudge, A.M. (Ed). W.B. Saunders, Philadelphia PA. Pp. 210-216.
Mayer, R. Crocodile farming. Research, development and on-farm monitoring. Rural Industries
Research and Development Corporation, Canberra. Project Number DAQ-188A. 1998:pp92
42
McInerney J. Liver enzymes and pathology in runt crocodiles (C. porosus). Proceedings of the
Association of Reptilian and Amphibian Veterinarians. 1994:57-58.
Merchant M, Britton A. Characterisation of serum complement activity of saltwater (Crocodylus
porosus) and freshwater (Crocodylus johnstoni) crocodiles. Comparative Biochemistry and
Physiology 2006:143A;488-493.
Merchant M, Mills K, Leger N, Jerkins E, Vliet K, McDaniel N. Comparison of innate immune
activity of all known living crocodilian species. Comparative Biochemistry and Physiology
2006:143B;133-137.
Peucker S, Mayer R. Runt observation studies. Crocodile Research Bulletin. 1995;1:57-62.
Richardson KC, Webb GJW, Manolis SC. 2002. Crocodiles: Inside Out. Surrey Beatty and Sons:
Chipping Norton, Australia.
Ricklefs RE, Cadena CD. Heritability of longevity in captive populations of nondomesticated
mammals and birds. The Journals of Gerontology. 2008:63A;435-446.
Riggio V, Finocchiaro R, Bishop S C. Genetic parameters for early lamb survival and growth in
Scottish Blackface sheep. Journal of Animal Science 2008:86;1758-1764.
Schmidt RE, Reavill DR, Phalen DN. Lymphatic and haematopoietic System. In: Pathology of pet
and aviary birds. Iowa State University Press, Ames, 2003:131-147.
Schobitz B, Reul JMHM, Holsboer F. The role of the hypothalamic-pituitary-adrenocortical system
during inflammatory conditions. Critical Reviews in Neurobiology 1994:8;263-291.
Southey, B.R., S.L. Rodriguez-Zas, and K.A. Leymaster. Survival analysis of lamb mortality in a
terminal sire composite population. Journal of Animal Science 2001:79;2298-2306.
Southey, B.R., S.L. Rodriguez-Zas, and K.A. Leymaster. Competing risks analysis of lamb mortality
in a terminal sire composite population. Journal of Animal Science 2004:82;2892-2899.
Steyermark AC, Spotila JR. Effects of maternal identity and incubation temperature on snapping turtle
(Chelydra serpentine) metabolism. Physiological and Biochemical Zoology 2000:73;298-306.
Thompson K. Bones and joints. In Maxie MG editor. Jubb, Kennedy, and Palmer’s Pathology of
Domestic Animals, Volume 5, Elsevier, Philadelphia, 2007:1-184.
Turton JA, Ladds PW, Manolis SC, Webb GJW. Relationship of blood corticosterone,
immunoglobulin and haematological values in young crocodiles (Crocodylus porosus) to water
temperature, clutch of origin and body weight. Australian Veterinary Journal 1997;75:114-119.
Webb G. The crocodile as a production unit. In: Proceedings of the Intensive Tropical Animal
Production Seminar. Townsville (Australia), 19-20 July 1989.
Webster EL, Torpy DJ, Elenkov IJ, Chrousos GP. Corticostropin-releasing hormone and
inflammation. Annals of the New York Academy of Sciences 1998:840;21-32.
Welsh CS, Garrick DJ, Enns RM, Nicoll GB. Threshold model analysis of lamb survivability in
Romney sheep. New Zealand Journal of Agricultural Research 2006;49: 411-418.
Valli VEO. Hematopoietic system. In Maxie MG editor. Jubb, Kennedy, and Palmer’s Pathology of
Domestic Animals, Volume 5, Elsevier, Philadelphia, 2007:107-324.
43
Improving Australias Crocodile Industry Productivity
— Understanding runtism and survival—
RIRDC Publication No. 09/135
BySally Isberg, Cathy Shilton and Peter Thomson
This project assessed the incidence of different causes of
juvenile saltwater crocodile deaths on an Australian crocodile
farm. In addition, a pilot histopathology study was conducted
to determine if there are any primary causes for runting in
captive saltwater crocodiles.
This information is targeted at Australian crocodile producers
to enhance their production efficiency by reducing juvenile
mortalities, particularly from runting.
The Rural Industries Research and Development Corporation
(RIRDC) manages and funds priority research and translates
results into practical outcomes for industry.
Our business is about developing a more profitable, dynamic
and sustainable rural sector. Most of the information we
produce can be downloaded for free or purchased from our
website: www.rirdc.gov.au, or by phoning 1300 634 313 (local
call charge applies).
Contact RIRDC:
Level 2
15 National Circuit
Barton ACT 2600
PO Box 4776
Kingston ACT 2604
Ph: 02 6271 4100
Fax: 02 6271 4199
Email: rirdc@rirdc.gov.au
web: www.rirdc.gov.au
Most RIRDC books can be freely downloaded
or purchased from www.rirdc.gov.au or
by
phoning 1300 634 313 (local call charge
applies).
www.rirdc.gov.au
ResearchGate has not been able to resolve any citations for this publication.
Book
Full-text available
Profit from farmed crocodile is essentially a function of the returns and costs from the average lifetime productivity of the herd. The aim of a genetic improvement program is to improve the total economic value of the herd, and consequently maximise profit. To date, no research has been conducted to evaluate the potential of a genetic improvement program in the Australian crocodile industry. By implementing a selection program based on reproductive performance, juvenile growth rates and juvenile survival rates, the resultant superior breeding animals will increase the profitability of crocodile farms. The major benefits to the industry will be decreasing overhead costs by growing animals to marketable size in a quicker time, increasing profitability by offsetting some of the production costs per animal and increasing the number of animals obtained from the farm each year. The major aim of this project was to create a practical genetic improvement program for immediate adoption by the Australian crocodile industry, to be called CrocPLAN.
Article
Full-text available
Chapter
Normal Structure Immunity Disease Additional Reading
Article
Lamb survivability is a major determinant of the reproductive efficiency of the flock. The objective of this study was to estimate variance components for direct, maternal, and permanent environmental components of lamb survivability. Threshold model procedures were used to estimate variance components from 25 362 lamb survival observations representing 136 sires from two Romney flocks in New Zealand with an average lamb survival of 80.8%. Lambs were born between 1997 and2000. Fixed effects fitted in the model included sex and contemporary group, defined as the interaction of age of dam, flock, and birth date class. Median heritabili‐ties were 0.106 ± 0.010 for direct, 0.082 ± 0.005 for maternal, and a median proportion of phenotypic variation 0.098 ± 0.003 for permanent environmental effects. The direct‐maternal correlation was –0.75 ± 0.03. Estimated breeding values on the underlying scale were back‐transformed to the phenotypic probability scale. For direct lamb survivability, the highest performing sire's lambs had a 14.8% greater chance of survival than the lowest performing sire's lambs for flocks with an average lamb survivability of 80%. Although heritabilities for both components of lamb survivability are low, selection for this trait can be accomplished and is of economic importance for the sheep breeder.
Article
Incidence of visual signs of bacterial infection and mortality, from causes related to bacterial infection, were recorded on a weekly basis in growing meat rabbits from 5 to 10 weeks of age. Heritability of Weekly Incidence of disease was highest in weeks 9 and 10 (0.05±0.02 and 0.06±0.02, respectively with linear model, and 0.10±0.06 and 0.12±0.05, respectively with a threshold model). Common litter effects accounted for 5–20% of the variance of disease incidence, while maternal genetic variance was small (0–3%). Individuals from small litters at weaning had higher disease incidence, and disease incidence reduced as litter parity of the doe increased (P
Article
1.1. Slow-growing juvenile Nile crocodiles were injected with recombinant bovine growth hormone (rbGH) once a week for 6 weeks and then re-treated after 4 weeks.2.2. The feed intake of the treated crocodiles was 26 g/kg/meal during the three periods, while the intakes of the controls were 15, 20 and 2 g/kg.3.3. The treated gained 2.3 and 0.9%/week in weight during the first and second injection period and the feed conversion efficiencies were 28 and 13%. The controls lost weight.4.4. The treated animals grew at rates of 0.98 and 0.43%/week during the first and the second injection period.5.5. Bovine GH enhances growth in juvenile crocodiles and seems to have less adverse effects than human GH.
Article
Sixty-six captive-reared, juvenile alligators maintained in environmentally controlled chambers were randomly assigned to four treatments consisting of stocking densities of 0.35, 0.18, 0.12, and 0.09 m2/alligator. Alligators were maintained at these densities for 3.5 months. At approximately 4 week intervals the alligators were weighed, measured, and bled via cardiac puncture. Duplicate aliquots of plasma were assayed for corticosterone. All of the alligators continued to grow during the experiment, but alligators maintained at the lowest stocking density were significantly heavier and grew significantly faster (P < 0.01) and had significantly lower plasma corticosterone (P < 0.01) than alligators maintained at the highest stocking density. These results indicate that crowding (i.e., high-density rearing) of juvenile alligators inhibits maximum growth and that the subsequent stress associated with this crowding results in chronically elevated plasma corticosterone. Plasma corticosterone showed a strong negative correlation with change in body weight; the faster the rate of growth, the lower the hormone levels. These results also show that a single blood sample taken at monthly intervals is sufficient to monitor chronically elevated plasma corticosterone in this species.