- Email: [email protected]
Quantification of plasma corticosterone in juvenile farmed saltwater crocodiles (Crocodylus porosus) using current Australian Code of Practice guidelines
Accepted Manuscript Quantification of plasma corticosterone in juvenile farmed saltwater crocodiles (Crocodylus porosus) using current Australian Code of Practice guidelines Sally R. Isberg, John W. Finger Jr., Peter C. Thomson PII: DOI: Reference:
S0016-6480(18)30206-5 https://doi.org/10.1016/j.ygcen.2018.08.020 YGCEN 13019
To appear in:
General and Comparative Endocrinology
Received Date: Revised Date: Accepted Date:
29 March 2018 6 August 2018 22 August 2018
Please cite this article as: Isberg, S.R., Finger, J.W. Jr., Thomson, P.C., Quantification of plasma corticosterone in juvenile farmed saltwater crocodiles (Crocodylus porosus) using current Australian Code of Practice guidelines, General and Comparative Endocrinology (2018), doi: https://doi.org/10.1016/j.ygcen.2018.08.020
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Quantification of plasma corticosterone in juvenile farmed saltwater crocodiles (Crocodylus porosus) using current Australian Code of Practice guidelines
Sally R. Isberg1,2, John W. Finger Jr.2,3,4 and Peter C. Thomson2 1
Centre for Crocodile Research, PO Box 329, Noonamah, NT 0837 Australia
2
School of Life and Environmental Sciences, Faculty of Veterinary Sciences, University of Sydney,
NSW 2006, Australia 3
Department of Environmental Health Science, University of Georgia, Athens, GA, 30602, USA
4
Department of Biological Sciences, Auburn University, Auburn, AL, 36849, USA
* Corresponding author email: [email protected]
Abstract Saltwater crocodiles (Crocodylus porosus) across three size categories (hatchlings, grower and harvest-size) were repeatedly blood sampled on two farms in the Northern Territory, Australia to determine reference plasma corticosterone (CORT; crocodilian stress hormone) levels. The mean CORT values for hatchlings (<1 year old), growers (1-3 years) and harvest-size individuals (2+ years) were 1.65 ± 0.15 ng/ml, 2.73 ± 0.21 ng/ml and 2.19 ± 0.16 ng/ml, respectively. No inter-farm differences within the hatchling or harvest-size crocodiles were detected, but growers on Farm 2 had significantly lower plasma CORT than those on Farm 1. However, the grower growth rate coefficients were the same across both farms so the repeated blood sampling design most likely contributed to the difference in CORT values rather than any management procedures. Plasma corticosterone levels significantly increased with time of day. Substantial variation in plasma CORT was observed at each sampling which is not unprecedented in the literature but requires further elucidation. Irrespective, as CORT values were generally low, our results suggest that the farming environment and husbandry practices, as implemented under the Australian industry Code of Practice, are effective as baseline animal welfare measures although they should be viewed as a foundation for further welfare research and not considered static.
Keywords: saltwater crocodile; production; corticosterone; welfare
Introduction In northern Australia, saltwater crocodiles are farmed for their belly skin, with defects reducing the skins’ value. As an emergent industry (Isberg et al., 2004), many of the factors affecting production are unknown. For example, the effects of on-farm stressors, such as those associated with management or housing, on growth or health have been largely un-investigated. Moreover, relatively little information is available in regards to animal welfare standards although Codes of Practice have been developed for ethical management and propagation. It is customary on Australian crocodile farms to house crocodiles in three size categories: hatchlings, grow out (or grower), and harvest size. Previously, Finger et al. (2015; 2016) investigated plasma corticosterone (CORT, the main reptilian stress hormone), growth and immune function in farmed hatchling crocodiles. They reported low CORT levels that were not associated with detrimental effects on growth or immune function. In fact, while CORT levels were highly variable at the first out of three samplings, the variability decreased over time indicating that hatchlings habituated to the farming environment. Importantly, their results endorsed the management strategies adopted within the Code of Practice on the humane treatment of wild and farmed Australian crocodiles (NRMMC, 2009). In addition to these studies in hatchlings, a few studies have reported the CORT levels of crocodiles in the finishing phase of production (Franklin et al., 2003; Isberg and Shilton, 2013; Pfitzer et al., 2014). During the finishing phase of production, crocodiles are often housed individually to allow any skin defects to heal, without the risk of more being added by conspecifics, to ensure the hide is accepted into the export market (Isberg and Shilton, 2013). Previous research has provided evidence that individual housing is not detrimental, as CORT levels are similar between communallyand individually-housed harvest-size crocodiles (1.6-1.8 m; Isberg and Shilton, 2013). However, except for studies on capture techniques (Franklin et al., 2003; Pfitzer et al., 2014), there have been no other studies investigating welfare in harvest-size individuals. Moreover, there have been no studies on the welfare of grower crocodiles. As such, there is a deficit of information regarding the welfare of farmed crocodiles especially in relation to the intermediate size classes. Furthermore, no study has been conducted on the larger size cohorts to repeatedly measure CORT and thus, disentangle seasonal influences and the effect on growth. Most importantly, no study has repeatedly measured CORT between different farms. all of which are obliged to adhere to the Code of Practice on the humane treatment of wild and farmed Australian crocodiles (hereafter, simply Code of Practice; NRMMC, 2009) although management regimes may differ. Therefore, this study aims to quantify baseline CORT levels, and thus provide a basis for welfare standards, across the different size categories within the Australian saltwater crocodile production system.
Methods Experimental animals Crocodiles were sampled from two crocodile farms in the Northern Territory of Australia. Both farms strictly adhere to the requirements within the Code of Practice (NRMMC, 2009). This Code describes minimum standards for hatchlings and raising stock. In this study, hatchlings were similarly defined as being less than one year of age (<1yo) but we divided the raising stock into growers and harvest-size crocodiles. Growers were defined to be between one and three years old whereas the time that a crocodile enters an individual pen is determined by its size and can be anywhere from two years onwards (2+yo; 1.6+ m). These animals are referred to herein as harvestsize crocodiles. The Code of Practice (NRMMC, 2009) specifies hatchling stocking densities should not exceed 10-15 individuals/m2. The hatchling pen design on Farm 1 was described by Finger et al. (2013, 2015, 2016). Thirty-five crocodiles were stocked into pens that were 116.5 cm wide and 209.5 cm long (14 crocodiles/m2) within a larger shed on Farm 1. After the second sampling, to reduce the biomass within each pen, the larger crocodiles were moved into larger pens (182 cm wide 300 cm long) and stocked at 55 crocodiles (10 crocodiles/m2). Although only reflected ambient light can enter the sheds, the rear of each pen was covered to provide crocodiles a hiding area and to retain heat whilst the front was an open-air feeding platform. Pen water was heated to 32°C and delivered ondemand from jets into the water body. On Farm 2, 70 hatchlings (15 crocodiles/m2) were housed in completely enclosed pens (156 cm wide 301 cm long) to retain heat and maintain darkness. Pen water was heated to 33°C by coiled pipes within the water body. Both farms fed their hatchlings 5 times per week in excess with a red meat mince fortified with 2% vitamin/mineral premix (Monsoon Crocodile Premix, Winnellie, NT, Australia) and 1.5% calcium carbonate. Residual food was removed the following morning and pens were cleaned thoroughly with a chlorine-based detergent before being re-filled with clean water. For growers approximately one metre long, the Code of Practice (NRMMC, 2009) specifies 2-4 individuals/m2 decreasing to 1-2 individuals/m2 as crocodiles approach 2 m long. The pen designs for both the grower crocodiles in communal pens and harvest-size crocodiles in individual pens as well as the husbandry regimes for both farms have been previously described (see Isberg and Shilton, 2013). Briefly, grower crocodiles on both farms were housed in enclosed communal pens of a similar design (60:40 land:water ratio), with a stocking density of 1.75 crocodiles/m2. Harvest-size crocodiles in individual pens on each farm were also similar, with a density of 0.82 crocodiles/m2 (70:30 land:water ratio). Grower crocodiles were initially fed a coarse red meat mince fortified with the same vitamins/minerals as hatchlings. Once grower crocodiles became large enough, they were weaned onto chicken heads. All harvest-size crocodiles were fed chicken heads. Both categories were fed in
excess. The grower pens were fed twice per week in the dry season and three times per week in the wet season, while the harvest-size crocodiles were fed once and twice, respectively. The following morning, excess food was removed and cleaned with a sodium hypochloride solution before being refilled with clean water. No heating of the water occurs in these pens. Capture and Sampling The sampling period commenced on the June 15, 2012 and finished on October 18, 2013. On sampling days, which were approximately three months apart, a timer was started upon entrance to hatchling and grower pens. At the first sampling, five crocodiles from each pen were randomly captured by hand (<1.2 m) or by the use of an electrical stunner (>1.2 m) and immediately blood sampled from the occipital sinus (see Franklin et al., 2003) using either a 23- or 16-gauge needle depending on the size of the animal. At each successive sampling, these same individuals were resampled. For crocodiles housed in individual pens, a timer was started immediately prior to opening the gate. The cumulative time required to obtain a blood sample (CumTime, in seconds) post pen entrance (grower crocodiles) or post pen opening (harvest size crocodiles) was noted at each sampling (Finger et al., 2015) with a maximum time of 20 minutes in each pen. After collection, blood was placed in lithium heparin tubes, gently inverted three times, and placed on ice. At the first sampling, all crocodiles were sexed by cloacal palpation and measured for head length (HL; in mm) as described by Isberg et al. (2005). Hatchling and grower crocodiles were tagged on the 100 scute (Isberg et al., 2004) with either a small animal tag (size 1005-1; National Band and Tag Company, Kentucky, USA) or a yellow sheep ear tag (Allflex, Queensland, Australia) that was engraved with a unique identifier. This allowed repeated measurements to be made from individual crocodiles to assess the true effect of time of year and animal size measured as head length (HL; mm). To identify these animals, clean water was required which often delayed capture of some crocodiles. Since harvest-size crocodiles were housed individually they did not require tagging. Both grower and individual pen water temperatures were recorded (± 1°C; T TEC-7, Nordex Pty Ltd, South Australia) at the time of sampling. Cloacal temperatures (± 1°C; AI368, Electro Chemical Engineering Pty Ltd, Australia) of harvest-size crocodiles was also recorded at each sampling. Animals were released immediately after sampling. By January 1, 2013 (Sampling 4 onwards), the hatchlings had all been moved into grower pens so were re-categorised as growers. As growers on Farm 2 became large enough, they entered the harvest-size category. The number of crocodiles in each category are listed in Table 1. Corticosterone assays Within two hours, all blood samples were centrifuged at 15,000 rpm for two minutes. Plasma was aliquoted and frozen at -20°C until analysis. Plasma CORT was determined using a high-
sensitivity CORT EIA kit (IDS Ltd., Tyne & Wear, UK) following manufacturer-specific protocols as described previously (see Finger et al., 2015; Isberg and Shilton, 2013). Statistical analyses Restricted maximum likelihood (REML) was used to analyse the CORT levels (ng/ml) for each size category. For all size categories, CORT needed to be log-transformed to meet the assumptions of normality. The following linear mixed model was used: logeCORTijkl = + HLHLi + TCumTimeij + oC Tempij + Farmk + Samplingj + CleanDaysl + AnimalIDi + where is the overall mean; HLi is the head lengths of the ith individual; HL is the regression coefficient associated with HL; CumTimeij is the cumulative time to obtain the blood sample from the ith individual at the jth sampling after entering the pen; T is the regression coefficient associated with CumTime; depending on the model Tempij is either the water (grower or individual pen), air temperature (grower or individual pen; www.bom.gov.au) or cloacal temperature (individual pen only) of the ith individual at the jth sampling; oC is the regression coefficient associated with Temp; Farmk is the fixed effect of the kth farm (k = 1, 2); Samplingj is the fixed effect of the jth sampling (j = 1,…,7); CleanDaysl is the binary effect of number of days since cleaning (l = 0 or 1); AnimalIDi is the random effect of the ith individual and indicates the observations were repeated measures [assumed N(0,
)]; and is the random residual effect [assumed N(0,
)].
To determine the impact of CORT on growth, the following linear mixed model was used: HLijk = + CORTadjCORTadjij + Farmk + Samplingj + AnimalIDi + where all parameters are the same as specified above with the exception of CORTadj ij. Finger et al. (2015, 2016) showed that CORT response can be influenced by human presence with animals sampled later responding with higher CORT levels. To account for this, the variable CORTadjij can be created using the residuals from a linear regression of CORT against time added to the mean CORT value, then back-transforming these, i.e. CORTadjij = exp(resij + mean) (Finger et al., 2015, 2016). A 5% significance level was chosen to evaluate explanatory variables by backward elimination from the full model. The significance of the AnimalID variance component was evaluated using a likelihood ratio test with a 5% significance level. The assumption of equal variances across sampling was tested using a heterogeneous residual variance REML model by a likelihood ratio test. All results are reported as back-transformed means and as standard errors (SE). Least significant difference (5% LSD) tests were used to test differences between fixed effects.
Results
Hatchlings The results of hatchling crocodiles at Farm 1 were presented in Finger et al. (2015, 2016). When combined with those results from Farm 2, the average plasma CORT value for hatchlings was 1.65 ± 0.15 ng/ml (n = 152) across the two farms. CumTime significantly affected CORT levels (P < 0.001; Figure 1), but neither Farm (P = 0.48) nor hatchling size (HL; P = 0.16) affected CORT. Mean CORT at the different sampling periods was on the 5% significance threshold (P = 0.05; Figure 2) and although not significant, mean CORT at sampling 2 is higher than the other sampling periods. Hatchling pen water temperature is environmentally controlled so it was not included in the analysis. Growers A total of 594 samples were taken from 246 individuals with 162 and 84 from Farms 1 and 2, respectively. The average grower CORT level was 2.73 ± 0.21 ng/ml. CumTime significantly affected CORT levels (P < 0.001; Figure 1) but grower size (P = 0.71) had no effect on CORT levels. AnimalID had a significant effect (P < 0.05) on grower CORT. The experimental design employed was to recapture the same individuals across all sampling periods to assess whether individual animals return consistent CORT values. This proved to be detrimental because to re-identify a particular crocodile, clean water is required which meant that animals were mainly sampled on the day of cleaning (81% and 35% of crocodiles sampled on day of cleaning at Farm 1 and Farm 2, respectively). This resulted in a significant interaction (P < 0.001) between Farm and CleanDays, whereby animals sampled on the same day as cleaning on Farms 1 and 2 had mean plasma CORT levels of 8.72 ± 0.60 ng/ml and 2.90 ± 0.54 ng/ml, respectively. Comparatively, if the animals were sampled the day after cleaning from the same respective farm, the CORT levels were 3.70 ± 0.44 ng/ml and 0.60 ± 0.09 ng/ml. For those animals sampled on the day of cleaning, while all attempts were made to access the animals at least two and a half hours after the pens were full of water to allow a return to baseline, often the cleaning process took longer than expected. Consequently, time of day significantly (P < 0.001) affected the resultant CORT levels. To confirm the effect of cleaning, blood samples were taken from randomly-chosen grower crocodiles from different pens on Farm 1 before the clean when the water was still full (n = 22), after the water was drained but before cleaning (n = 21), after the pen was cleaned whilst there was still no water (n = 21), 2½ hours after the pen had been filled (n = 21), and finally the next morning a day after cleaning (n = 12). Figure 3 shows the CORT response corresponding to these cleaning stages. Dropping the water induced a 4-fold increase in plasma CORT with an additional 2.8-fold increase from the cleaning process which also occurs when the water is drained. These levels remained elevated even after 2.5 hours of full water but had declined to pre-cleaning levels by the following morning.
There were no seasonal (Sampling; P = 0.40; Figure 2) effects on grower CORT values although Figure 2 shows unequal variation across the sampling periods that could not be accounted for using the variables evaluated. Pen water temperature also did not significantly affect grower CORT values (P = 0.74) but given that the majority of the animals were sampled on the day of cleaning, the recorded water temperature reflects the temperature of the fresh water used to refill the pen and not the water temperatures experienced by the animals prior to cleaning. Harvest-size crocodiles housed in individual pens Samples were repeatedly taken from 98 individuals from Farm 1 with an additional 44 individuals included from Farm 2. A total of 608 samples were taken throughout the study period with a mean CORT level of 2.19 ± 0.16 ng/ml. CumTime had no effect on CORT levels (P = 0.32; Figure 1). Unlike animals in the communal pens, animals housed in individual pens were able to be sampled more expediently (mean CumTime = 74.4 seconds). The minimum amount of time to take a sample was 31 seconds and the longest time was 801 secs when the electrical stunner failed, although it should be noted that only seven samples took longer than 3 minutes (180 secs) to obtain. Time of day significantly affected CORT levels (P < 0.001), with CORT increasing the later in the day the sample was taken. Neither water (P = 0.42) nor cloacal (P = 0.93) temperature significantly affected CORT. Time of day was not significant in determining the cloacal temperature of harvest-size crocodiles (P = 0.08) but water temperature was significant (P < 0.001; CloacalTemp = 3.09 + 0.90×WaterTemp; Figure 4). There was no effect of length of time housed in an individual pen (P = 0.84) on CORT levels but there was a significant effect of AnimalID (P < 0.05). A heterogeneous residual variance model was required to account for the unequal residual variances between sampling periods. Figure 2 shows there was a significant decline in CORT values over the first four sampling periods (5% LSD) followed by a sudden rise at Sampling 5 (April-May 2013) when many of the Farm 2 growers were transferred into individual pens. CORT returned to previous levels at the 6th sampling (July-August 2013). Impact on growth Table 2 shows the linear regression coefficients of HL in the different size categories. The hatchlings at Farm 2 were larger at the initial sampling compared to those at Farm 1 and grew at a slightly faster rate (HL: Farm 1 = 0.40 cm/month; Farm 2 = 0.48 cm/month). Growers on the two farms are growing at the same rate (0.33 cm/month). Within the individual pens, Farm 1 harvest-size crocodiles were larger at the initial sampling but Farm 2 crocodiles were growing at 1.87 times the rate.
CumTime-adjusted CORT values were used as an explanatory variate for growth (i.e. HL) but in each size category had no effect (hatchlings P = 0.09; growers P = 0.62; harvest-size P = 0.73). On the other hand, farm and sampling periods were highly significant (P < 0.001) for all size categories including their interaction in both the grower and harvest-size crocodile analyses.
Discussion In this study, we repeatedly sampled the same juvenile saltwater crocodiles on two Northern Territory farms to understand the effects of seasonal and inter-farm variations on plasma CORT across size categories. This repeated sampling allowed us to examine if individual crocodiles displayed consistently high, low, or stochastic CORT values. Finger et al. (2015) reported stochastic CORT values in hatchling crocodiles that could not be attributed to AnimalID. The addition of hatchlings from another farm did not alter this outcome, suggesting that hatchling CORT is controlled by an, as of yet, elusive influence. In contrast, there was significant variation between crocodiles in both the grower and harvest size categories. This could represent better habituation of some animals compared to others in the captive situation. In the case of the grower crocodiles though, since the majority of observations were made on the day of cleaning, the AnimalID effect might actually represent an individual’s ability to tolerate the cleaning regime. For the harvest-size crocodiles housed in individual pens, the AnimalID effects might be more representative of long-term habituation which may have compromises on economic indicators such as blemish healing as discussed within Isberg and Shilton (2013). Analogous to the observation in hatchling crocodiles observed in this study and in Finger et al. (2015), CumTime had a significant effect on grower crocodile CORT even though all crocodiles were sampled within 3 minutes of capture. This suggests that the presence of people within a grower pen and/or the capture of conspecifics increases CORT levels (Finger et al. 2015). Because we repeatedly sampled the same individuals, the CumTime required to find these same individuals at successive samplings was higher than the first sampling. In fact, to rapidly identify individuals and prevent an even higher impact of CumTime on CORT, we sampled individuals on the same day as pen cleaning. Therefore, pen cleaning is an obvious confounder. Although CORT values from Farm 1 were higher than those from Farm 2 regardless of cleaning, this may have been influenced by the greater proportion of crocodiles on Farm 1 (81%) being sampled on cleaning days than on Farm 2 (35%). For those grower crocodiles sampled on the day of cleaning, all attempts were made to access the animals more than 2.5 hours after the pens were full of water, thereby allowing crocodiles time to settle after cleaning disturbances. But, as shown in Figure 3, the resultant elevated CORT levels due to cleaning did not return to baseline levels until the morning following cleaning. In addition, the later in the day the sample was taken, the longer the cleaning process had taken and the longer crocodiles
had been without full water. Consequently, time of day significantly affected plasma CORT values of grower crocodiles. In contrast to grower crocodiles, we were able to rapidly sample crocodiles housed in individual pens (see Franklin et al., 2003; Isberg and Shilton, 2013). As such, capture time had no effect on harvest-size crocodile CORT levels. Similar to grower crocodiles, however, CORT increased as the time of day progressed. This is in contrast to what has been observed in juvenile alligators. Lance and Lauren (1984) observed a biphasic peak in alligator CORT at 0800 hours and again at 2000 hours. Over this study period, the earliest and latest samples were taken at 0719 and 1628 hours, respectively, and we observed an increasing linear trend in CORT over the course of the day. However, this increase may not have been necessarily due to abiotic influences (i.e., photoperiod, temperature, etc.) alone. In fact, increased CORT with increased time of day likely reflects the higher probability of human disturbance, as husbandry-related activities and human-crocodile interactions increased with increasing time of day as well. In contrast to grower crocodiles, since individual pens were not cleaned on sampling days, it is unlikely that cleaning affected CORT levels. It may be the case that various size/age classes of crocodilians display circadian differences in CORT release, but this requires future investigation. There were seasonal (Sampling) effects on harvest-size crocodile CORT, but not on grower or, contrary to Finger et al. (2015), hatchling CORT. Sampling was used to model the combined effect of abiotic environmental factors (i.e., ambient air temperature, photoperiod, etc.) of a particular time of year on CORT (Finger et al., 2015). However, using “Sampling” as a collective method for extracting environmental effects may not be optimal. Indeed, even though water temperature was kept constant (i.e., provided with heaters) for hatchling crocodiles, Sampling significantly affected CORT values as reported by Finger et al. (2015), although this was not substantiated herein when combined with Farm 2’s hatchling CORT data. Furthermore, despite a strong positive relationship between water and cloacal temperature (Figure 4), no effect of water temperature on grower and harvest-size crocodile CORT was observed. Whilst other environmental factors (i.e., photoperiod or air temperature) may have contributed to this Sampling effect, husbandry practices may have confounded this as well. Similar to previous results (Isberg and Shilton, 2013), the time a crocodile spent in an individual pen did not affect CORT levels. However, using Sampling as the predictor variable, there was a significant decline in CORT values over the first four sampling periods which might indicate either habituation to individual housing or a seasonal decline in CORT. The increase in CORT at Sampling 5 is explained by grower animals at Farm 2 being moved into individual pens, as CORT levels at Sampling 6 are similar to those at Sampling 4 (Figure 2).
Heterogeneous variance models were required to meet the assumptions of unequal variances across sampling periods for harvest-size crocodiles, and although not significant, there was still considerable variability in grower crocodiles as well (Figure 2). Large variation in crocodilian plasma CORT is not unprecedented in the literature. In fact, in our previous study we also had to account for variance heterogeneity (Finger et al., 2015). In a study investigating how CORT implants affect hatchling alligator health, CORT values for the control alligators ranged from 3.8-42.8 ng/ml (Morici et al., 1997). In Nile crocodiles, CORT values ranged from 5.15-118 ng/ml at initial capture (Pfitzer et al., 2014). Whilst the statistical models used herein were able to account for much of the variability, particularly caused by CumTime, there were also many factors that could not be included in a model, such as agonistic interactions or dietary status of an individual. More investigation is required to better understand the factors contributing to this variability. More so, although individual tolerances for habituation have been useful in modelling these data, the utility of non-invasive measures such as faecal CORT could prove more useful (Ganswindt et al., 2014). Ultimately, one of the best indicators of poor management or chronic stress is its impact on economic predictors such as reduced growth rate (Elsey et al., 1990; Morici et al., 1997). Since high CORT values have been associated with reduced growth in many crocodilian species (e.g., Elsey et al., 1990) including saltwater crocodiles (Turton et al., 1997; Shilton et al., 2014), we examined the effect of time-adjusted CORT values on grower and harvest-size crocodile HL (Table 2). Similar to our previous hatchling study (Finger et al., 2015), we observed no effect of CumTime-adjusted CORT values on both growers and harvest-size crocodile growth suggesting that CORT levels are not suppressing growth. Furthermore, housing hatchlings in either complete darkness (Farm 2) or in sheds with reflected ambient light (Farm 1) had no effect on growth or plasma CORT.
Conclusions This is the first study to repeatedly sample juvenile farmed saltwater crocodiles from all size categories to assess CORT as a proxy of welfare. Since CORT values were generally low, along with no impact observed on growth, these results suggest that the farming environment and husbandry practices, as implemented under the Australian industry Code of Practice on the humane treatment of wild and farmed Australian crocodiles (NRMMC, 2009), are acceptable. Our results should provide a foundation for further welfare research, including the exploration of less invasive methods to assess the impact of husbandry.
Acknowledgements: We would like to extend a special thanks to Darwin and Lagoon Crocodile Farms for providing access to crocodiles used in this study. We would also like to thank Sue Aumann for help with ELISA kits
and to Berrimah Veterinary Laboratories for providing lab space for performing ELISAs. This project was partially funded by a grant from the Rural Industries Research and Development Corporation, Australia and protocols were approved by the University of Sydney Animal Ethics Committee (approval number: N00/5-2012/3/5729). JWF was funded by the Interdisciplinary Toxicology Program at the University of Georgia and the former Faculty of Veterinary Science, University of Sydney.
Conflicts of interest: The authors declare no conflicts of interest. References Elsey, R.M., Joanen, T., McNease, L., Lance, V., 1990. Growth rate and plasma corticosterone levels in juvenile alligators maintained at different stocking densities. J. Exp. Zool. 255, 30–36. Finger, J. W. Jr., Adams, A. L., Thomson, P. C., Shilton, C. M., Brown, G. P., Moran, C., Miles, L. G., Glenn, T. C., Isberg, S. R. 2013. Using phytohaemagglutinin to determine immune responsiveness in saltwater crocodiles (Crocodylus porosus). Aust. J. Zool. 61, 301-311. Finger, J.W. Jr., Thomson, P.C., Adams, A.L., Benedict, S. Moran, C., Isberg, S.R., 2015. Reference levels for corticosterone and immune function in farmed saltwater crocodiles (Crocodylus porosus) hatchlings using current Code of Practice guidelines. Gen. Comp. Endocrinol. 212, 63-72. Finger, J.W. Jr., Thomson, P.C., Isberg, S.R., 2016. Unexpected lower testosterone in faster growing farmed saltwater crocodile (Crocodylus porosus) hatchlings. Gen. Comp. Endocrinol. 226, 1-4. doi: 10.1016/j.ygcen.2015.11.016. Franklin, C.E., Davis, B.M., Peucker, S.K.J., Stephenson, H., Mayer, R., Whittier, J., Lever, J., Grigg G.C., 2003. Comparison of stress induced by manual restraint and immobilisation in the estuarine crocodile, Crocodylus porosus. J. Exp. Zool. 298, 86-92. Ganswindt, S.B., Myburgh, J.G., Cameron, E.Z., Ganswindt, A., 2014. Non-invasive assessment of adrenocortical function in captive Nile crocodiles (Crocodylus niloticus). Comp. Biochem. Physiol. 177A, 11-17. Isberg S.R., Shilton, C.M., 2013. Stress in farmed saltwater crocodiles (Crocodylus porosus): no difference between individually- and communally-housed animals. SpringerPlus. 2, 381. doi:10.1186/2193-1801-2-381. Isberg, S.R., Thomson, P.C., Nicholas, F.W., Barker, S.C., Moran, C., 2004. Farmed saltwater crocodiles-A genetic improvement program. Rural Industries Research and Development Corporation Paper No. 04/147.
Lance, V.A., Elsey, R.M., Trosclair, P.L., 2015. Sexual maturity in male American alligators in southwest Louisiana. South Am. J. Herpetol. 10, 58–63. Lance, V.A., Lauren, D., 1984. Circadian variation in plasma corticosterone in the American alligator, Alligator mississippiensis and the effects of ACTH injections. Gen. Comp. Endocrinol. 54, 1-7. Morici, L.A., Elsey, R.M., Lance, V.A., 1997. Effects of long-term corticosterone implants on growth and immune function in juvenile alligators, Alligator mississippiensis. J. Exp. Zool. 279, 156-162. Morpurgo, B., Gvaryahu, G., Robinzon, B., 1992. Effects of population density, size, and gender on plasma testosterone, thyroxine, haematocrit, and calcium in juvenile Nile crocodiles (Crocodylus niloticus) in captivity. Copeia 4, 1023–1027. NRMMC (National Resource Management Ministerial Council), 2009. Code of Practice on the humane treatment of wild and farmed Australian crocodiles. NRMMC, Canberra. Pfitzer, S., Ganswindt, A., Fosgate, G.T., Botha, P.J., Myburgh, J.G., 2014. Capture of farmed Nile crocodiles (Crocodylus niloticus): comparison of physiological parameters after manual capture and after capture with electrical stunning. Vet. Rec. doi:10.1136/vr.102438 Rooney, A.A., Crain, D.A., Woodward, A.R., Guillette, L.J., 2004. Seasonal variation in plasma sex steroid concentrations in juvenile American alligators. Gen. Comp. Endocrinol. 135, 25–34. Shilton, C., Brown, G.P., Chambers, L., Benedict, S., Davis, S., Aumann, S., Isberg, S.R., 2014. Pathology of runting in farmed saltwater crocodiles (Crocodylus porosus) in Australia. Vet. Pathol. doi: 10.1177/0300985813516642 Turton, J.A., Ladds, P.W., Manolis, S.C., Webb, G.J., 1997. Relationship of blood corticosterone, immunoglobulin and haematological values in young crocodiles (Crocodylus porosus) to water temperature, clutch of origin and body weight. Aust. Vet. J. 75, 114-119.
Figure Legends Figure 1. The effect of cumulative time (CumTime; secs) to procure a blood sample for CORT analysis. Although all blood samples were taken within three minutes of capturing an individual, CumTime represents the cumulative time before the sample was taken after entering the pen. Black circles are hatchlings, open circles are growers and grey circles are harvest-size crocodiles in individual pens.
Figure 2. Box-plots for model-adjusted plasma corticosterone (CORTadj; ng/ml) showing medians, first and third quartiles, minimum and maximum values between sampling periods. A. Hatchlings; B. Growers; C. Harvest-size crocodiles housed in individual pens. * and # indicates a significant difference (P <0.05) from the previous sampling period.
Figure 3. CORT (ng/ml) response of grower crocodiles at various stages of the cleaning process. * indicates a significant difference (P <0.05) from baseline levels before cleaning.
Figure 4. Relationship between crocodile cloacal and water temperatures in harvest-size crocodiles in individual pens (r2 = 0.86; P <0.001).
Table 1. Summary of the number of hatchlings (<1 year old), growers (1-3 years old) and individual pen (2+ years old) crocodiles sampled at each sampling period (1-7) from Farm 1 and Farm 2. In between Samplings 3 and 4, the animals changed age-category whereby hatchlings became growers. Some grower animals from Farm 2 entered the individual pens.
Sampling 1 Months Jun-Jul Year Hatchlings Farm 1 40 Farm 2 Growers Farm 1 Farm 2 Individual pens Farm 1 98 Farm 2 -
2 Aug - Sep 2012
3 Nov - Dec
4 Feb
5 6 Apr - May Jul - Aug 2013
7 Oct
39 20
38 15
-
-
-
-
90 38
73 39
97 29
79 18
66 23
32 -
95 20
88 19
92 25
74 21
52 11
-
Table 2. Linear regression equations for predicting HL over sampling periods at Farm 1 and 2. Regression coefficients are given as 0 + xx where x is sampling. r2 values are also presented
Hatchling
Growers
Individual pens
Farm 1
5.69 + 1.20x; r2 = 0.99
11.44 + 1.02x; r2 = 0.99
23.54 + 0.38x; r2 = 0.99
Farm 2
10.27 + 1.44x; r2 = 1.00
11.76 + 1.02x; r2 = 0.96
22.46 + 0.71x; r2 = 0.99
Highlights of Paper:
We defined corticosterone in different size categories of farmed saltwater crocodiles The average plasma corticosterone (CORT) values for hatchlings and growers in communal pens were 1.65 ± 0.15 ng/ml and 2.73 ± 0.21 ng/ml, respectively, and for harvest-size crocodiles in individual pens was 2.19 ± 0.16 ng/ml No seasonal effects on CORT were observed Farmed saltwater crocodiles are not inherently stressed Validation of faecal CORT should be pursued in saltwater crocodiles as a noninvasive method for continually assessing the impact of husbandry regimes
S0016-6480(18)30206-5 https://doi.org/10.1016/j.ygcen.2018.08.020 YGCEN 13019
To appear in:
General and Comparative Endocrinology
Received Date: Revised Date: Accepted Date:
29 March 2018 6 August 2018 22 August 2018
Please cite this article as: Isberg, S.R., Finger, J.W. Jr., Thomson, P.C., Quantification of plasma corticosterone in juvenile farmed saltwater crocodiles (Crocodylus porosus) using current Australian Code of Practice guidelines, General and Comparative Endocrinology (2018), doi: https://doi.org/10.1016/j.ygcen.2018.08.020
This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.
Quantification of plasma corticosterone in juvenile farmed saltwater crocodiles (Crocodylus porosus) using current Australian Code of Practice guidelines
Sally R. Isberg1,2, John W. Finger Jr.2,3,4 and Peter C. Thomson2 1
Centre for Crocodile Research, PO Box 329, Noonamah, NT 0837 Australia
2
School of Life and Environmental Sciences, Faculty of Veterinary Sciences, University of Sydney,
NSW 2006, Australia 3
Department of Environmental Health Science, University of Georgia, Athens, GA, 30602, USA
4
Department of Biological Sciences, Auburn University, Auburn, AL, 36849, USA
* Corresponding author email: [email protected]
Abstract Saltwater crocodiles (Crocodylus porosus) across three size categories (hatchlings, grower and harvest-size) were repeatedly blood sampled on two farms in the Northern Territory, Australia to determine reference plasma corticosterone (CORT; crocodilian stress hormone) levels. The mean CORT values for hatchlings (<1 year old), growers (1-3 years) and harvest-size individuals (2+ years) were 1.65 ± 0.15 ng/ml, 2.73 ± 0.21 ng/ml and 2.19 ± 0.16 ng/ml, respectively. No inter-farm differences within the hatchling or harvest-size crocodiles were detected, but growers on Farm 2 had significantly lower plasma CORT than those on Farm 1. However, the grower growth rate coefficients were the same across both farms so the repeated blood sampling design most likely contributed to the difference in CORT values rather than any management procedures. Plasma corticosterone levels significantly increased with time of day. Substantial variation in plasma CORT was observed at each sampling which is not unprecedented in the literature but requires further elucidation. Irrespective, as CORT values were generally low, our results suggest that the farming environment and husbandry practices, as implemented under the Australian industry Code of Practice, are effective as baseline animal welfare measures although they should be viewed as a foundation for further welfare research and not considered static.
Keywords: saltwater crocodile; production; corticosterone; welfare
Introduction In northern Australia, saltwater crocodiles are farmed for their belly skin, with defects reducing the skins’ value. As an emergent industry (Isberg et al., 2004), many of the factors affecting production are unknown. For example, the effects of on-farm stressors, such as those associated with management or housing, on growth or health have been largely un-investigated. Moreover, relatively little information is available in regards to animal welfare standards although Codes of Practice have been developed for ethical management and propagation. It is customary on Australian crocodile farms to house crocodiles in three size categories: hatchlings, grow out (or grower), and harvest size. Previously, Finger et al. (2015; 2016) investigated plasma corticosterone (CORT, the main reptilian stress hormone), growth and immune function in farmed hatchling crocodiles. They reported low CORT levels that were not associated with detrimental effects on growth or immune function. In fact, while CORT levels were highly variable at the first out of three samplings, the variability decreased over time indicating that hatchlings habituated to the farming environment. Importantly, their results endorsed the management strategies adopted within the Code of Practice on the humane treatment of wild and farmed Australian crocodiles (NRMMC, 2009). In addition to these studies in hatchlings, a few studies have reported the CORT levels of crocodiles in the finishing phase of production (Franklin et al., 2003; Isberg and Shilton, 2013; Pfitzer et al., 2014). During the finishing phase of production, crocodiles are often housed individually to allow any skin defects to heal, without the risk of more being added by conspecifics, to ensure the hide is accepted into the export market (Isberg and Shilton, 2013). Previous research has provided evidence that individual housing is not detrimental, as CORT levels are similar between communallyand individually-housed harvest-size crocodiles (1.6-1.8 m; Isberg and Shilton, 2013). However, except for studies on capture techniques (Franklin et al., 2003; Pfitzer et al., 2014), there have been no other studies investigating welfare in harvest-size individuals. Moreover, there have been no studies on the welfare of grower crocodiles. As such, there is a deficit of information regarding the welfare of farmed crocodiles especially in relation to the intermediate size classes. Furthermore, no study has been conducted on the larger size cohorts to repeatedly measure CORT and thus, disentangle seasonal influences and the effect on growth. Most importantly, no study has repeatedly measured CORT between different farms. all of which are obliged to adhere to the Code of Practice on the humane treatment of wild and farmed Australian crocodiles (hereafter, simply Code of Practice; NRMMC, 2009) although management regimes may differ. Therefore, this study aims to quantify baseline CORT levels, and thus provide a basis for welfare standards, across the different size categories within the Australian saltwater crocodile production system.
Methods Experimental animals Crocodiles were sampled from two crocodile farms in the Northern Territory of Australia. Both farms strictly adhere to the requirements within the Code of Practice (NRMMC, 2009). This Code describes minimum standards for hatchlings and raising stock. In this study, hatchlings were similarly defined as being less than one year of age (<1yo) but we divided the raising stock into growers and harvest-size crocodiles. Growers were defined to be between one and three years old whereas the time that a crocodile enters an individual pen is determined by its size and can be anywhere from two years onwards (2+yo; 1.6+ m). These animals are referred to herein as harvestsize crocodiles. The Code of Practice (NRMMC, 2009) specifies hatchling stocking densities should not exceed 10-15 individuals/m2. The hatchling pen design on Farm 1 was described by Finger et al. (2013, 2015, 2016). Thirty-five crocodiles were stocked into pens that were 116.5 cm wide and 209.5 cm long (14 crocodiles/m2) within a larger shed on Farm 1. After the second sampling, to reduce the biomass within each pen, the larger crocodiles were moved into larger pens (182 cm wide 300 cm long) and stocked at 55 crocodiles (10 crocodiles/m2). Although only reflected ambient light can enter the sheds, the rear of each pen was covered to provide crocodiles a hiding area and to retain heat whilst the front was an open-air feeding platform. Pen water was heated to 32°C and delivered ondemand from jets into the water body. On Farm 2, 70 hatchlings (15 crocodiles/m2) were housed in completely enclosed pens (156 cm wide 301 cm long) to retain heat and maintain darkness. Pen water was heated to 33°C by coiled pipes within the water body. Both farms fed their hatchlings 5 times per week in excess with a red meat mince fortified with 2% vitamin/mineral premix (Monsoon Crocodile Premix, Winnellie, NT, Australia) and 1.5% calcium carbonate. Residual food was removed the following morning and pens were cleaned thoroughly with a chlorine-based detergent before being re-filled with clean water. For growers approximately one metre long, the Code of Practice (NRMMC, 2009) specifies 2-4 individuals/m2 decreasing to 1-2 individuals/m2 as crocodiles approach 2 m long. The pen designs for both the grower crocodiles in communal pens and harvest-size crocodiles in individual pens as well as the husbandry regimes for both farms have been previously described (see Isberg and Shilton, 2013). Briefly, grower crocodiles on both farms were housed in enclosed communal pens of a similar design (60:40 land:water ratio), with a stocking density of 1.75 crocodiles/m2. Harvest-size crocodiles in individual pens on each farm were also similar, with a density of 0.82 crocodiles/m2 (70:30 land:water ratio). Grower crocodiles were initially fed a coarse red meat mince fortified with the same vitamins/minerals as hatchlings. Once grower crocodiles became large enough, they were weaned onto chicken heads. All harvest-size crocodiles were fed chicken heads. Both categories were fed in
excess. The grower pens were fed twice per week in the dry season and three times per week in the wet season, while the harvest-size crocodiles were fed once and twice, respectively. The following morning, excess food was removed and cleaned with a sodium hypochloride solution before being refilled with clean water. No heating of the water occurs in these pens. Capture and Sampling The sampling period commenced on the June 15, 2012 and finished on October 18, 2013. On sampling days, which were approximately three months apart, a timer was started upon entrance to hatchling and grower pens. At the first sampling, five crocodiles from each pen were randomly captured by hand (<1.2 m) or by the use of an electrical stunner (>1.2 m) and immediately blood sampled from the occipital sinus (see Franklin et al., 2003) using either a 23- or 16-gauge needle depending on the size of the animal. At each successive sampling, these same individuals were resampled. For crocodiles housed in individual pens, a timer was started immediately prior to opening the gate. The cumulative time required to obtain a blood sample (CumTime, in seconds) post pen entrance (grower crocodiles) or post pen opening (harvest size crocodiles) was noted at each sampling (Finger et al., 2015) with a maximum time of 20 minutes in each pen. After collection, blood was placed in lithium heparin tubes, gently inverted three times, and placed on ice. At the first sampling, all crocodiles were sexed by cloacal palpation and measured for head length (HL; in mm) as described by Isberg et al. (2005). Hatchling and grower crocodiles were tagged on the 100 scute (Isberg et al., 2004) with either a small animal tag (size 1005-1; National Band and Tag Company, Kentucky, USA) or a yellow sheep ear tag (Allflex, Queensland, Australia) that was engraved with a unique identifier. This allowed repeated measurements to be made from individual crocodiles to assess the true effect of time of year and animal size measured as head length (HL; mm). To identify these animals, clean water was required which often delayed capture of some crocodiles. Since harvest-size crocodiles were housed individually they did not require tagging. Both grower and individual pen water temperatures were recorded (± 1°C; T TEC-7, Nordex Pty Ltd, South Australia) at the time of sampling. Cloacal temperatures (± 1°C; AI368, Electro Chemical Engineering Pty Ltd, Australia) of harvest-size crocodiles was also recorded at each sampling. Animals were released immediately after sampling. By January 1, 2013 (Sampling 4 onwards), the hatchlings had all been moved into grower pens so were re-categorised as growers. As growers on Farm 2 became large enough, they entered the harvest-size category. The number of crocodiles in each category are listed in Table 1. Corticosterone assays Within two hours, all blood samples were centrifuged at 15,000 rpm for two minutes. Plasma was aliquoted and frozen at -20°C until analysis. Plasma CORT was determined using a high-
sensitivity CORT EIA kit (IDS Ltd., Tyne & Wear, UK) following manufacturer-specific protocols as described previously (see Finger et al., 2015; Isberg and Shilton, 2013). Statistical analyses Restricted maximum likelihood (REML) was used to analyse the CORT levels (ng/ml) for each size category. For all size categories, CORT needed to be log-transformed to meet the assumptions of normality. The following linear mixed model was used: logeCORTijkl = + HLHLi + TCumTimeij + oC Tempij + Farmk + Samplingj + CleanDaysl + AnimalIDi + where is the overall mean; HLi is the head lengths of the ith individual; HL is the regression coefficient associated with HL; CumTimeij is the cumulative time to obtain the blood sample from the ith individual at the jth sampling after entering the pen; T is the regression coefficient associated with CumTime; depending on the model Tempij is either the water (grower or individual pen), air temperature (grower or individual pen; www.bom.gov.au) or cloacal temperature (individual pen only) of the ith individual at the jth sampling; oC is the regression coefficient associated with Temp; Farmk is the fixed effect of the kth farm (k = 1, 2); Samplingj is the fixed effect of the jth sampling (j = 1,…,7); CleanDaysl is the binary effect of number of days since cleaning (l = 0 or 1); AnimalIDi is the random effect of the ith individual and indicates the observations were repeated measures [assumed N(0,
)]; and is the random residual effect [assumed N(0,
)].
To determine the impact of CORT on growth, the following linear mixed model was used: HLijk = + CORTadjCORTadjij + Farmk + Samplingj + AnimalIDi + where all parameters are the same as specified above with the exception of CORTadj ij. Finger et al. (2015, 2016) showed that CORT response can be influenced by human presence with animals sampled later responding with higher CORT levels. To account for this, the variable CORTadjij can be created using the residuals from a linear regression of CORT against time added to the mean CORT value, then back-transforming these, i.e. CORTadjij = exp(resij + mean) (Finger et al., 2015, 2016). A 5% significance level was chosen to evaluate explanatory variables by backward elimination from the full model. The significance of the AnimalID variance component was evaluated using a likelihood ratio test with a 5% significance level. The assumption of equal variances across sampling was tested using a heterogeneous residual variance REML model by a likelihood ratio test. All results are reported as back-transformed means and as standard errors (SE). Least significant difference (5% LSD) tests were used to test differences between fixed effects.
Results
Hatchlings The results of hatchling crocodiles at Farm 1 were presented in Finger et al. (2015, 2016). When combined with those results from Farm 2, the average plasma CORT value for hatchlings was 1.65 ± 0.15 ng/ml (n = 152) across the two farms. CumTime significantly affected CORT levels (P < 0.001; Figure 1), but neither Farm (P = 0.48) nor hatchling size (HL; P = 0.16) affected CORT. Mean CORT at the different sampling periods was on the 5% significance threshold (P = 0.05; Figure 2) and although not significant, mean CORT at sampling 2 is higher than the other sampling periods. Hatchling pen water temperature is environmentally controlled so it was not included in the analysis. Growers A total of 594 samples were taken from 246 individuals with 162 and 84 from Farms 1 and 2, respectively. The average grower CORT level was 2.73 ± 0.21 ng/ml. CumTime significantly affected CORT levels (P < 0.001; Figure 1) but grower size (P = 0.71) had no effect on CORT levels. AnimalID had a significant effect (P < 0.05) on grower CORT. The experimental design employed was to recapture the same individuals across all sampling periods to assess whether individual animals return consistent CORT values. This proved to be detrimental because to re-identify a particular crocodile, clean water is required which meant that animals were mainly sampled on the day of cleaning (81% and 35% of crocodiles sampled on day of cleaning at Farm 1 and Farm 2, respectively). This resulted in a significant interaction (P < 0.001) between Farm and CleanDays, whereby animals sampled on the same day as cleaning on Farms 1 and 2 had mean plasma CORT levels of 8.72 ± 0.60 ng/ml and 2.90 ± 0.54 ng/ml, respectively. Comparatively, if the animals were sampled the day after cleaning from the same respective farm, the CORT levels were 3.70 ± 0.44 ng/ml and 0.60 ± 0.09 ng/ml. For those animals sampled on the day of cleaning, while all attempts were made to access the animals at least two and a half hours after the pens were full of water to allow a return to baseline, often the cleaning process took longer than expected. Consequently, time of day significantly (P < 0.001) affected the resultant CORT levels. To confirm the effect of cleaning, blood samples were taken from randomly-chosen grower crocodiles from different pens on Farm 1 before the clean when the water was still full (n = 22), after the water was drained but before cleaning (n = 21), after the pen was cleaned whilst there was still no water (n = 21), 2½ hours after the pen had been filled (n = 21), and finally the next morning a day after cleaning (n = 12). Figure 3 shows the CORT response corresponding to these cleaning stages. Dropping the water induced a 4-fold increase in plasma CORT with an additional 2.8-fold increase from the cleaning process which also occurs when the water is drained. These levels remained elevated even after 2.5 hours of full water but had declined to pre-cleaning levels by the following morning.
There were no seasonal (Sampling; P = 0.40; Figure 2) effects on grower CORT values although Figure 2 shows unequal variation across the sampling periods that could not be accounted for using the variables evaluated. Pen water temperature also did not significantly affect grower CORT values (P = 0.74) but given that the majority of the animals were sampled on the day of cleaning, the recorded water temperature reflects the temperature of the fresh water used to refill the pen and not the water temperatures experienced by the animals prior to cleaning. Harvest-size crocodiles housed in individual pens Samples were repeatedly taken from 98 individuals from Farm 1 with an additional 44 individuals included from Farm 2. A total of 608 samples were taken throughout the study period with a mean CORT level of 2.19 ± 0.16 ng/ml. CumTime had no effect on CORT levels (P = 0.32; Figure 1). Unlike animals in the communal pens, animals housed in individual pens were able to be sampled more expediently (mean CumTime = 74.4 seconds). The minimum amount of time to take a sample was 31 seconds and the longest time was 801 secs when the electrical stunner failed, although it should be noted that only seven samples took longer than 3 minutes (180 secs) to obtain. Time of day significantly affected CORT levels (P < 0.001), with CORT increasing the later in the day the sample was taken. Neither water (P = 0.42) nor cloacal (P = 0.93) temperature significantly affected CORT. Time of day was not significant in determining the cloacal temperature of harvest-size crocodiles (P = 0.08) but water temperature was significant (P < 0.001; CloacalTemp = 3.09 + 0.90×WaterTemp; Figure 4). There was no effect of length of time housed in an individual pen (P = 0.84) on CORT levels but there was a significant effect of AnimalID (P < 0.05). A heterogeneous residual variance model was required to account for the unequal residual variances between sampling periods. Figure 2 shows there was a significant decline in CORT values over the first four sampling periods (5% LSD) followed by a sudden rise at Sampling 5 (April-May 2013) when many of the Farm 2 growers were transferred into individual pens. CORT returned to previous levels at the 6th sampling (July-August 2013). Impact on growth Table 2 shows the linear regression coefficients of HL in the different size categories. The hatchlings at Farm 2 were larger at the initial sampling compared to those at Farm 1 and grew at a slightly faster rate (HL: Farm 1 = 0.40 cm/month; Farm 2 = 0.48 cm/month). Growers on the two farms are growing at the same rate (0.33 cm/month). Within the individual pens, Farm 1 harvest-size crocodiles were larger at the initial sampling but Farm 2 crocodiles were growing at 1.87 times the rate.
CumTime-adjusted CORT values were used as an explanatory variate for growth (i.e. HL) but in each size category had no effect (hatchlings P = 0.09; growers P = 0.62; harvest-size P = 0.73). On the other hand, farm and sampling periods were highly significant (P < 0.001) for all size categories including their interaction in both the grower and harvest-size crocodile analyses.
Discussion In this study, we repeatedly sampled the same juvenile saltwater crocodiles on two Northern Territory farms to understand the effects of seasonal and inter-farm variations on plasma CORT across size categories. This repeated sampling allowed us to examine if individual crocodiles displayed consistently high, low, or stochastic CORT values. Finger et al. (2015) reported stochastic CORT values in hatchling crocodiles that could not be attributed to AnimalID. The addition of hatchlings from another farm did not alter this outcome, suggesting that hatchling CORT is controlled by an, as of yet, elusive influence. In contrast, there was significant variation between crocodiles in both the grower and harvest size categories. This could represent better habituation of some animals compared to others in the captive situation. In the case of the grower crocodiles though, since the majority of observations were made on the day of cleaning, the AnimalID effect might actually represent an individual’s ability to tolerate the cleaning regime. For the harvest-size crocodiles housed in individual pens, the AnimalID effects might be more representative of long-term habituation which may have compromises on economic indicators such as blemish healing as discussed within Isberg and Shilton (2013). Analogous to the observation in hatchling crocodiles observed in this study and in Finger et al. (2015), CumTime had a significant effect on grower crocodile CORT even though all crocodiles were sampled within 3 minutes of capture. This suggests that the presence of people within a grower pen and/or the capture of conspecifics increases CORT levels (Finger et al. 2015). Because we repeatedly sampled the same individuals, the CumTime required to find these same individuals at successive samplings was higher than the first sampling. In fact, to rapidly identify individuals and prevent an even higher impact of CumTime on CORT, we sampled individuals on the same day as pen cleaning. Therefore, pen cleaning is an obvious confounder. Although CORT values from Farm 1 were higher than those from Farm 2 regardless of cleaning, this may have been influenced by the greater proportion of crocodiles on Farm 1 (81%) being sampled on cleaning days than on Farm 2 (35%). For those grower crocodiles sampled on the day of cleaning, all attempts were made to access the animals more than 2.5 hours after the pens were full of water, thereby allowing crocodiles time to settle after cleaning disturbances. But, as shown in Figure 3, the resultant elevated CORT levels due to cleaning did not return to baseline levels until the morning following cleaning. In addition, the later in the day the sample was taken, the longer the cleaning process had taken and the longer crocodiles
had been without full water. Consequently, time of day significantly affected plasma CORT values of grower crocodiles. In contrast to grower crocodiles, we were able to rapidly sample crocodiles housed in individual pens (see Franklin et al., 2003; Isberg and Shilton, 2013). As such, capture time had no effect on harvest-size crocodile CORT levels. Similar to grower crocodiles, however, CORT increased as the time of day progressed. This is in contrast to what has been observed in juvenile alligators. Lance and Lauren (1984) observed a biphasic peak in alligator CORT at 0800 hours and again at 2000 hours. Over this study period, the earliest and latest samples were taken at 0719 and 1628 hours, respectively, and we observed an increasing linear trend in CORT over the course of the day. However, this increase may not have been necessarily due to abiotic influences (i.e., photoperiod, temperature, etc.) alone. In fact, increased CORT with increased time of day likely reflects the higher probability of human disturbance, as husbandry-related activities and human-crocodile interactions increased with increasing time of day as well. In contrast to grower crocodiles, since individual pens were not cleaned on sampling days, it is unlikely that cleaning affected CORT levels. It may be the case that various size/age classes of crocodilians display circadian differences in CORT release, but this requires future investigation. There were seasonal (Sampling) effects on harvest-size crocodile CORT, but not on grower or, contrary to Finger et al. (2015), hatchling CORT. Sampling was used to model the combined effect of abiotic environmental factors (i.e., ambient air temperature, photoperiod, etc.) of a particular time of year on CORT (Finger et al., 2015). However, using “Sampling” as a collective method for extracting environmental effects may not be optimal. Indeed, even though water temperature was kept constant (i.e., provided with heaters) for hatchling crocodiles, Sampling significantly affected CORT values as reported by Finger et al. (2015), although this was not substantiated herein when combined with Farm 2’s hatchling CORT data. Furthermore, despite a strong positive relationship between water and cloacal temperature (Figure 4), no effect of water temperature on grower and harvest-size crocodile CORT was observed. Whilst other environmental factors (i.e., photoperiod or air temperature) may have contributed to this Sampling effect, husbandry practices may have confounded this as well. Similar to previous results (Isberg and Shilton, 2013), the time a crocodile spent in an individual pen did not affect CORT levels. However, using Sampling as the predictor variable, there was a significant decline in CORT values over the first four sampling periods which might indicate either habituation to individual housing or a seasonal decline in CORT. The increase in CORT at Sampling 5 is explained by grower animals at Farm 2 being moved into individual pens, as CORT levels at Sampling 6 are similar to those at Sampling 4 (Figure 2).
Heterogeneous variance models were required to meet the assumptions of unequal variances across sampling periods for harvest-size crocodiles, and although not significant, there was still considerable variability in grower crocodiles as well (Figure 2). Large variation in crocodilian plasma CORT is not unprecedented in the literature. In fact, in our previous study we also had to account for variance heterogeneity (Finger et al., 2015). In a study investigating how CORT implants affect hatchling alligator health, CORT values for the control alligators ranged from 3.8-42.8 ng/ml (Morici et al., 1997). In Nile crocodiles, CORT values ranged from 5.15-118 ng/ml at initial capture (Pfitzer et al., 2014). Whilst the statistical models used herein were able to account for much of the variability, particularly caused by CumTime, there were also many factors that could not be included in a model, such as agonistic interactions or dietary status of an individual. More investigation is required to better understand the factors contributing to this variability. More so, although individual tolerances for habituation have been useful in modelling these data, the utility of non-invasive measures such as faecal CORT could prove more useful (Ganswindt et al., 2014). Ultimately, one of the best indicators of poor management or chronic stress is its impact on economic predictors such as reduced growth rate (Elsey et al., 1990; Morici et al., 1997). Since high CORT values have been associated with reduced growth in many crocodilian species (e.g., Elsey et al., 1990) including saltwater crocodiles (Turton et al., 1997; Shilton et al., 2014), we examined the effect of time-adjusted CORT values on grower and harvest-size crocodile HL (Table 2). Similar to our previous hatchling study (Finger et al., 2015), we observed no effect of CumTime-adjusted CORT values on both growers and harvest-size crocodile growth suggesting that CORT levels are not suppressing growth. Furthermore, housing hatchlings in either complete darkness (Farm 2) or in sheds with reflected ambient light (Farm 1) had no effect on growth or plasma CORT.
Conclusions This is the first study to repeatedly sample juvenile farmed saltwater crocodiles from all size categories to assess CORT as a proxy of welfare. Since CORT values were generally low, along with no impact observed on growth, these results suggest that the farming environment and husbandry practices, as implemented under the Australian industry Code of Practice on the humane treatment of wild and farmed Australian crocodiles (NRMMC, 2009), are acceptable. Our results should provide a foundation for further welfare research, including the exploration of less invasive methods to assess the impact of husbandry.
Acknowledgements: We would like to extend a special thanks to Darwin and Lagoon Crocodile Farms for providing access to crocodiles used in this study. We would also like to thank Sue Aumann for help with ELISA kits
and to Berrimah Veterinary Laboratories for providing lab space for performing ELISAs. This project was partially funded by a grant from the Rural Industries Research and Development Corporation, Australia and protocols were approved by the University of Sydney Animal Ethics Committee (approval number: N00/5-2012/3/5729). JWF was funded by the Interdisciplinary Toxicology Program at the University of Georgia and the former Faculty of Veterinary Science, University of Sydney.
Conflicts of interest: The authors declare no conflicts of interest. References Elsey, R.M., Joanen, T., McNease, L., Lance, V., 1990. Growth rate and plasma corticosterone levels in juvenile alligators maintained at different stocking densities. J. Exp. Zool. 255, 30–36. Finger, J. W. Jr., Adams, A. L., Thomson, P. C., Shilton, C. M., Brown, G. P., Moran, C., Miles, L. G., Glenn, T. C., Isberg, S. R. 2013. Using phytohaemagglutinin to determine immune responsiveness in saltwater crocodiles (Crocodylus porosus). Aust. J. Zool. 61, 301-311. Finger, J.W. Jr., Thomson, P.C., Adams, A.L., Benedict, S. Moran, C., Isberg, S.R., 2015. Reference levels for corticosterone and immune function in farmed saltwater crocodiles (Crocodylus porosus) hatchlings using current Code of Practice guidelines. Gen. Comp. Endocrinol. 212, 63-72. Finger, J.W. Jr., Thomson, P.C., Isberg, S.R., 2016. Unexpected lower testosterone in faster growing farmed saltwater crocodile (Crocodylus porosus) hatchlings. Gen. Comp. Endocrinol. 226, 1-4. doi: 10.1016/j.ygcen.2015.11.016. Franklin, C.E., Davis, B.M., Peucker, S.K.J., Stephenson, H., Mayer, R., Whittier, J., Lever, J., Grigg G.C., 2003. Comparison of stress induced by manual restraint and immobilisation in the estuarine crocodile, Crocodylus porosus. J. Exp. Zool. 298, 86-92. Ganswindt, S.B., Myburgh, J.G., Cameron, E.Z., Ganswindt, A., 2014. Non-invasive assessment of adrenocortical function in captive Nile crocodiles (Crocodylus niloticus). Comp. Biochem. Physiol. 177A, 11-17. Isberg S.R., Shilton, C.M., 2013. Stress in farmed saltwater crocodiles (Crocodylus porosus): no difference between individually- and communally-housed animals. SpringerPlus. 2, 381. doi:10.1186/2193-1801-2-381. Isberg, S.R., Thomson, P.C., Nicholas, F.W., Barker, S.C., Moran, C., 2004. Farmed saltwater crocodiles-A genetic improvement program. Rural Industries Research and Development Corporation Paper No. 04/147.
Lance, V.A., Elsey, R.M., Trosclair, P.L., 2015. Sexual maturity in male American alligators in southwest Louisiana. South Am. J. Herpetol. 10, 58–63. Lance, V.A., Lauren, D., 1984. Circadian variation in plasma corticosterone in the American alligator, Alligator mississippiensis and the effects of ACTH injections. Gen. Comp. Endocrinol. 54, 1-7. Morici, L.A., Elsey, R.M., Lance, V.A., 1997. Effects of long-term corticosterone implants on growth and immune function in juvenile alligators, Alligator mississippiensis. J. Exp. Zool. 279, 156-162. Morpurgo, B., Gvaryahu, G., Robinzon, B., 1992. Effects of population density, size, and gender on plasma testosterone, thyroxine, haematocrit, and calcium in juvenile Nile crocodiles (Crocodylus niloticus) in captivity. Copeia 4, 1023–1027. NRMMC (National Resource Management Ministerial Council), 2009. Code of Practice on the humane treatment of wild and farmed Australian crocodiles. NRMMC, Canberra. Pfitzer, S., Ganswindt, A., Fosgate, G.T., Botha, P.J., Myburgh, J.G., 2014. Capture of farmed Nile crocodiles (Crocodylus niloticus): comparison of physiological parameters after manual capture and after capture with electrical stunning. Vet. Rec. doi:10.1136/vr.102438 Rooney, A.A., Crain, D.A., Woodward, A.R., Guillette, L.J., 2004. Seasonal variation in plasma sex steroid concentrations in juvenile American alligators. Gen. Comp. Endocrinol. 135, 25–34. Shilton, C., Brown, G.P., Chambers, L., Benedict, S., Davis, S., Aumann, S., Isberg, S.R., 2014. Pathology of runting in farmed saltwater crocodiles (Crocodylus porosus) in Australia. Vet. Pathol. doi: 10.1177/0300985813516642 Turton, J.A., Ladds, P.W., Manolis, S.C., Webb, G.J., 1997. Relationship of blood corticosterone, immunoglobulin and haematological values in young crocodiles (Crocodylus porosus) to water temperature, clutch of origin and body weight. Aust. Vet. J. 75, 114-119.
Figure Legends Figure 1. The effect of cumulative time (CumTime; secs) to procure a blood sample for CORT analysis. Although all blood samples were taken within three minutes of capturing an individual, CumTime represents the cumulative time before the sample was taken after entering the pen. Black circles are hatchlings, open circles are growers and grey circles are harvest-size crocodiles in individual pens.
Figure 2. Box-plots for model-adjusted plasma corticosterone (CORTadj; ng/ml) showing medians, first and third quartiles, minimum and maximum values between sampling periods. A. Hatchlings; B. Growers; C. Harvest-size crocodiles housed in individual pens. * and # indicates a significant difference (P <0.05) from the previous sampling period.
Figure 3. CORT (ng/ml) response of grower crocodiles at various stages of the cleaning process. * indicates a significant difference (P <0.05) from baseline levels before cleaning.
Figure 4. Relationship between crocodile cloacal and water temperatures in harvest-size crocodiles in individual pens (r2 = 0.86; P <0.001).
Table 1. Summary of the number of hatchlings (<1 year old), growers (1-3 years old) and individual pen (2+ years old) crocodiles sampled at each sampling period (1-7) from Farm 1 and Farm 2. In between Samplings 3 and 4, the animals changed age-category whereby hatchlings became growers. Some grower animals from Farm 2 entered the individual pens.
Sampling 1 Months Jun-Jul Year Hatchlings Farm 1 40 Farm 2 Growers Farm 1 Farm 2 Individual pens Farm 1 98 Farm 2 -
2 Aug - Sep 2012
3 Nov - Dec
4 Feb
5 6 Apr - May Jul - Aug 2013
7 Oct
39 20
38 15
-
-
-
-
90 38
73 39
97 29
79 18
66 23
32 -
95 20
88 19
92 25
74 21
52 11
-
Table 2. Linear regression equations for predicting HL over sampling periods at Farm 1 and 2. Regression coefficients are given as 0 + xx where x is sampling. r2 values are also presented
Hatchling
Growers
Individual pens
Farm 1
5.69 + 1.20x; r2 = 0.99
11.44 + 1.02x; r2 = 0.99
23.54 + 0.38x; r2 = 0.99
Farm 2
10.27 + 1.44x; r2 = 1.00
11.76 + 1.02x; r2 = 0.96
22.46 + 0.71x; r2 = 0.99
Highlights of Paper:
We defined corticosterone in different size categories of farmed saltwater crocodiles The average plasma corticosterone (CORT) values for hatchlings and growers in communal pens were 1.65 ± 0.15 ng/ml and 2.73 ± 0.21 ng/ml, respectively, and for harvest-size crocodiles in individual pens was 2.19 ± 0.16 ng/ml No seasonal effects on CORT were observed Farmed saltwater crocodiles are not inherently stressed Validation of faecal CORT should be pursued in saltwater crocodiles as a noninvasive method for continually assessing the impact of husbandry regimes