20. Preference of goats and sheep for browse species under field conditions
R. S. Pande* P. D. Kemp† J. Hodgson
(Published in: New Zealand Journal of Agricultural Research, 2002, Vol. 45: 97–102
0028–8233/02/4502–0097 $7.00/0 © The Royal Society of New Zealand 2002 New Zealand Journal of Agricultural Research www.informaworld.com/smpp/.../title~db=all~content=g919964384)
The preference of goats and sheep for browse species in mixed browse and pasture conditions in New Zealand was determined using 6 goats and 6 sheep observed for 2-h periods on 6 occasions between February and December. The goats and sheep were given free access to established nurseries. The 2 nurseries contained 6 types of leguminous shrubs, 3 non-leguminous shrub species, and 2 erect grass species, growing in rows with a grass-dominant temperate pasture. Preference was determined with an index that took into account the proportion of browsing observations and the relative abundance of each browse species. Goats browsed in 45% and sheep in 12% of the observations. The three most preferred species for goats and sheep were the same, but the order of preference differed. The overlap coefficient for all browse species (where 0 = no overlap and 1 = identical diets) ranged from 0.47 in winter to 0.79 in summer. The results are contrasted with those from a cafeteria-style indoor experiment that used the same browse and animal species.
Keywords grazing ecology; goats; sheep; browsing; preference; diet selection; browse species
Grazing animals continuously discriminate between vegetation types and plant parts (Hodgson 1986). Such activity is usually greater under range conditions and maximal when animals are browsing (Van Soest 1982). Goats are regarded as more efficient users of browse than sheep (Van Soest 1982; Devendra & Burns 1983), but scant information is available on the preference of goats and sheep for the browse species commonly available in New Zealand. Lambert et al. (1989a) compared the preference of goats and sheep for a range of browse species in a cafeteria-style indoor experiment. The multi-species nurseries used to provide the browse material in the experiment by Lambert et al. (1989a) were used in the field experiment reported here, which was designed to study the preference for browse species and the browsing activity of goats and sheep on field sites with multiple browse and pasture species available.
MATERIALS AND METHODS
The experiment was carried out at the Ballantrae Hill Research Station, AgResearch Grasslands, (175°50E, 40°18S, 150 m a.s.l.), 20 km from Palmerston North, New Zealand. Observations took place at intervals from February 1989 to December 1989 inclusive.
There were two established nurseries each containing 10 different browse species, with two forms of gorse (Table 1). Short spine gorse is an ecotype with shorter spines and a denser bush than typically found on gorse. The browse species were 3–4 years of age and grown in hedges consisting of 5–8 plants per hedge. The hedges were usually 5 m long and were between 0.4 and 1.7 m high. At each nursery, hedges of short spine gorse, black locust, and ceanothus appeared only once and hedges of †Author for correspondence. the other eight browse species were replicated three times.
Table 1 Browse species in the nurseries.
Common name Scientific name
Chamaecytisus palmensis (Christ) Bisby et K.Nicholls
Broom Cytisus scoparius (L.) Link.
Tree medic Medicago arborea L.
Black locust Robinia pseudoacacia L.
Gorse Ulex europaeus L.
Short spine gorse Ulex europaeus L.
Tauhinu Ozothamnus leptophyllus (G.Forst.) Breitw. et J.M.Ward
Ceanothus Ceanothus griseus (Trel.) McMinn
Manuka Leptospermum scoparium J.R. et G.Forst.
Toetoe Cortaderia fulvida (Buchanan) Zotov
Pampas Cortaderia selloana (Schult. et Schult.f.) Asch. et Graebn.
Main pasture species growing underneath the browse species were browntop (Agrostis capillaris), crested dogstail (Cynosurus cristatus), cocksfoot (Dactylis glomerata), perennial ryegrass (Lolium perenne), Yorkshire fog (Holcus lanatus), and white clover (Trifolium repens).
Six dry, angora-cross female goats and 6 dry, pure Romney female sheep were used. The animals were 17 months of age at the start of the experiment, and the mean body weight was 18.2 ± 0.4 kg and 43.5 ± 1.1 kg for goats and sheep, respectively. The animals were kept overnight in an adjacent shearing shed prior to each observation period, and provided with water but no feed, to encourage feeding during the observation period. The previous grazing experiences of the animals were on hill pastures.
Table 2 Animal observation and nursery schedule.
Observation date (1989) Nursery 1 Nursery 2 Season
22 Feb sheep goats summer
23 Mar goats sheep autumn
27 Apr sheep goats autumn
23 May sheep goats winter
22 Jun goats sheep winter
24 Nov goats sheep summer
The goats and sheep were allowed free choice of all plant species in their allocated nursery. Each observation period was 2 h (10.00–12.00 h) starting from release of the animals into the nurseries.Observation dates were scheduled to cover summer, autumn, and winter seasons, and were arranged in a sequence of approximately 4-week intervals between 22 February and 22 June 1989, with a final observation on 24 November 1989. The animal species were alternated between nurseries (Table 2).
The preferences of the goats and sheep for browse species were quantified on the basis of the animal activities; browsing, grazing, and idling. The interval sampling technique was used to record animal activities at 2-minute intervals (Hodgson 1982). A shearing shed about 200 m from the nurseries was used as the observation post so as to minimise the influence of the observer on animal activities (Jamieson & Hodgson 1979). Distribution of the browsing activities in relation to the arrangement of the browse species in each nursery and the defoliation of plant parts are described in detail in Pande (1990).
The preference of goats and sheep for the individual browse species was adjusted for both the difference in the proportion of time spent browsing by goats and sheep, and the different abundance of the browse species. These adjustments were made by calculating, for each nursery, the proportion of browsing observations (PBO) each animal species spent browsing each browse species, and by calculating the relative abundance value (RAV) for each browse species. The RAV for each replicate of each browse species was calculated from the length, width, height, and density score of the plants in each hedge (Whittaker 1975). Density of the foliage of the browse plants was ranked visually from 1 (lowest) to 10 (highest).
The PBO and RAV were used to calculate the browsing preference index (BPI) given below (see Pande 1990).
BPI = (PBOi/ RAVi/ ni=1 (PBOi/ RAVi )
An overlap coefficient (C) using the simplified Morisita Index (Horn 1966) was used to measure the overlap for preferred browse species between sheep and goats. The Cranges from 0 (no overlap) to 1 (identical diets).
C=2ni-1 Xi Yi/ni=1Xi +ni=1 Yi
Xi = PBO for animal species X, for browse species i
Yi = PBO for animal species Y, for browse species i.
The effect of animal species, browse species, nursery, and season and their interactions on RAV, PBO, and BPI were analysed with the SAS General Linear Model (SAS 1988).
Goats and sheep were browsing during 45 and 12% of observations, respectively (Table 3). Conversely, sheep grazed more often than goats (Table 3). Both animal species spent a similar time idling, but this was in 8% or less of all observations (Table 3). There was no significant effect of season on the observations of animal activities.
Table 3 The percentage of observations of animal activities in the categories of browsing, grazing, and idling, expressed as means for the six observation periods. NS, not significant; *, P < 0.05 for within row mean comparison
Animal activities Goats Sheep
Browsing 44.7 11.6*
Grazing 48.6 80.5
Idling 6.7 8.0 NS
SEM = ±1.3
Proportional browsing observations (PBO)
There was a significant effect of browse species on PBO, and the animal species by browse species interaction was also significant (P < 0.001; Pande 1990). The other main effects and interactions were not significant. Goats were observed to browse tree medic most often, whereas sheep most often browsed black locust (Table 4). Tree medic, black locust, tagasaste, and ceanothus were often browsed by both goats and sheep, whereas tauhinu, toetoe, and pampas were rarely browsed by either animal species (Table 4). Short spine gorse and gorse were regularly browsed by goats but more rarely by sheep (Table 4).
Table 4 The proportion of browsing observations (PBO), the relative abundance value (RAV), and the browsing preference index (BPI) for each browse species for goats and sheep over 1 year.
Species Goat Sheep
PBO RAV BPI PBO RAV BPI
Tree medic 0.268 0.001 0.837 0.215 0.001 0.790
Short spine gorse 0.249 0.003 0.154 0.056 0.002 0.053
Black locust 0.178 0.007 0.066 0.398 0.013 0.149
Tagasaste 0.151 0.076 0.007 0.121 0.072 0.007
Ceanothus 0.092 0.042 0.006 0.125 0.044 0.007
Broom 0.048 0.059 0.003 0.093 0.058 0.014
Gorse 0.143 0.252 0.001 0.041 0257 0.002
Manuka 0.012 0.089 0.001 0.009 0.085 0.001
Tauhinu 0.002 0.111 0.000 0.013 0.107 0.001
Toetoe 0.006 0.176 0.000 0.018 0.175 0.001
Pampas 0.002 0.182 0.000 0.000 0.184 0.000
SEM 0.008 0.031 0.024 0.008 0.031 0.024
Relative abundance value (RAV)
The RAV of the browse species differed significantly (P < 0.0001). The RAV was unaffected by season, nursery, or animal species. The only significant interaction effect was nursery by browse species (P < 0.001), reflecting a difference in the number and hedge size of some browse species between nurseries (Pande 1990). The four most abundant browse species in terms of RAV were gorse, pampas, toetoe, and tauhinu, whereas tree medic, short spine gorse, and black locust were the least abundant (Table 4).
Browsing preference index (PBI)
The main effect of browse species on BPI was significant (P < 0.001), but the main effects of animal species, season, and nursery were not significant (Pande 1990). The nursery by browse species and the animal species by nursery by browse species interactions were significant (P < 0.001; Pande 1990). For both goats and sheep the greatest browsing preference, based on BPI, was for tree medic. Goats preferred short spine gorse and then black locust after tree medic, whereas sheep preferred black locust and then short spine gorse (Table 4). The two animal species were similar in their preference for the other browse species (Table 4).
The BPI gave different results for the preference of goats and sheep for some browse species than the PBO due to the differences in abundance as measured by RAV (Table 4). In particular, the PBO ranked gorse as the fifth most preferred browse species by goats, yet due to its high RAV gorse was ranked seventh using the BPI (Table 4). Similarly, black locust was the highest ranked browse species for sheep using the PBO but was ranked third using the BPI (Table 4).
Overlap coefficient (C)
The overlap coefficients for the preferred browse species by goats and sheep were 0.79 (November) and 0.71 (February) in summer, 0.62 (March) and 0.64 (April) in autumn, and 0.47 (May) and 0.58 (June) in winter, respectively. Hence, the greatest overlap between goats and sheep was in summer and the least in winter.
The results supported the view that goats are intermediate browsers and sheep are intermediate grazers (Van Soest 1982). Nevertheless, when only the browsing activity of the goats and sheep was considered there was a high degree of overlap on the preferred browse species. Both animal species showed the greatest preference for tree medic, short spine gorse, and black locust according to the BPI. Although the overlap coefficient (C) for goats and sheep was high, the degree of overlap was seasonally dependent. The availability of species such as black locust was lower in winter, and sheep were observed to browse ceanothus more in winter which suggests that the palatability of at least this species was seasonal.
The use of the PBI enabled the preference of goats and sheep for the individual browse species to be ranked independently of the differences in the abundance of the browse species. Although the most preferred species were the least abundant this appeared to be a coincidental result. The BPI, compared with using the proportion of PBO, gave a different preference ranking to some browse species, but the general trend was similar for both indices. The adjustment for the abundance of a browse species (RAV) affected the preference ranking of species like gorse and black locust, but largely demonstrated that tree medic, despite its low abundance, was the most preferred species. The BPI would appear to be a useful method for adjusting animal preference observations made on mixtures of browse species of varying abundance.
The browse species could be readily placed into one of three preference groups of high, moderate, and low preference for both animal species, although the ranking of individual browse species showed some variation between goats and sheep. The high preference group was tree medic, short spine gorse, and black locust, the moderate group comprised tagasaste, ceanothus, and broom, and members of the poor group were gorse, manuka, tauhinu, toetoe, and pampas. This ranking relates well to the digestibility of the browse species, with the notable exception of short spine gorse (Lambert et al. 1989b), but corresponds more poorly, especially for sheep, with the rankings obtained from a cafeteria-style indoor experiment (Lambert et al. 1989a).
Lambert et al. (1989b) reported the in vivo digestibility of tree medic, tagasaste, and black locust as being in the range 76–81%; ceanothus and broom 71–73%; short spine gorse, tauhinu, and gorse 63–67%; and manuka, toetoe, and pampas 47–48%. The high preference for short spine gorse by both goats and sheep in our experiment Pande et al.—Preference of goats and sheep for browse species 101 suggested that digestibility was not the only factor affecting its palatability.
Many factors such as digestibility, nutrient content, smell, taste, and physical properties (e.g., hairs, thorns) influence the palatability of plant species, and animal preference is also affected by factors such as access to and the availability of plant material (Crawley 1983). No attempt was made in this experiment to directly determine the factors that influenced animal preference though some observations can be made. The high digestibility of the most preferred browse species would have provided positive post-ingestive consequences to the sheep and goats, thereby reinforcing the animals’ preference (Bryant et al. 1991).
Additionally, tree medic, black locust, and short spine gorse are all legumes with consequent reasonable levels of protein. At least some of the browse species contained potentially toxic secondary metabolites, with manuka known to contain alkaloid and phenolic compounds (Corbett & McDowell 1958), but specific defence chemicals have not been isolated. The differences in physical structure between the browse species would have affected access to young leaves, and species such as gorse are spiny on the mature stems.
The high preference for gorse observed by Lambert et al. (1989a), as opposed to the poor preference in our experiment, was probably the result of the chopping process used by Lambert et al. (1989a) which may have improved the palatability of gorse through, for example, a decrease in its spiny nature, and, perhaps, poor access to young shoots of gorse in our experiment despite the relative abundance of the plants. Thus, the comparison between our outdoor experiment and the indoor experiment of Lambert et al. (1989a), both based on the same browse nurseries, highlighted the need to examine as many of the factors involved in animal preference as possible before conclusions can be drawn on the relative ranking of plant species.
Lambert et al. (1989a) concluded that goats had a greater preference for short spine gorse, manuka, and ceanothus than sheep, and that sheep had a greater preference than goats for broom and pampas. In our experiment goats also had a higher preference for short spine gorse than sheep, whereas sheep had a greater preference for black locust than goats. There were no significant differences in the preference by goats and sheep for individual browse species outside the three most highly preferred species, tree medic, short spine gorse, and black locust, but some of the trends were similar to the results of Lambert et al. (1989a). For example, sheep tended to show a greater preference for broom than goats, and both sheep and goats exhibited a low preference for tauhinu, toetoe, and pampas (cf. Lambert et al. 1989a).
One explanation for the greater browse preference differences between goats and sheep reported by Lambert et al. (1989a), compared with those reported here, is that Lambert et al. (1989a) always imposed choices between pairs of feeds, whereas in the present experiment all 11 browse species and types were always available. In the outdoor experiment animals were observed to mainly browse the most preferred species, thereby making it difficult to collect sufficient data on the least preferred species to detect possible preference differences between goats and sheep for these browse species. Lacher et al. (1982) also reported that when preferred browse species were available, less preferred species were either rejected or rarely consumed regardless of their abundance. Whether browsed by goats or sheep, a mixture of the browse species examined would be difficult to maintain if it contained highly and poorly preferred browse species. Low preference browse species would seem unlikely to be browsed even if more abundant than the highly preferred species.
R. S. Pande thanks Food and Agriculture Organisation and New Zealand Ministry of External Relations and Trade for financial support. The support of AgResearch in allowing the use of browse nurseries and experimental animals at Ballantrae Hill Station is gratefully acknowledged.
• Bryant, J. P.; Provenza, F. D.; Pastor, J.; Reichardt, P. B.; Clausen, T. P.; du Toit, J. T. 1991: Interactions between woody plants and browsing mammals mediated by secondary metabolites. Annual Review of Ecology and Systematics 22: 431–446.
• Corbett, R. E.; McDowell, M. A. 1958: Extractives from the New Zealand Myrataceae: III. Journal of the Chemical Society 168: 3715– 3716.
• Crawley, M. J. 1983: Herbivory: the dynamics of animal plant interactions. Oxford, Blackwell Scientific.
• Devendra, C.; Burns, M. 1983: Goat production in the tropics. UK, CABI. Pp. 90–176. 102 New Zealand Journal of Agricultural Research, 2002, Vol. 45
• Hodgson, J. 1982: Ingestive behaviour. In: Leaver, J. D. ed. Herbage intake handbook. UK, British Grassland Society. Pp. 113–138.
• Hodgson, J. 1986: Grazing behaviour and herbage intake. In: Frame, J. ed. Grazing. British Grassland Society Occasional Symposium 19: 51–65.
• Horn, H. S. 1966: Measurement of “overlap” in comparative ecological studies. The American Naturalist 100: 419–424.
• Jamieson, W. S.; Hodgson, J. 1979: The effects of daily herbage allowance and sward characteristics upon the ingestive behaviour and herbage intake of calves under strip-grazing management. Grass and Forage Science 34: 261–271.
• Lacher, T. E.; Willing, M. R.; Mares, M. A. 1982: Food preference as a function of resource abundance with multiple prey types: an experimental analysis of optimal foraging theory. The American Naturalist 120: 297–314.
• Lambert, M. G.; Jung, G. A.; Fletcher, R. H.; Budding, P. J.; Costall, D. A. 1989a: Forage shrubs in North Island hill country. 2. Sheep and goat preferences. New Zealand Journal of Agricultural Research 32: 485–490.
• Lambert, M. G.; Jung, G. A.; Harpster, H. W.; Budding, P. J.; Wewala, G. S. 1989b: Forage shrubs in North Island hill country. 3. Forage digestibility. New Zealand Journal of Agricultural Research 32: 491–497.
• Pande, R. S. 1990: Feed value of tagasaste (Chamaecytisus palmensis) for goats and preferential browsing activities by goats and sheep in multi-species shrub/pasture conditions. Unpublished MAgSc thesis, Massey University, Palmerston North, New Zealand.
• SAS 1988: SAS user’s guide. Cary, NC, SAS Institute.
• Van Soest, P. J. 1982: Nutrition ecology of the ruminants. Cornallis, UK, O & B Books.
• Whittaker, R. H. 1975: Communities and ecosystems. NY, MacMillan.