01_front.pdf - Massey Research Online

(1)

Copyright is owned by the Author of the thesis. Permission is given for

a copy to be downloaded by an individual for the purpose of research and

private study only. The thesis may not be reproduced elsewhere without

the permission of the Author.

(2)

THE USE OF DOGS TO DETECT NEW ZEALAND REPTILE SCENTS

A thesis presented in partial fulfihnent of the requiretnents for the degree of Master of Science in Zoology at Massey University,

Pahnerston North, New Zealand.

Clare Melody Browne

2005

(3)

ABSTRACT

This study examined the ability o f do mest ic dogs ( Canis familiaris) to detect the scent o f t h e Cook Strait tuatara (Sphenodon puncta/us), Marlborough green gecko (Naultinus manukanus) and forest gecko (Hoplodactylus granulatus).

H andlers from two local dog training clubs with a total of 20 dogs participated in this study. The dogs ' capacity to detect human and repti les scents was evaluated in a series of tria ls. Each trial required the dogs to identi fy a d i fferent target scent, and consist ed of nine rep l icate scent discriminat ion exerc ises. In the exercises the dogs were presented w ith a l ine of c loths. One or more of the c loths contained scent and the dogs were commanded to locate a spec ific scented c loth. Tuatara and gecko seats, sloughed skins and paper towels captive individuals had been s itting on were used to imbue the c loths with reptile scent .

The dogs were able to ident ify human, tuatara and gecko scents with average success rates o f up to 96. 3%, 93 . 7% and 86. 7%, respective ly. The dogs could detect fresh reptile seats, seats that had been exposed in nat ive forest for two weeks and discriminate between several di ffcrcnt rept ile scents. The detection successes were signi ficantly h igher than would be expected if the dogs were select ing c loths at random (p ⁼ 0.05). The average results of each trial and the success rates of ind ividual dogs were s ignificantly d i fferent at both dog clubs (p ⁼0.000).

The results indicate that the methods used in this study arc a good model for scent discrimination research, and dogs cou ld be used to detect tuatara and gecko species for conservation work. Dogs may provide an alternative to the visual methods currently used to locate these reptiles.

11

(4)

ACKNOWLEDGEMENTS

Thank- you to my supervisors : Pro fessor Kevin Stafford and Associate Professor Robin Fordham. Your support ( Robin) and unrelenting harassment ( Kevin) helped me complete this thesis in the end.

I would like to thank a l l o f the hand lers and their dogs t hat participated in this research; I could not have done it w itho ut your help. Thanks to everyone at the Tararua A llbrceds Dog Training Club: Karen Belcher with Abbie; B ronwen Burnette with Topaz; Sue Cooper with Celtic; C heric Gear with Bonn; Robyn How with Po lly; Meredith P itcher w ith Rowan; Sandra M ohekey with Runa; and J anet Young w ith Kayla and Tasha.

Thank- you also to everyone from the Fei lding Dog Training Club: Robyn Annand with Chelsea and Willy; L inda Avery w ith Cassie; Me ! B lair with E isa and Spice; Amber Gorinski with Titan and Zeus; L inda Gunness wit h Sambuca; D iane Lock with Kayla;

Maggy Lud low with Zane; E i leen McQuil lan with Azzie; and Janet Young w ith Kayla and Tasha (again).

My vo lunteers helped me in all weather cond itions, mostly without complaint. Thank-you to I an Johnston; Let it ia Morrison; Rowena Teal; K iryn Weaver; Mark Hamer; Caro l N icho l son; Nadine F letcher; Kymberly Eagleson; Stacey Horton; Bridget Wrenn;

Monique Jansen van Rensburg; Jeraldine Teng; Fiona Sanggang; Andrea 0' eill; Brad S iebert ; and Claire Russell. Again, this project could not have been done without your help.

Thanks to everyone who co llected reptile samples for me : Bruce Benseman, Rhys M i l l s and the other staff a t N ga Manu N ature Reserve; Kelly Hare and Sue Keal l at V ictoria University of Wel l ington; Roger and Barbara Watkins; H eather Barton ; B arbara B lanchard at Wellington Zoo; and Dominique Fortis at Otorohanga Kiwi House and Native B ird Park. This research was dependent on getting sufficient reptiles samp les, and you were all very generous with the t ime and e ffo rt you put into collecting them for me.

Thanks a lso to all the people at Massey U nivers ity who carried c loths around for me.

Ill

(5)

I had usefu l discussions about my methods, analysis, tuatara, p lants and dog obedience w ith Doug Armstrong; Sue Cooper; L indsay H azley; E d M inot; A lasdair Noble; and J i l l Rapson. A lasdair Robertson gave m e t ips o n how t o win favour with m y thesis markers.

I would l ike to thank the other conservation dog handlers and assoc iated people I have spoken to over the past few years: Graeme Atk ins; I sabe l Castro; John Cheyne; Dave Crouchley; Lance Dew ; John McLennan; Jonathan M i les; Keri N ei lson ; Murray Potter;

S teve Sawyer; Mandy Tocher; and Adele S maill. Thank-you to Linda Kerley for giving me good advice on dog training; and Deborah Smith and Al ice Whitelaw for sharing i n formation on the ir own research with conservat ion dogs. Thanks also to l an Mc Lean for g iving me t ips on training a detection-dog.

Thank-you to Yvonne Gray fro m AgResearch for supplying weather informat ion ; Brian Carter at the Carter Observatory in We l lington for sunlight hours; and Barbara J ust for sorting things out for me. Thanks to Or Paddy Ryan for a l lowing me to use his fantastic photographs, and Louise J une for braving the dogs and tak ing photographs for me.

Rowena Teal tried to take some photographs too.

Thanks Ian for your support throughout the ent ire process, for standing in the ram watching dogs and reading hundreds of drafts of this thesis. Thanks to Fleur Maseyk and Dorothce Durpoix for reading dra fts for me as well. Thanks to my parents for the food and encouragement. My friends in the Wildlife Eco logy lab made me realise that my fie ldwork could have been so much worse. Thanks to my dog for giving me my flrst grey hair.

F inancial assistance was very gratefully received from the Julie A l ley Bursary; the I nst itute of Veterinary, Animal and Bio medical Sc iences and the Eco logy Group, I nstitute of Natural Resources, Massey U niversity. Thanks also to Lesley Dawson from H i l ls Pet N utrition N Z Ltd. for providing a mountain o f dog food. The Massey U niversity A nimal Ethics Committee gave approval for the experiments described in this thesis.

IV

(6)

LIST OF FIGURES

Figure 3.1: Tararua Allbrccds Dog Training C lub grounds. 32

Figure 3.2: E xamp les of the tuatara samples used to scent target and decoy 34 c loths. A tuatara scat is on the upper left, a p iece of tuatara skin is on the lower

left, and a tuatara-scented paper towel ( fo lded) is on the right.

Figure 3.3: Dog 2 being encouraged to smell one of a pair of target c loths. The 3 7 second target cloth has been placed i n the l ine.

Figure 3.4: Dog 5 sn iffing along the line of c loths, searching for the target c loth. 3 7

Figure 3.5: Dog 3 retrieving t he target cloth. 3 8

Figure 3.6: Dog 5 present ing the target c loth to its handler. 3 8

Figure 3 . 7: The tuatara seats were placed i n this young planted nat ive forest. 42

Figure 3.8: Tuatara seats sitting in the forest ( in the pouch on the right ), with 42 rain gauge and high/low thermo meter.

Figure 3.9: The temperature range, amount of rainfa l l and sunlight hours the 5 5

tuatara seats were exposed to i n young planted native forest between 20 September and 1 8 October 2 003.

Figure 3.10: Causes of failed scent exerc ises for each dog, across all scent trials. 5 5 n ⁼total number o f fai led exerc ises.

Figure 4.1: Feilding Dog Training C lub grounds. 7 7

X

(12)

Figure 4.2: E xamples o f t he gecko samples u sed to scent target and decoy 7 7 c loths. A Marlborough green gecko-scented paper towel ( fo lded) is on the upper

left, half of a M arlborough green g ecko skin is a long the bottom, a M arlborough green gecko scat is on the upper r ight, and a forest gecko scat is on the lower right.

Figure 4.3: Dog 1 2 and its handler facing away from the line o f c loths while one 78 o f a pair of target c loths is p laced in the l ine.

Figure 4.4: Dog 4 sniffing alo ng the line of c loths, searching for the target c loth. 78

Figure 4.5: Dog 1 5 retrieving the target cloth. 78

Figure 4.6: Gecko seats sitting in t he forest (in the two pouches on the left), with 82

ra in gauge and high/low thermo meter. One pouch is covered by a p lastic container to protect it from direct rain fall.

Figure 4.7: The temperature range, amount of rainfall and sunlight hours the 97

M arlborough green gecko seats were exposed to in the fo rest between 2 1 September and 1 9 October 2003 . (The covered seats did not receive any direct ra infall. )

Figure 4.8: Causes of failed scent exerc ises for each dog, across all scent trials. 97 n ⁼total number of failed exerc ises.

XI

(13)

LIST OF TABLES

Table 2 . 1: Protected native species and introduced pest spec ies t hat conservation 1 7 dogs (protected spec ies dogs and predator dogs) have been trained to locate.

Table 3 . 1: The dogs that participated in this study fro m the Tararua Al lbrecds 3 1

Dog Training C lub. The dogs received various levels of obedience training, inc luding scent discrimination training, prior to this study.

Table 3.2: The success of all dogs in the seven different scent trials at the 45

Tararua Allbrceds Dog Training Club. The dogs were required to ident ify a different target scent in each tria l. T he results arc calcu lated as the average percent correct .

Table 3 .3: Trial I results. The target cloths were scented with the dogs' own 47 handler' s scent. I indicates a success ful scent exercise, 0 indicates a failed scent

exerc1se.

Table 3 . 4: Trial 2 results. The target cloths were scented with un familiar 48

people ' s scent. l indicates a success fu l scent exercise, 0 indicates a failed scent exerc1se.

Table 3 .5: Trial 3 results. The target cloths were scented with t uatara-sccntcd 49

paper towels. I indicates a success fu l scent exercise, 0 indicates a failed scent exerc1sc .

Table 3 .6: Trial 4 results. The target cloths were scented with tuatara seats. 50 l indicates a successfu l scent exercise, 0 indicates a failed scent exercise.

Table 3. 7: Trial 5 results. The target cloths were scented with tuatara skins. 5 1 l indicates a succes s fu l scent exercise, 0 indicates a failed scent exercise.

Xll

(14)

Table 3.8: Trial 6 results. The target c loths were scented with e ither tuatara seats 5 2

or skins. I indicates a successful scent exercise, 0 indicates a fai led scent exerc ise, D indicates a fai led scent exercise when a decoy cloth was retrieved instead of the target cloth.

Table 3.9: Trial 6 results, showing the two groups the dogs were spl it into. The 53 target cloths were scented w ith eit her tuatara seats or sk ins. The results are

calculated as the average percent correct.

Table 3.10: Trial 7 results. The t arget cloths were scented with weathered tuatara 54 seats. I indicates a success fu l scent exercise, 0 indicates a fai led scent exerc ise.

Table 3.11: Comparison of this study with other assessments of the ability o f 6 1

dogs to detect bio logical scents . All studies tested the dogs in experimenta l s ituations, using similar methods.

Table 4.1: The dogs that part ic ipated in this study fro m the Fei lding Dog 75

Training C lub. The dogs received various levels of obedience and agil ity training, including scent discriminat ion training, prior to this study.

Table 4.2: The success of all dogs in the seven di fferent scent trials at the 84 Fe i lding Dog Training C lub. The dogs were required to identify a di fferent target

scent in each trial. The resu lts arc calcu lated as the average percent correct.

Table 4.3: Trial I results. The target cloths were scented w ith the dogs' own 86

handler's scent. l indicates a success fu l scent exercise, 0 indicates a fai led scent exerctse.

Table 4.4: Tria l 2 results. T he target c loths were scented with unfamiliar 87

people 's scent . l indicates a successfu l scent exercise, 0 indicates a fai led scent exerctse.

X Ill

(15)

Table 4.5: Trial 3 results. The target cloths were scented with Marlborough 8 8

green gecko-scented paper towels, seats o r skins. I indicates a successful scent exerc ise, 0 indicates a fai led scent exerc ise.

Table 4.6: Trial 3 results, showing the three groups the dogs were split into. The 8 8 target cloths were scented with M arlborough green gecko seats, s kins, or paper

towels. T he resu lts arc calculated as the average percent correct .

Table 4. 7: Trial 4 results. The target cloths were scented with either 89 M arlborough green gecko seats or s k i ns. I indicates a success fu l scent exercise,

0 indicates a fa i led scent exerc ise.

Table 4.8: Trial 4 results, showing the two groups the dogs were split into. The 90 target c loths were scented with e ither Marlborough green gecko seats or skins.

The results arc calcu lated as the average percent correct.

Table 4.9: Tria l 5 resu lts. The target cloths were scented with eit her 9 1

M arlborough green gecko seats or sk ins. I indicates a success fu l scent exerc ise, 0 indicates a failed scent exerc ise, D indicates a failed scent exercise when a decoy c loth was retrieved instead o f the target cloth.

Table 4.10: Trial 5 results, showing t he two groups the dogs were split into. The 92 target c loths were scented with either Marlboro ugh green gecko seats or skins.

The results arc ea lcu lated as the average percent correct.

Table 4.11: Trial 6 resu lts. T he target cloths were scented with either 93 M arlbo rough green gecko seats or forest gecko seats. I indicates a successful

scent e xercise, 0 indicates a failed scent exercise, D indicates a fai led scent exercise when a decoy c loth was retrieved instead of the target c loth.

Table 4.12: Tria l 6 results, showing the two groups the dogs were spl it into. The 94 target c loths were scented with e it her Marlborough green gecko seats or forest

gecko seats. The results arc calculated as the average percent correct.

XIV

(16)

Table 4.13 : Trial 7 results. The target c loths were scented w it h weathered 95

M arlborough green gecko seats. I indicates a successfu l scent exercise, 0 indicates a fai led scent exerc ise, D indicates a fai led scent exerc ise w hen a decoy c loth was retrieved instead of t he target c loth.

Table 4.14: Trial 7 results, showing the two groups the dogs were spl it into. The 96 target c loths were scented with weathered M arlborough green gecko seats that

had been e ither completely exposed or covered. The results arc ca lculated as the average percent correct.

XV