With the increase in the number of fast-moving vehicles and the simultaneous development of road building technology, roadside bird mortality has become an increasingly important environmental issue that has the potential to do serious damage to already vulnerable bird populations. This project was therefore initiated after an alarming number of owls were found dead along the N17 and R550 roads in the rural areas between Springs and Devon in the East Rand highveld of Gauteng Province. Five hundred and fifty four owls of four species, namely Marsh Owls, the Red Data listed Grass Owl, Barn Owl and the Spotted Eagle Owl were collected on the stretches of the R550 and N17 during the period between October 2001 and September 2003. They accounted for 53.6%, 27.4%, 17.5% and 1.3% of the mortalities respectively. Unidentified species of owls accounted for the remaining 0.2%. It was found that the monthly mortality rates of the birds varied throughout the year, with the greatest losses being suffered during July, as opposed to relatively lower mortalities occurring during the warmer months. All carcasses were collected, their GPS locations plotted on a map, and ‘hotspot areas’ identified as places of highest incidences of mortalities of the owls. Vehicle-induced mortalities are then discussed in relation to these hotspots, in terms of vegetation and habitat descriptions, daily vehicle counts along the route, as well as fixtures found along the route. It was found that traffic density was indirectly proportional to owl mortalities, with higher traffic speeds definitely having an increasingly detrimental effect on the owl mortalities. Weather conditions also play a role in mortality counts, with the mortalities being significantly negatively correlated to rainfall. Moon phases were also related to the times of highest mortalities, however, this factor did not play a significant role in influencing road mortalities. Another factor that was studied was the influence of differing tarmac road surface temperatures as opposed to gravel road verge surface temperatures, and how these temperatures differed from the ambient temperature. It was found that there was not a big enough difference in the temperatures that would warrant (the previously thought notion) that the owls were attracted to the roads at night to gain heat. Gravel roads had very low incidences of owl mortalities with the highest mortalities recorded along tarmac roads that are bordered by open grasslands or cattle grazing paddocks. It was shown that tarmac roads, bordered by croplands, had a lesser effect on the owl mortalities. Another factor influencing the road mortalities of the owls is grain that is spilled on the road during transport. This initially seemed to be the major factor in attracting granivorous rodents to the roads, and in turn, attracting the owls to prey on them. Rodents identified from recovered pellets and the stomach contents of dead owls confirmed the fact that the majority of rodent prey items were indeed granivorous species, namely Mastomys natalensis and Rhabdomys pumilio. This is in disagreement with previous studies that indicated that a large proportion of the prey species of the Grass Owl was Otomys irroratus, a grass-eating species, even though a rodent trapping study to determine prey abundance within the area indicated a healthy population of O. irroratus. Further studies into prey items of the owls that were dissected indicated that the majority of the prey items recovered were not caught directly on the road as it was already partially digested, suggesting that the prey was caught prior to the owl being killed. From the pellet analyses, other prey items were also found to form part of the owls’ diets. It was found that insects formed largely the diet of the Marsh Owl during the spring, summer and autumn months, with them resorting to smaller rodents during the winter months. Spotted Eagle Owls also preyed almost exclusively on insects. Grass Owls, on the other hand, preyed on small mammals exclusively, with the very rare exception of some insects also being taken. Of all of the dead owls recovered on the roads, post mortems were carried out on only 78 of the carcasses. All of the dead owls examined were in good health prior to death. Various morphometrics of the examined owls were noted. Comparisons of body mass showed that females were larger than males for most species. This was also found for most other measurements as well. It was also found that, according to body mass comparisons, Barn Owls and Marsh Owls were significantly similar. Conducting ANOVA analysis on other morphometrics to determine gender differences, it was found that Grass Owl males were significantly different to females in terms of body mass and length. Marsh Owl males were significantly different to females in terms of body mass and tail length; with Barn Owl males being significantly different in terms of tarsus length to females. Except for Spotted Eagle Owl tarsus lengths all other measurements were in favour of females being significantly larger. These findings were also confirmed when applying the Dimorphism Index to all morphometrics measurements, especially body mass. The degree of parasite infestation was also studied during post mortem examinations. Very few cestodes and nematodes were found, with too few to have an effect on the overall health of the birds prior to death. The vegetation type was studied at transects that coincided with hotspot and non-hotspot sites. Using the PRIMER statistical software package, hotspot sites were found to have highest plant cover and diversity, whereas nonhotspot sites showed lowest plant cover and diversity, generally dominated by Hyparrhenia hirta. During these studies, the degree of available nesting habitat was determined and nesting sites were identified, using the ropedragging technique to flush out roosting and nesting owls that would otherwise have been impossible to find in the thick cover. It was found that Grass Owls preferred a habitat rich in thick grass cover that was relatively high (0.75 m–1 m). It was found that the grass species preferred by these owls were Eragrostis curvula, Paspalum sp., Setaria sp., Sporobolis sp., with few other small herbaceous plants. Marsh Owls, on the other hand, seemed not to be too partial regarding roosting and nesting sites, with them roosting and breeding in more mixed vegetation grasslands that had sparser cover, not reaching the height of the grass cover typical of the habitat preferred by the Grass Owls. Opposed to this were the non-hotspot vegetation sites. These sites were found to have vegetation cover unsuited to both the Marsh and Grass owls, with mixed H. hirta grassland not forming the dense cover, or the height, needed by those two species of owls. Foraging owls were also observed, with the vegetation type in the immediate vicinity noted. Vegetation types similar to breeding areas were noted in these foraging areas. Habitat preferences as well as breeding performance were noted for both grassland species of owls, and found to be directly related to land usage in terms of varying agricultural practices and regimes. Fallow, undisturbed lands were found to be highly productive for the owls. Lands planted with Eragrostis sp. were also found to be very productive, but only if left undisturbed for a period of time sufficient to allow the grassland owls to colonise it. Maize-planted fields were found to be utilised only as foraging fields and no breeding of owls was found to take place close to these fields. After extensive nest searching, it was found that both Marsh and Grass owls were breeding from late March to early June, with the Barn Owls breeding in October and again in March. Grass Owls occur in the study area because of the presence of a natural corridor of suitable habitat that runs parallel with the Blesbokspruit. This favourable habitat of the study area is thus conducive to high population density of grassland owls utilising these uncultivated patches of dense and tall vegetation. The high incidence of mortalities on the road in the study area is due to the concomitant high population densities. This healthy population seems to be sustaining the losses occurring on roads. Owls also seem to be gathering in larger numbers in hotspot zones because of the easy available prey, which are attracted to these high productive areas. Agricultural practices in the area lead to the spillage of grain on the road during transportation. Potential prey species foraging on the roads expose themselves to the nocturnal hunters offering an easy dinner. This process leaves these owls vulnerable to vehicle collisions. The overall population size may be larger than previously thought, not with standing the high mortalities already recorded. The small patches of viable habitat in the study area remains suitable for the breeding of the two grassland owl species allowing for such high densities to occur in the area. The Grass Owl, nonetheless, remains severely threatened as it already occurs as a high priority species for conservation concern in the Gauteng Province. This study provides the first assessment of this owl species of this scale in South Africa and this will ultimately promote the long-term survival of these owls. / Dr. V. Wepener

With the increase in the number of fast-moving vehicles and the simultaneous development of road building technology, roadside bird mortality has become an increasingly important environmental issue that has the potential to do serious damage to already vulnerable bird populations. This project was therefore initiated after an alarming number of owls were found dead along the N17 and R550 roads in the rural areas between Springs and Devon in the East Rand highveld of Gauteng Province. Five hundred and fifty four owls of four species, namely Marsh Owls, the Red Data listed Grass Owl, Barn Owl and the Spotted Eagle Owl were collected on the stretches of the R550 and N17 during the period between October 2001 and September 2003. They accounted for 53.6%, 27.4%, 17.5% and 1.3% of the mortalities respectively. Unidentified species of owls accounted for the remaining 0.2%. It was found that the monthly mortality rates of the birds varied throughout the year, with the greatest losses being suffered during July, as opposed to relatively lower mortalities occurring during the warmer months. All carcasses were collected, their GPS locations plotted on a map, and ‘hotspot areas’ identified as places of highest incidences of mortalities of the owls. Vehicle-induced mortalities are then discussed in relation to these hotspots, in terms of vegetation and habitat descriptions, daily vehicle counts along the route, as well as fixtures found along the route. It was found that traffic density was indirectly proportional to owl mortalities, with higher traffic speeds definitely having an increasingly detrimental effect on the owl mortalities. Weather conditions also play a role in mortality counts, with the mortalities being significantly negatively correlated to rainfall. Moon phases were also related to the times of highest mortalities, however, this factor did not play a significant role in influencing road mortalities. Another factor that was studied was the influence of differing tarmac road surface temperatures as opposed to gravel road verge surface temperatures, and how these temperatures differed from the ambient temperature. It was found that there was not a big enough difference in the temperatures that would warrant (the previously thought notion) that the owls were attracted to the roads at night to gain heat. Gravel roads had very low incidences of owl mortalities with the highest mortalities recorded along tarmac roads that are bordered by open grasslands or cattle grazing paddocks. It was shown that tarmac roads, bordered by croplands, had a lesser effect on the owl mortalities. Another factor influencing the road mortalities of the owls is grain that is spilled on the road during transport. This initially seemed to be the major factor in attracting granivorous rodents to the roads, and in turn, attracting the owls to prey on them. Rodents identified from recovered pellets and the stomach contents of dead owls confirmed the fact that the majority of rodent prey items were indeed granivorous species, namely Mastomys natalensis and Rhabdomys pumilio. This is in disagreement with previous studies that indicated that a large proportion of the prey species of the Grass Owl was Otomys irroratus, a grass-eating species, even though a rodent trapping study to determine prey abundance within the area indicated a healthy population of O. irroratus. Further studies into prey items of the owls that were dissected indicated that the majority of the prey items recovered were not caught directly on the road as it was already partially digested, suggesting that the prey was caught prior to the owl being killed. From the pellet analyses, other prey items were also found to form part of the owls’ diets. It was found that insects formed largely the diet of the Marsh Owl during the spring, summer and autumn months, with them resorting to smaller rodents during the winter months. Spotted Eagle Owls also preyed almost exclusively on insects. Grass Owls, on the other hand, preyed on small mammals exclusively, with the very rare exception of some insects also being taken. Of all of the dead owls recovered on the roads, post mortems were carried out on only 78 of the carcasses. All of the dead owls examined were in good health prior to death. Various morphometrics of the examined owls were noted. Comparisons of body mass showed that females were larger than males for most species. This was also found for most other measurements as well. It was also found that, according to body mass comparisons, Barn Owls and Marsh Owls were significantly similar. Conducting ANOVA analysis on other morphometrics to determine gender differences, it was found that Grass Owl males were significantly different to females in terms of body mass and length. Marsh Owl males were significantly different to females in terms of body mass and tail length; with Barn Owl males being significantly different in terms of tarsus length to females. Except for Spotted Eagle Owl tarsus lengths all other measurements were in favour of females being significantly larger. These findings were also confirmed when applying the Dimorphism Index to all morphometrics measurements, especially body mass. The degree of parasite infestation was also studied during post mortem examinations. Very few cestodes and nematodes were found, with too few to have an effect on the overall health of the birds prior to death. The vegetation type was studied at transects that coincided with hotspot and non-hotspot sites. Using the PRIMER statistical software package, hotspot sites were found to have highest plant cover and diversity, whereas nonhotspot sites showed lowest plant cover and diversity, generally dominated by Hyparrhenia hirta. During these studies, the degree of available nesting habitat was determined and nesting sites were identified, using the ropedragging technique to flush out roosting and nesting owls that would otherwise have been impossible to find in the thick cover. It was found that Grass Owls preferred a habitat rich in thick grass cover that was relatively high (0.75 m–1 m). It was found that the grass species preferred by these owls were Eragrostis curvula, Paspalum sp., Setaria sp., Sporobolis sp., with few other small herbaceous plants. Marsh Owls, on the other hand, seemed not to be too partial regarding roosting and nesting sites, with them roosting and breeding in more mixed vegetation grasslands that had sparser cover, not reaching the height of the grass cover typical of the habitat preferred by the Grass Owls. Opposed to this were the non-hotspot vegetation sites. These sites were found to have vegetation cover unsuited to both the Marsh and Grass owls, with mixed H. hirta grassland not forming the dense cover, or the height, needed by those two species of owls. Foraging owls were also observed, with the vegetation type in the immediate vicinity noted. Vegetation types similar to breeding areas were noted in these foraging areas. Habitat preferences as well as breeding performance were noted for both grassland species of owls, and found to be directly related to land usage in terms of varying agricultural practices and regimes. Fallow, undisturbed lands were found to be highly productive for the owls. Lands planted with Eragrostis sp. were also found to be very productive, but only if left undisturbed for a period of time sufficient to allow the grassland owls to colonise it. Maize-planted fields were found to be utilised only as foraging fields and no breeding of owls was found to take place close to these fields. After extensive nest searching, it was found that both Marsh and Grass owls were breeding from late March to early June, with the Barn Owls breeding in October and again in March. Grass Owls occur in the study area because of the presence of a natural corridor of suitable habitat that runs parallel with the Blesbokspruit. This favourable habitat of the study area is thus conducive to high population density of grassland owls utilising these uncultivated patches of dense and tall vegetation. The high incidence of mortalities on the road in the study area is due to the concomitant high population densities. This healthy population seems to be sustaining the losses occurring on roads. Owls also seem to be gathering in larger numbers in hotspot zones because of the easy available prey, which are attracted to these high productive areas. Agricultural practices in the area lead to the spillage of grain on the road during transportation. Potential prey species foraging on the roads expose themselves to the nocturnal hunters offering an easy dinner. This process leaves these owls vulnerable to vehicle collisions. The overall population size may be larger than previously thought, not with standing the high mortalities already recorded. The small patches of viable habitat in the study area remains suitable for the breeding of the two grassland owl species allowing for such high densities to occur in the area. The Grass Owl, nonetheless, remains severely threatened as it already occurs as a high priority species for conservation concern in the Gauteng Province. This study provides the first assessment of this owl species of this scale in South Africa and this will ultimately promote the long-term survival of these owls. / Dr. V. Wepener