South Africa harbors 13 tortoise species within its borders, more than any other country (Ernst et al. 2000). Three species are listed by the International Union for Conservation of Nature (IUCN) as Lower Risk/near threatened or Endangered (Branch 1996). Due to the delayed maturity, iteroparous reproduction, and sexual differences in behaviors in chelonians, mortality by road traffic may impact the demography and dynamics of populations (Steen and Gibbs 2004; Aresco 2005a; Gibbs and Steen 2005; Boarman and Sazaki 2006; Beaudry et al. 2008). Although this threat has not been evaluated for tortoises in South Africa, the ranges of species throughout the country, in combination with a relatively dense road network, suggests that roads may affect some populations.

The greater padloper, Homopus femoralis, is not listed by the IUCN, but is endemic to South Africa and Lesotho (Branch 2008). It is a small, relatively flat species that reaches maximum straight-line carapace length and shell height of 168 and 70 mm, respectively (Boycott and Bourquin 2000; Branch 2008), making it inconspicuous and vulnerable when crossing roads. Its slow reproductive rate may increase population sensitivity to individuals killed by traffic: females produce clutches that contain up to 3 eggs (Branch 1999; Boycott and Bourquin 2000; Kuperus and Loehr 2009), and captive observations suggest that only 1 clutch is produced annually (Loehr 2009 and unpubl. data). There are no data available on habitat use or population density and dynamics for H. femoralis. In this paper, I report the number and distribution of dead H. femoralis along a road, and identify an opportunity to mitigate road mortality.

Methods

The study area consisted of 8.4 km of a tarred 2-lane trunk road in the Karoo (Beaufort West municipality, coordinates deposited at the Scientific Services Unit, CapeNature, Western Cape, South Africa), and included the road shoulders (Fig. 1). Shoulders had widths of ca. 5 m and were mostly bare, with few isolated small shrubs and annuals. The dense vegetation that grew > 5 m from the tar, bordering the road shoulders, was not included in the study area. The road, which was selected based on occasional observations of crossing H. femoralis, had low traffic volume, averaging only 62 vehicles per day projected for 2010 (Provincial Government of the Western Cape 2010).

View Figure

• Download as Powerpoint
• Download Figure

From 18 February to 21 February 2010, 2 people inspected the road and its shoulders on foot (average walking speed 2.9 km/hr) for H. femoralis remains, each person covering one side of the road. Homopus femoralis shells were readily distinguishable from remains of the sympatric tortoises Psammobates tentorius and Stigmochelys pardalis, based on the coloration of the scutes (plain brown-yellowish, or dark edged, but not dark radiated or dotted), the small, depressed shell shape, and the presence of 4 claws on the forelimbs. We carefully inspected scattered shell fragments to determine if they originated from 1 or more individuals.

For each H. femoralis, I recorded the geographic coordinates (Garmin eTrex H, Garmin International, Olathe, KS). All coordinates were plotted on an aerial photograph (ArcMap 9.3, ESRI, Redlands, CA) to evaluate their distribution.

Results

Forty-eight dead H. femoralis were identified (Fig. 2). Remains included few recently killed individuals (Fig. 1) and many mummified tortoises (Fig. 3) and shell fragments originating from tortoises that appeared to have died several years earlier (see Dodd 1995). In addition to H. femoralis, we found 3 dead P. tentorius.

View Figure

• Download as Powerpoint
• Download Figure

View Figure

• Download as Powerpoint
• Download Figure

Tortoise remains were not distributed homogeneously. The northwestern portion of the road had a relatively high number of remains (Fig. 2). This portion had a length of 1.5 km (18% of the total road length) and accounted for 27 dead H. femoralis (56% of the total for the species). It was located in a relatively even area with small shrubs, well distant from rocky terrain.

Discussion

The road that I studied appears to be responsible for relatively high mortality of H. femoralis, despite the low traffic. Although local avian predators may kill tortoises by dropping them on hard surfaces, the road did not appear to serve this purpose during fieldwork in spring and summer 2006–2011 (V.J.T. Loehr, pers. obs.), and mummified tortoises had their extremities still attached (e.g., Fig. 3), suggesting that mortality was the result of collisions with road traffic. The number of shells counted was high because it had accumulated over several years, but the small shell size and exposure to scavengers (e.g., Antworth et al. 2005), sun, precipitation, wind, and turbulence from passing traffic probably facilitated shell fragmentation and decomposition, leading to underestimation. Moreover, tortoises that were injured by traffic and died at greater distances from the road were not counted. The large number of dead H. femoralis might relate to a high population density or result from reproductive or dispersal behavior. Furthermore, live tortoises used the road to drink during rainfall, a time when drivers have limited sight.

Herpetological studies have shown that road mortality is often spatially heterogeneous (Langen et al. 2007; Beaudry et al. 2008). This appears also to be the case for H. femoralis; most tortoises were positioned over a relatively short stretch of road. Without data on H. femoralis populations and mortality on adjacent roads, I cannot draw conclusions regarding mitigation needs, but the road design at the locality of highest mortality would enable mitigation relatively easily. Culverts under the road already facilitate the drainage of incidental rain water and may be adapted for tortoise use by adding barriers or fences along the road, which would lead tortoises to the culverts (Guyot and Clobert 1997; Aresco 2005b). Mitigation measures for minor roads may provide an important contribution to population viability (van Langevelde et al. 2009).

The methods used in this study did not permit an attempt to identify temporal variation in road mortality or population effects. Nevertheless, the methodology used provides a quick (5.8 man-hrs for 8.4 km of road) and easy (remains were easily identifiable, and fieldwork did not depend on tortoise activity) way to determine if an existing road may potentially affect local tortoise populations. For the planning of costly measures, additional study should reveal population estimates and dynamics, preferably over a number of years, to determine if road mortality has a detrimental effect on populations.