Grazing Behaviour of the Icelandic Horse
Anna Gudrun Thorhallsdottir, Gudni Agustsson, and Johann Magnusson
Agricultural University Hvanneyri, 311 Borgarnes, Iceland. 
Voice 354- 4370000       Fax 354-4370048 
annagudrun@hvanneyri.is

Introduction

            For the last decades, there has been a growing interest for horses in Iceland and horse numbers have increased considerably.  With increasing numbers, there has also been increased interest and need for information on the grazing behaviour of the Icelandic horse, as grazing problems is more often seen.   For the last 10 years, several studies have been conducted on the grazing behaviour of the Icelandic horse at the Agricultural University at Hvanneyri in Iceland.  The aim of these studies has both been to describe diet selection and grazing behaviour of free roaming horses and to analyse factors affecting the grazing behaviour in a more controlled experimental set up.  The questions that we have been asking are how long time do horses spend grazing each day, when do they graze, what do they graze and how do factors like time of day, season, and quality and quantity of the pasture affect their grazing behaviour, measured at movement between feeding stations (Charnov, 1976; McNair, 1982; Stephens and Krebs, 1986; Jiang and Hudson, 1993), where feeding stations are defined as the area an ungulate can feed without moving its forelegs (Novellie, 1978).  The studies that are described in this paper were conducted on free roaming horses in north of Iceland in 1992 and on horses grazing different pastures in west of Iceland in 1994 and 1995. 

Methods

Study 1.  The first study was conducted at Kolkuos in Skagafjordur, N-Iceland in late summer and early fall in 1992.  The grazing behaviour of 72 horses, mares with foals and sub-adults, 1-4 years of age, in an 520 ha area of diverse plant communities was recorded four times, for 48 hours each time.  The diverse plat communities consisted of 37% dry heath, 14% barren land, 12% grassland, 11% natural mire, 8% drained mire, 8% semi-bogs, 8% sec. succession, and 2% sand.  The first recording was in early august and the last was in the end of September.  Recordings on in what plant community each horse was found were made every two hours. while daylight lasted and whether it was grazing, resting, or walking.  The number of horses grazing or resting t in each plant community at each time was used to estimate activities within the day, plant community preferences and forage ratio (Putman et al 1987).  

Study 2.  The second study was conducted at Hestur and Stora-Drageyri, W-Iceland in the summers of 1994 and 1995.  In each location, a 16 hectare area was divided into three enclosures, on 7.1 hectare (light grazing pressure), 5.0 hectare (medium grazing pressure) and 3.8 hectare (heavy grazing pressure).  In each enclosure, 5 adult horses grazed from middle of June to middle of September. The vegetation type at Stora-Drageyri consisted mainly of a dry grassland, previously cultivated that had not received any form of fertilizer for over 20 years.  Number of species was rather limited, around 10 species, dominated by grasses such as Agrostis sp., Poa sp. and Festuca sp.  The dominating vegetation type at Hestur was species rich drained mire, with over 20 common species, including different Carex sp. and Eriophorum sp., grasses like Calamagrostis sp. and Deschampsia sp., and different dicots.  Detailed measurements were made on the grazing behaviour of each individual horse, every other hour for three days, for three periods, in July, August and September. In July, measurements were taken for the full 24 hours, and while daylight lasted in August and September.  Each horse was observed for 4 min each time, or a total of 48 min every 24 hours and its behaviour recorded.   Special attention was made on grazing behaviour and movements between grazing patches and feeding stations.  Measurements were recorded with help of the software Observer 2.0 and analysed as a split plot design, with repeated measures on the horses.

Results

Study 1.  In August, on average 54% of the horses were found grazing, while the proportion was up to 80% in late September.  In July, shorter rest periods was seen between grazing activities while in late September grazing activities were seen in longer periods, the shorter rest periods were no longer present, but longer rest periods between the longer grazing periods.  On average, 62% of the horses were observed grazing.

Only 39% of the 520 hectare were used by the horses for grazing – the grassland, the natural mire, the drained mire and semi-bog.  The natural mire was used more for grazing earlier in the period, while the drier communities, the drained mire and the grassland were used progressively more later in the fall. The grassland had the highest forage ratio (density of horses grazing at each observation) in all the periods.  The drained mire came second in August and the semi-bog in September.  The dry heath was mostly used for resting in all periods.

Study 2.  There were significant differences found between the grazing behaviour of the horses, measured as time spent on each feeding station, at the two locations Hestur and Stora-Drageyri in both years (1994: p<0.001; 1995: p=0.006).  The horses spent longer time at each feeding station at Hestur than at Stora-Drageyri.  At both locations, the horses in the enclosure with the heaviest grazing pressure (3.8 ha) spent the longest time at each feeding station, before moving to the next and the horses in the enclosure with the lightest grazing pressure the shortest time at each feeding station (1994: p=0.0164; 1995: p<0.001).  

The time at each feeding station was longer later in the summer in both years (1994 and 1995: p<0.001).  The time was longer in August than in July in both locations, and the longest in Stora-Drageyri in September in both years, but about the same in Hestur in August and September in both years (location*month 1994: p=0.0768; 1995: p=0.0305).  

Discussion

Results from Kolkuos in N-Iceland showed that 62% of the horses were on average observed grazing, ranging from 54% in early August to 80% in late September Other studies by Thorhallsdottir (unpublished data), using vibrarecorders have shown that the Icelandic horses spend 60-80% of their time grazing, the longer the poorer the pasture.  These results are in good agreement with observations by studies by Jonsdottir (1988) on Icelandic horses and Duncan (1987) on free roaming horses in Camargue in France. As the forage value of the plants drops towards the fall (Thorsteinsson and Olafsson, 1969), the longer it takes to acquire enough nutrition from grazing.  To meet nutritional requirements, the horses thus need to graze longer later in the season. 

The horses showed marked preferences for certain plant communities for grazing and these preferences changed as the season proceeded.  Grasses and sedges (Carex sp. and Eriophorum sp.) were heavily selected for and the dry heath, dominated by small scrubs was largely ignored for grazing.  Wetter areas, like the natural mire was grazed much less than the much drier one, the drained mire. The wetter areas were also grazed  earlier than later in the season.  According to Thorsteinsson and Olafsson (1969), that measured the nutritional value of different plant species at different times and locations, the nutritional value of sedges drops faster with the season than that of grasses.  Similar grazing pattern by horses were seen in the study of Jonsdottir (1988).  She also reported resting to take place mainly in a dry heath community.

According to the marginal value theorem, a forager should leave a patch and move to the next its rate of energy gain falls below that of the general environment (Charnov, 1976).  Time spent at each feeding station has been used as a measure of the quality of the patches, predicting longer giving-up times for better patches (McNair, 1982).   However, as Redhead and Tyler (1988) pointed out, the main variable to look at is the rate of depletion of a patch.  If it takes longer to reach the average capture rate of a poor environment than that of a rich environment, the animals should remain longer in a bad patch of a poor environment than in a good patch of a rich environment (Redhead and Tyler, 1988).

The time spent on each feeding station was longer at Hestur than at Stora-Drageyri and the time was longest in the most heavily grazed enclosures and shortest in the lightest grazed enclosure.  In the heaviest grazed enclosure, the environment was indeed poor, and little to gain by moving to a new feeding station.  Moving is energy requiring process and the gain by moving has to be more than the loss by moving.  It was an interesting result that the time at each feeding station was significantly longer at Hestur than at Stora-Drageyri.  The pasture at Hestur had much more species diversity, with different types of sedges and forbs, while the pasture at Stora-Drageyri was dominated by different grass species. As seen in study 1, horses prefer to graze areas that have a dominance of grasses rather than sedges.  The overall more species diversity found at Hestur does not seem to have overridden the importance of more preferred plants at Stora-Drageyri, and that the horses considered the pasture at Hestur of overall lower quality than that of Stora-Drageyri.  The same results were obtained by Novellie (1978), working with ungulates in South Africa, where time at each feeding station got longer as the overall nutritional quality of the forage got lower and the animals were not able to select as much as earlier. 

Time spent at each feeding station became longer as the season proceeded, and was longest in Stora-Drageyri in September.  At Hestur the time was longer in August than July, but about the same in August and September. Overall time was, though, always longer at Hestur.  This is in good agreement with the understanding that time spent at each feeding station is a measurement of the overall quality of the pasture, where abundance of preferred plants leads to short time at each feeding station and little of preferred plants leads to long time.  In the fall, more or less all the plants are of equal poor quality and little to gain by moving.  The fact that the time did not increase from August to September at Hestur might be understood as the maximum time had already been reached in August, because of the overall poorer pasture.  It is well documented that sedges loose their nutritional value earlier in the season than grasses (Thorsteinsson and Olafsson 1969), supporting that the pasture at Hestur was indeed a poorer pasture.

Acknowledgements

We thank Lena Fernlund, Asdis Ingimarsdottir and two Danish students for assistance with the fieldwork, Susan Durham for help with data analysis and The Icelandic Agricultural Production Fund (Framleidnisjodur landbunadarins) for financial assistance. 

References

Charnov EL (1976) Optimal foraging: The marginal value theorem.  Theor Pop Biol 9: 129-136.

Duncan P (1987) Horses and Grasses.  The Nutritional Ecology of Equids and Their Impact on the Camargue.  Springer – Verlag.  New York. 

Jiang Z, Hudson RJ (1993) Optimal grazing of wapiti (Cervus elaphus) on grassland: patch and feeding station departure rules.  Evol  Ecol  7: 488-498.

Jonsdottir S (1988) Some features of behaviour, selective grazing and interaction of sheep and horses.  Dissertation no. 188189.  University of Wales.  78p.

McNair JN (1982) Optimal giving up time and the marginal value theorem. Am Nat 119: 511-529.

Novellie PA (1978) Comparison of the foraging strategies of blesbok and springbok on the Transvaal highveld.  S  African Journal of Wildlife Research  8: 37-144.

Redhead E and Tyler PA (1988) An experimental analysis of the optimal foraging behaviour in patchy environments.  Quart  J  Exper Psychol 40B: 93-102.

Stephens DW and Krebs JR (1986) Foraging theory. Princeton, NJ: Princeton University Press. 

Thorsteinsson I  and Olafsson G (1969)  Efnainnihald og meltanleiki nokkurra uthagaplantna II: rannsoknir a islenskum beitilöndum (Nutritional value and digestibility of range plants – studies on Icelandic rangelands).  Isl. Landbunandarrannsoknir (J Agr Res). 1:45-63. (in Icelandic).