Echinococcosis, which includes cystic and alveolar echinococcosis (CE and AE respectively) is a major parasitic zoonosis of public health importance in western China, where 380 000 echinococcosis patients are estimated to live.1-3 In western Sichuan Province, human alveolar echinococcosis (AE) prevalence can reach 9.4% (56/593) in some Tibetan communities.4 Human AE is particularly important due to its high mortality in untreated patients (over 94% with 10 years after diagnosis), due to the invasion of the liver by E. multilocularis larval lesions and extension or metastases to other organs.5,6 Transmission to humans occurs when eggs of the tapeworm, excreted by the final hosts (usually foxes, but also dogs), are ingested accidentally. The larva primary target organ is the liver where it proliferates slowly, but also spreads into extrahepatic structures and even metastasizes to distant organs.3 Small mammal intermediate hosts can also be infected. Thus the parasite develops into a metacestod in the liver and produces large numbers of protoscolices.7 The cycle is completed when a final host ingests an infected small mammal and protoscolices develop into adult worms in the intestine.
Budke et al8 have found the highest prevalence (13%–33%) of E. multilocularis in domestic dogs on record in Tibetan communities of western Sichuan Province, China. Based on the analysis of the results of mass-screening of human populations, Wang et al9-11 supported the hypothesis that overgrazing is a key factor for E. multilocularis transmission to dogs and humans. One possible explanation was that fencing was modifying grazing patterns, making it more conducive to overgrazing in unprotected areas and higher grass within the fenced areas. It was thus hypothesized that grass height would have consequences on the population density of small mammal intermediate host such as Ochotona spp. and Microtus spp., which could then consequently increase the transmission pressure from dogs to humans. Craig et al12 and Wang et al13 found that owned dogs are a major transmission source to humans in Gansu and on the eastern Tibetan Plateau, respectively. On the plateau, the black lipped pika (Ochotona curzoniae) was thought as one of the possible major intermediate host in the parasite life cycle.14 Indeed, Qiu et al15 showed that prevalence in plateau pika could reach 6.7% (5/75). However, this picture is now complicated by the recent discovery of a new species of Echinococcus, i.e. E. shiqucus,16,17 that is known to circulate between the Tibetan fox Vulpes ferrilata and the plateau pika, and could have been mistaken with E. multilocularis in the past, resulting in a lower E. multilocularis prevalence than previously believed.
The current research was conducted to better understand the link between grass height and the distribution of small mammal species (Ochotona spp. and Microtus spp.) and to compare this distribution to dog infection and compare dog and fox diet in the area.
The study was carried out in 6 townships in Shiqu County of Ganzi Tibetan Autonomous Prefecture in northwest Sichuan Province, China, from 2001 to 2007. The selection of the investigation sites was based on documented high prevalence of AE in humans by mass ultrasound screening, the existence of overgrazing impacts and reported outbreaks of small mammals.4,9-11,15
Shiqu County is located on the eastern Tibetan Plateau at 97°20'00"–99°15'28"E and 32°19'28"–34°20'40"N, has a population of 71 000 (98% ethnically Tibetan), and borders Qinghai Province in the east, north and west and the Tibet Autonomous Region in the south. It covers an area of 25 100 km2, with a mean elevation of 4200 meters, 19 000 km2 of which is classed as grazing pasture. The weather is affected by a monsoon climate, with a mean annual rainfall of 596 mm, and is characterized by wide temperature differences (average temperature –1.6°C). Winter is longer than summer and frost conditions persist throughout the year.18
The study was approved by the Ethical Committee of Sichuan Center for Disease Control and Prevention as well as those of all collaborating investigators.
Small mammal density investigations were performed in June 2007, in 39 settlements, located in 24 villages of 4 Tibetan pastoral townships. The 4 townships were Mengsha, Yiniu, Xiazha and Eduoma which cover 1 434 km2, 955 km2, 834 km2 and 1503 km2 respectively. Small mammal populations were monitored using index methods. These methods are based on the detection of surface indices of small mammals (i.e., holes and feces),19-22 and are usually employed to link small mammal indices and habitats on a large scale.14,23-25 Ochotona spp. burrows (>10 cm in diameter, always with round feces with 0.5 cm in diameter) were differentiated from Microtus spp. like burrows (<5 cm in diameter, 0.2 cm sized flat-ellipse feces) by their sizes and feces. Sampling was performed by 2 investigators walking at approximately 1 kilometer per 1.5 hours along a 2 kilometers transect drawn across each settlement. Along this transect, 50 counts of small mammal burrows were performed; each count covered an area 40 meters long and 10 meters wide (about 400 m2). For valley, transects began at a point 20 meters to the right of the first house in the direction of entrance into the settlement. If it was impossible to walk in this direction (for example, facing a very steep hill), the starting point changed to the left of the first house. For settlements on flatland, transects went north beginning from a point 20 meters away from the northernmost house of the settlement.11 For the first count of each transect, grass height was measured using a ruler every 2 meters. Along each transect, stray dogs were counted within a 200 meters distance. The land pattern for the settlements is defined as flatland or valley accordingly.
To obtain infection data in owned dogs, in settlements where transects were done in 2 townships (Yiniu and Xiazha), all dogs were sampled through arecoline purgation (May, September 2002), according to the recommendations of OIE/WHO,7 and/or through collecting ground feces when accessible and available.8 Purgation and feces samples were taken to the Sichuan Center for Disease Control and Prevention in Chengdu, where helminths were removed, counted, and placed in 10% formal saline or 95% ethanol. Species identification of worms was done at the Sichuan Center for Disease Control and Prevention. Copro-PCR testing was performed at the School of Environmental and Life Sciences, University of Salford (UK), using species-specific primers for E. multilocularis DNA amplification based on the method of Dinkel et al26 as modified by van der Giessen et al.27 Owned dogs were always tied and released at night, while dogs were considered to be “stray dogs” if they were observed to be free–roaming during day-time.9
For dietary analysis, dog feces were sampled in 3 townships, i.e. Tuanjie (May and July 2002), in Qiwu and Yiniu (May 2006) in and around villages, and fox feces were collected in the vicinity of Tuanjie in July 2001 and 2002. For dietary analysis, feces were processed according to Zabala and Zuberogoitia,28 with a preliminary step of decontamination using an autoclave (30 minutes at 120°C 1.1 bar-in humid atmosphere). The non-digested remains (bones, skulls, teeth and insects) were identified using reference collections. Complementarily, small mammal hairs found in dog feces collected in Tuanjie were cross-sectioned and prepared on microscope slides. Briefly, hairs were fixed in 3% PBS glutaraldehyde, dehydrated using 4 successive ethanol baths, then impregnated using Technovite®7100 (Hereus Kulzer, Wehrheim, Germany).
Dietary analysis results were expressed as frequency of occurrence of components (tooth, hair and bone). Assuming the longer fencing the higher grass quality, the link between grass height (independent variable, considered as a proxy of grazing pressure) and the burrow density of small mammals and the year when the pasture was fenced (dependent variables), was modeled using a generalized linear model (negative binomial distribution, considering the overdispersion of data). Burrow density of small mammal (Ochotona spp. and Microtus spp. respectively) were modeled against years of pasture fencing and land characteristics (valley or flatland) using a generalized linear model (negative binomial distribution).
Arcview GIS 3.2 (ESRI Redlands, CA) was used to link the transect data with nearby owned dogs within a 4 kilometers diameter circle (Figure). Infection in owned dogs (dependent variable: not infected coded as 0 and infected as 1) was modeled against the maximum and median burrow density of Ochotona spp. and Microtus spp. using Logistic regression. Abundance of worms (i.e. the number of E. multilocularis worms or worm burden) collected at purgation from owned dogs against dog age and sex, and burrow density of Microtus spp. and Ochotona spp using generalized linear model with a negative binomial distribution and a stepwise approach.
|view in a new window |
Figure. Illustration of geographic positions of infected dogs and uninfected dogs produced by Arcview 3.2.
Grass height and burrow density of small mammals
The average grass height was 4.33 cm (SD=0.830, n=31) and the median burrow density of small mammals was 0 (maximum=48) per 400 square meters. The lower the grass height, the higher burrow density was found in the months of June and July, when average grass height was equal to or higher than 3 cm (χ2=4.670, P=0.031, coefficient= –1.570, 95% confidence interval was from –3.032 to –0.107, n=27) (Table 1).
|view in a new window |
Table 1. Relationship between grass heights and burrow density of small mammals
For 1809 valid counts, the ratio of burrows for Ochotona spp. vs. Microtus spp. was 8.2 (34 698/4223). The ratio of burrows for Microtus spp. in open pastures vs. fenced pastures was 8.6 (3785/438). Regarding Microtus spp., burrow density greater than 126 burrows per 400 square meters were observed in open pastures only, with the highest density being 500 burrows per 400 square meters. In open pastures the median burrow density was 3 (maximum=200) and 0 (maximum=500) for Ochotona spp. and Microtus spp. respectively. In fenced pastures the median burrow density was 10.5 (maximum=145) and 0 (maximum=126) for Ochotona spp. and Microtus spp. respectively.
The longer time pastures have been fenced led to lower burrow density of Ochotona spp. in pastures (χ2=3.980, P=0.046, coefficient = –0.042, 95% confidence interval –0.083 to –0.001). The procedure also found land pattern (χ2=52.680, P <0.0001, coefficient = –0.872, 95% confidence interval –1.107 to –0.636) and the interaction between fencing and land pattern (χ2=16.580, P <0.001, coefficient=–0.872, 95% confidence interval (0.087 to 0.025)) were also statistically significant (Table 2). When using burrow density of Microtus spp. as dependent variable in replacement of burrow density of Ochotona spp. while other variables kept the same, the procedure failed to find the burrow density of Microtus spp. in open pastures to be statistically higher than that in the fenced pastures (P=0.499).
|view in a new window |
Table 2. Factors influencing burrow density of Ochotona spp.
Relationship between small-mammal relative density and E. multilocularis infection in owned dogs
A total of 228 owned dogs were examined. The average abundance of worms (worm burden) was 39.64 (SD=352.73, maximum=5000). The worm burden in dogs was statistically significantly related to the maximum density of Ochotona spp. (χ2=5.250, P=0.022, coefficient=0.028, 95% confidence interval 0.004 to 0.051) (Table 3). Our modeling approach failed to detect significant relationships between the worm burden and the maximum burrow density of Microtus spp. (P=0.455), or owned dog age (P=0.052) and sex (P=0.059).
|view in a new window |
Table 3. Relationship between burrow density of small mammals and worm abundance of owned dogs
E. multilocularis prevalence was 14.8% based on copro-PCR. The prevalence in owned dogs was positively correlated to the number of stray dogs seen within a 200 meter radius (Wald χ2=8.375, P=0.004, odds ratio=1.198, 95% confidence interval 1.060–1.353).
Diet of definitive hosts
No significant differences in the proportion of food components in feces for owned dogs among 3 sites were detected (χ2=4.608, P=0.595). Small mammals were detected in 28.8% of all feces, but few remains allowing species or genus identification (e.g. teeth) were found. Ochotona spp. teeth were found in 1 dog feces in Qiwu. No Ochotona spp. hairs were identified in the 8 dog feces in which small mammal hairs were found in Tuanjie. No insectivore remains were identified in feces. Large bones, indicating predation on larger mammals (e.g. cattle), were present in 81.4% of all feces. (Table 4).
|view in a new window |
Table 4. Occurrence of food components in dog feces (n (%))
Significant differences were found in the proportions of food components for foxes between 2001 and 2002 (χ2=16.180, P=0.012). Ochotona spp. and Microtus spp. had the largest frequency of occurrence, in comparison to insectivores and Cricetulus kamensis. Both Ochotona spp. and Microtus spp. displayed large differences in their occurrence between years (Table 5).
|view in a new window |
Table 5. Occurrence of food components in fox feces (n (%))
Based on the dog data collected in 2001 and 2002 and small mammal data in 2007, this study is grounded in the assumption that the distribution of small mammal indices has been relatively stable during the last 6 years. Small mammals are known to be potentially cyclic and this is particularly the case for the genus Microtus and possibly Ochotona. Although such cycles cannot be excluded on the Tibetan plateau, multi-annual cycles have 3–6 years durations and holes (especially Ochotona) may last several years in the frozen soils of high altitude. This time discrepancy is, however, one of the limitations of our study.
Although foxes are viewed as one of the major sources for human exposure to E. multilocularis in Europe and Japan,3 dogs, in particular the owned dogs, have been found to be a major risk factor in the eastern Qinghai-Tibetan plateau.12,13 The proximity of domestic dogs to humans, the large number of owned dogs and the high number of stray dogs were assumed to be key reasons why dogs are more important than foxes as a risk factor for human AE.13 Therefore, to understand how these owned dogs become infected and what factors could be influential is essential to understanding the transmission of E. multilocularis to humans and subsequently provide information for control options.
Previous studies assumed that overgrazing was an important factor promoting dog E. multilocularis infection.9-11,14 It was shown that overgrazing promoted overall abundance of small mammals, e.g. Ochotona spp., Microtus spp. and Cricetulus kamensis,10,25 with the larger fenced pastures leading to higher burrow density of small mammals in the area of a township. The overall abundance of small mammals was then linked to the dog E. multilocularis copro-PCR infection.10,11 Overgrazing measured by fenced pastures, which assumed more fenced pasture led to higher abundance of small mammals in the open pastures, was further linked to human AE prevalence.9
The current study showed that grass height was negatively related to the burrow abundance of small mammals (Ochotona curzoniae mostly), and thus further supports previous assumptions that overgrazing increased the abundance of small mammals. Index methods were not designed to estimate small mammal population biomasses, but the huge contrast between the number of Ochotona and Microtus-like indices suggest that the biomass of Ochotona population is by far higher than that of microtine rodents. Even considering that the infection of larval stage of E. multilocularis in this Ochotona curzoniae was lower than that of Microtus fuscus,16,17 the former spp. role in transmission could not be underestimated due to its huge population. Here we found that the worm abundance of E. multilocularis in owned dogs was positively related to the maximum burrow density of Ochotona spp. and not to Microtus indices. This further supports a possible important role of Ochotona spp. in transmission to dogs.
The overgrazing hypothesis included both Ochotona spp., Microtus spp. and Cricetulus kamensis. Previous reports indicated that population increases of Ochotona curzoniae, Ochotona cansus, Microtus fuscus and Cricetulus kamensis could be associated with overgrazing.25,29,30 This research seemed to support that lower grass height, an indicator of grazing pressure, led to higher burrow density of small mammals. For smaller small mammals (mostly Microtus spp.), burrow density could reach 500 burrows per 400 square meters in open (non-fenced) pastures, which also indicates a patch distribution of Microtus fuscus, a typical pattern as previously described.31
No statistical relationship was found between the burrow density of smaller small mammals (mostly Microtus spp.) and E. multilocularis in owned dogs by the study but its role can not be ruled out because Microtus fuscus, Pitymys irene and Cricetulus kamensis were all found to harbour the larva stage infection of E. multilocularis in the liver (Raoul, unpublished data.).16,17 Furthermore, our data suggest that dog predation towards small mammals is primarily focused on species belonging to genus other than Ochotona (i,e., Microtus spp. and/or Cricetulus spp.). The main limitation however in dietary analysis of canid feces was the availability of components allowing species or genus identification, e.g., teeth. Compared to fox feces, dog feces analyzed in this study rarely displayed such material (hairs and small bones were the main components found). The shape of Ochotona spp. hairs in cross-section is characteristic of the genus, and cannot be mistaken for Microtus spp. or Cricetulus spp. (Raoul, unpublished data). However, the two latter genera cannot so far be distinguished using hair cross-section. Two fox species, the red fox (Vulpes vulpes) and the Tibetan fox (Vulpes ferrilata), are sympatric in the study area (authors' personal observations). Fox feces collected could not be formally attributed to one or to the other species. These results should therefore be taken as a global picture of food consumed by the two fox species, but if constant predation behavior of owned dogs across years is assumed, it tends to suggest that the predation of Ochotona spp. was not as frequent as the predation of Microtus spp. by owned dogs. The relative importance of Microtus spp. and/or Cricetulus spp. consumption, i.e., the other genera recorded in the area,25 is still to be clarified. Ochotona spp. and Microtus spp. were the main small mammal prey of foxes in the area. Their importance in the diet varied according to year, which may be attributed to fox functional response to prey density variation.14 Such variation in prey consumption by Vulpes vulpes according to prey availability has been shown in Europe for the water vole Arvicola terrestris, a grassland species.32 The hamster Cricetulus kamensis, although present in various habitats, was recorded in 3 out of 298 feces. Interaction with Ochotona spp. and Microtus spp. in fox prey choice may explain this low frequency of occurrence. Insectivores were not trapped in this area,25 suggesting a very low density and/or a distribution focused to spatially very restricted habitats. This could explain the quasi-absence of insectivores in fox and dog diet. Moreover, the aversion of Vulpes vulpes for insectivores is now well documented.33
The number of stray dogs observed in the vicinity was found to be positively associated with E. multilocularis infection in owned dogs measured by copro-PCR. Stray dogs were previously assumed as playing a major role in the transmission of E. multilocularis.13 This is the first time that a correlation has been found between the density of stray dog and the infection of owned dogs, which has been suggested to be a major transmission source of alveolar echinococcosis.12,13
The conclusions are that burrow density of pika (Ochotona spp.) is a statistically significant factor associated with E. multilocularis infection in owned dogs while voles (Microtus spp. and Cricetulus kamensis) are not. On the other hand, dietary analyses suggested that the role of Ochotona spp. might be less important than other species in the transmission of E. multilocularis to owned dogs. The huge biomass of pika in the high pastures may be able to sustain efficient transmission of E. multilocularis, despite relatively low prevalence and low average consumption by owned dogs. How a wildlife cycle including potentially two species of foxes and a domestic cycle including owned and stray dogs combines with several intermediate host categories such as Ochotona and Microtus is still unclear. However, this study indicates that more attention should be paid to quantifying the ecology of the definitive and intermediate host categories defined in this study.
1. Ministry of Health, China. Report on current situation of important human parasitic diseases. 2006 (Accessed February 5, 2009 at http://www.moh.gov.cn).
2. Budke CM, Qiu JM, Zinsstag J, Wang Q, Torgerson PR. Use of disability adjusted life years in the estimation of the disease burden of echinococcosis for a high endemic region of the Tibetan Plateau. Am J Trop Med Hyg 2004; 71: 56-64.
3. Vuitton DA, Zhou H, Bresson-Hadni S, Wang Q, Piarroux M, Raoul F, et al. Epidemiology of alveolar echinococcosis with particular reference to China and Europe. Parasitology 2003; 127 Suppl: s87-s107.
4. Li TY, Qiu JM, Yang W, Craig PS, Chen XW, Xiao N, et al. Echinococcosis in Tibetan populations, Western Sichuan Province, China. Emerging Infect Dis 2005; 11: 1866-1873.
5. Ammann RW, Eckert J. Clinical diagnosis and treatment of echinococcosis in humans. In: Thompson RCA, Lymbery AJ, eds. Echinococcus and hydatid disease. Wallingford: CAB International; 1995: 411-463.
6. Ammann RW, Eckert J. Cestodes: Echinococcus. Gastroenterol Clin N Am 1996; 25: 655-689.
7. Eckert J, Deplazes P, Graig PS, Gemmell MA, Gottstein B, Heath D, et al. Echinococcosis in animals: clinical aspects, diagnosis and treatment. In: Eckert J, Gemmell MA, Meslin FX, eds. WHO/OIE Manual on Echinococcosis in humans and animals. Paris: OIE: 2001: 75.
8. Budke CM, Campos-Ponce M, Wang Q, Torgerson PR. A canine purgation study and risk factor analysis for echinococcosis in a high endemic region of the Tibetan plateau. Vet Parasitol 2005; 127: 43-49.
9. Wang Q, Vuitton DA, Qiu JM, Giraudoux P, Xiao YF, Schantz PM, et al. Fenced pasture: a possible risk factor for human alveolar echinococcosis in Tibetan pastoralist communities of Sichuan, China. Acta Tropica 2004; 90: 285-293.
10. Wang Q, Vuitton DA, Xiao YF, Budke CM, Campos-Ponce M, Schantz PM, et al. Pasture types and Echinococcus multilocularis, Tibetan communities. Emerging Infec Dis 2006; 12: 1008-1009.
11. Wang Q, Xiao YF, Vuitton DA, Schantz PM, Raoul F, Budke C, et al. Impact of overgrazing on the transmission of Echinococcus multilocularis in Tibetan pastoral communities of Sichuan Province, China. Chin Med J 2007:120: 237-242.
12. Craig PS, Giraudoux P, Shi D, Bartholomot B, Barnish G., Delattre P. An epidemiological and ecological study of human alveolar echinococcosis transmission in south Gansu, China. Acta Trop 2000; 77: 167-77.
13. Wang Q, Qiu JM, Schantz PM, He JG, Ito A, Liu FJ. Risk factors for development of human hydatidosis among people whose family raising livestock in Western Sichuan Province, China. Chin J Parasite Dis Parasitol (Chin) 2001; 19: 289-293.
14. Giraudoux P, Pleydell D, Raoul F, Qurere JP, Wang Q, Yang Y, et al. Transmission ecology of Echinococcus multilocularis: what are the ranges of parasite stability among various host communities in China. Parasitol Int 2006; 55 Suppl: s237-s246.
15. Qiu JM, Liu FJ, Schantz P, Ito A, Carol D, He JG. Epidemiological survey of hydatidosis in Tibetan areas of Western Sichuan Province. XXXIII Archivos Internacionales de la Hidatidosis 1999; 84.
16. Xiao N, Nakao M, Qiu JM, Budke CM, Giraudoux P, Craig, PS, et al. Short report: dual infection of animal hosts with different Echinococcus species in the eastern Qinghai-Tibet Plateau region of China. Am J Trop Med Hyg 2006; 75: 292-294.
17. Xiao N, Qiu J, Nakao M, Li T, Yang W, Chen XW, et al. Echinococcus shiquicus n. sp., a taeniid cestode from Tibetan foxes and Plateau pikas in China. Intern J Parasitol 2005; 35: 693-701.
18. Editorial Commission of Shiqu County Record. Shiqu county record. Chengdu: The People's Publication House of Sichuan Province; 2000: 55-137.
19. Hansson L. Field signs as indicators of vole abundance. J Applied Ecol 1979; 16: 339-347.
20. Giraudoux P, Pradier B, Delattre P, Deblay S, Salvi D. Estimation of water vole abundance, by using surface indices. Acta Theriologica; 1995: 40: 77-96.
21. Fichet-Calvet E, Jomâa I, Giraudoux P, Ashford RW. Estimation of sand rat abundance by using surface indices. Acta Theriologica 1999: 44: 353-352.
22. Quere JP, Raoul F, Giraudoux P, Delattre P. An index method applicable at landscape scale to estimate relative population densities of the common vole (Microtus arvalis). Rev Ecol-terr Life 2000; 55: 25-32.
23. Giraudoux P, Delattre P, Habert M, Quéré JP, Deblay S, Defaut R. Population dynamics of fossorial water vole (Arvicola terrestris scherman): a land usage and landscape perspective. Agricult Ecosyst Envir 1997; 66: 47-60.
24. Raoul F, Defaut R, Michelat D, Montadert M, Pépin D, Quere JP. Landscape effects on the populations dynamics of small mammal communities: a preliminary analysis of prey-resource variations. Rev Ecol-terr Life 2001; 56; 339-352.
25. Raoul F, Quere JP, Rieffel D, Bernard N, Takahashi K, Scheifler R, et al. Distribution of small mammals in a pastoral landscape of the Tibetan plateaus (Western Sichuan, China) and relationship with grazing practices. Mammalia 2006; 42: 214-225.
26. Dinkel A, von Nickisch-Rosenegk M, Bilger B, Merli M, Lucius R, Romig T. Detection of Echinococcus multilocularis in the definitive host: coprodiagnosis by PCR as an alternative to necropsy. J Clin Microbiol 1998; 36: 1871-1876.
27. van der Giessen JW, Rombout YB, Franchimont JH, Limper LP, Homan WL. Detection of Echinococcus multilocularis in foxes in The Netherlands. Vet Parasitol 1999; 82: 49-57.
28. Zabala J, Zuberogoitia I. Badger, Meles meles (Mustelidae, Carnivora), diet assessed through scat-analysis: a comparison and critique of different methods. Folia Zoologica 2003; 52: 23-30.
29. Li FM. Out of control of the rodents in Shiqu county and handling strategies. Sichuan Grassland 1995; 3: 27-30.
30. Hou XM. The current situation of rodents and its control methods in the resource area of Qingnan pasture. Sichuan Grassland 2001; 1: 28-31.
31. Zhang YM, Zhang ZB, Liu JK. Burrowing rodents as ecosystem engineers: the ecology and management of plateau zokors Myospalax fontanierii in alpine meadow ecosystems on the Tibetan Plateau. Mammal Review 2003; 33: 284-294.
32. Weber JM, Aubry S. Predation by foxes, Vulpes vulpes, on the fossorial form of the water vole, Arvicola terrestris scherman, in western Switzerland. J Zool Lond 1993; 229: 553-559.
33. Artois M. Le Renard roux (Vulpes vulpes Linnaeus, 1758). In: Encyclopédie des carnivores de France. Artois M, Delatre P, eds. SFEPM: Nort/Erdre; 1989: 90.