Livestock Research for Rural Development 24 (5) 2012 Guide for preparation of papers LRRD Newsletter

Citation of this paper

Greenbriar suppression with goat mob browsing

L L Boggs, J P Muir* and J W Dunn

Southwestern Oklahoma State University, Weatherford, OK USA
* Texas AgriLife Research, 1229 North U.S. Highway 281, Stephenville, TX USA


Little information is available on the suppression of greenbriar (Smilax spp.), an invasive native vine of North America, by goats.  Goats will readily browse greenbriar but little is known about how permanent this vegetation removal is and if it positively affects the subsequent development of the herbaceous canopy.  When land managers are unwilling to spray herbicides or unable to afford chemical or mechanical means of controlling invasive plants, they may be willing to hire goats to achieve the same goal. Wooded paddocks in north-central Texas were split into mechanically cut (to allow goat access to all greenbriar leaves) and uncut (goats given access only up to 2 m).  


Days to 95% leaf removal of 12 kids/ha required an average 48 days while paddocks with 24 and 36 kids/ha averaged 30 days (P<0.05).  Long term suppression of greenbriar did not vary by stocking rate or mechanical cut (P>0.05).  However, herbaceous grass percentage of herbage mass did increase (P< 0.05) in all browsed paddocks compared to unbrowsed areas. Our results indicate that complete removal of greenbriar leaves can be achieved within a short time but destruction of plants may require long term browsing or use of herbicides on regrowth following browsing.  Short duration mob browsing may provide flock owners a source of income in exchange for ridding land of unwanted greenbriar herbage, but suppression will only be temporary since plants are not destroyed during such short periods of herbage removal.

Keywords: biological control, mechanical control, Smilax


Kiely et al (2004) estimated that in 2000, U.S. agriculture producers applied $5 million in herbicides to kill unwanted plants (Table 1). Not only is this a costly venture, it can also lead to weed resistance after repeated applications and exposure of non-target populations to harmful chemicals.  Alternative biological methods for weed control include the use of ungulates in non-crop situations such as forested areas.

Table 1. User expenditures on pesticides in the U.S.A. during 2000 by pesticide type and market sector (Kiely et al 2004)

Market Sector
















Mil $



Mil $



Mil $



Mil $



Mil $






























































Using goats for brush control is not a new strategy.  As early as the 1950’s and 1960’s goats were being recognized as effective and economical in control of brush and small hardwoods in pasture areas (Magee 1957; Wilson 1969).   In addition, Luginbuhl et al (1999) indicated goats are highly effective and economical biological control agents.  Studies conducted over four growing seasons showed controlled defoliation by goats increased desirable vegetative cover in experimental plots from 65 to 85% while at the same time multi-flora rose bushes (Rosa multiflora Thumb.) were practically eliminated.


Browsing habits of goats have produced similar results with other noxious weedy plants.  Toranno et al (1999) indicated 75-80% thistle control was achieved using goats at lower than maximum desired stocking rates.  The recommended stocking rate of goats is 1 goat/ha for every 3% increase in thistle population to achieve desired maximum thistle control.  Only 17 goats were used in the 7 ha study done by Toranno et al (1999) when the recommended maximum stocking rate was 42 goats.


In Brazil, goats select browse and forbs over grass during both wet and dry seasons (Pfister and Malechek 1986).  Because of this preference, they have been successfully used to turn Brazilian caatinga, where woody vegetation dominates, into grasslands where cattle thrived Pfister and Malechek 1986).  In the U.S.A., Luginbuhl et al (1996) reported that goat browse studies conducted in hardwood forests of Arkansas showed 65% of goat diet from July to December consisted of vines, and the vines remained an important part of the goat diet year round. 


Several studies conducted in Texas have indicated that stocking rate appears to have both positive and negative effects on weed and brush control.  Mellado et al (2003) reported that heavily stocking pastures resulted in negative effects on both the land and the goats browsing it.  Goats were forced, due to overgrazing, to alter their diet selection and consume more resinous, toxic and coarse species of plants resulting in negative daily weight gains and lower fertility.  In addition, shrub and grass cover in overgrazed pastures was reduced.  Owens et al (1991) determined that stocking rates of 0.8 goats/ha appeared optimum for brush control experiments in Texas.


Very little information is available on the control of greenbriar, which is considered to be one the most troublesome weeds throughout much of northeastern and southeastern North America (USDA, 2011; Figure 1). Goats will readily browse greenbriar (Packard et al 2007) but suppression and regeneration dynamics are poorly understood. This study was initiated to evaluate the control of greenbriar using various goat stocking rates.

Figure 1: Greenbriar (Smilax sp.) is one of the most troublesome plants goats can help control in North America.
Figure 2: Artificially low browse lines were created by cutting greenbriar at ground level in one half of each paddock.  Goats were allowed to establish their own browse lines in the other half of the paddock.  Browse lines were only created in May and June of the first year of grazing.

Materials and Methods

Research was conducted on eight 0.4 ha paddocks at Stephenville TX (N 32 15’, W 98 12’, altitude 395 m) on Windthorst fine sandy loam (pH 6.6, 11 mg P kg-1, 196 mg K kg-1, 902 mg Ca kg-1, and 168 mg Mg kg-1 using the TAMU-EDTA extractant method, Hons et al., 1990).  The dominant upper canopy plant species included Smilax spp. (greenbriar), Quercus spp. (oak), Ulmus spp. (elm) and Prosopis spp.  (mesquite).  In addition, several broadleaf herbaceous weed species were present: Solanum spp. (nightshades), Cirsium spp. (thistle), and Tragia spp. (noseburn).  Where the canopy was sufficiently open, Panicum spp., Cynodon dactylon, and Poa arachnifera were present.  Native Texas legumes including Indigofera miniata (scarlet pea), Lespedeza spp., Acacia angustissima var. hirta (prairie acacia) and Rhynchosia minima were sparsely scattered throughout.


Each paddock was divided in two.  In half of each paddock, brush was left for goats to establish their own browse line; goats in these areas could only access greenbriar leaves up to 2 m, the approximate browse line. In the other half, brush (but not trees) was mechanically cut to 0.5 m using chainsaws, hedge trimmers and pruning instruments to guarantee goats access to all greenbriar leaves (Figure 2).  Initial greenbriar populations were sampled in May and June 2005 to establish a baseline vegetation survey in the mechanically cut and uncut halves of each paddock.  Individual plant counts and foliage dry matter (DM) were quantified on both cut and uncut sides of the paddocks in 10 random locations on each side.  Stakes were set in the center of each sample location and a 0.5 m string was attached.  Plant counts were taken from the circle around the stake within this radius up to a height of 2 m to include any browse goats might consume.  In the second and third years of the study, 10 locations within each half paddock were randomly sampled in the same manner.  Plants collected were separated by species, leaves removed from stems and dried for 48 hours at 55o C.  Dry matter weights were taken to determine herbage mass available per individual plant of each species in each half-paddock.  Only leaf samples were used to determine dry weight and nutritional value as the goats generally only consume the leaves.            


Initial plant composition was widely variable among paddocks.  While all eight paddocks contained the same basic group of plants (greenbriar, grass, hardwood trees and broadleaf species, mainly nightshade, thistle and noseburn), they were found in differing numbers in the individual paddocks.  For example, in initial plant sampling, Paddock 9 had nine distinct plant species identified in the browse zone.  Two of these samples showed different masses, 225 kg/ha of greenbriar but only 56  kg/ha of grass.    In contrast, Paddock 4 contained only four distinct species of plants in which greenbriar comprised 103  kg/ha and grasses 170 kg/ha. 


Boer x Spanish kid goats weighing approximately 20-25 kg were purchased from local markets and private producers.  All goats were weighed, wormed and then separated into groups to be placed into the paddocks.  Goat stocking rates of 0 (control), 12 (low stocking rate), 24 (medium stocking rate) and 36 goats/ha (high stocking rate) were established in two paddocks each for a total of eight paddocks (Figure 3).  Once goats were turned onto paddocks, they only received multi-mineral block supplement and water.  All goats were removed from their respective paddocks when the browse was visually determined to be 95-100% cleared (Figure 4).  

Long-term rainfall average from May to October is 452 mm at the trial site.  However, in 2005 it was only 61% (275 mm), in 2006 it rose to 68% (308 mm) and in 2007 it was 137% (623) of this long-term average during the same 6-mo growing season.


Analysis of variance (ANOVA) was run on greenbriar percentage of herbage and grass percentage of total herbage up to a 2-m height for 2005, 2006 and 2007.  Stocking rate and mechanical cutting were considered independent variables and the interaction of these two factors were tested and simple effects submitted to ANOVA only when this interaction was found to be not significant at P<0.05.  Where appropriate, individual mean separation using least square differences (P≤0.05) were calculated. Years were not considered a factor in the model since effects of browse, which were imposed on the same paddock every year, would be cumulative.


Figure 3:  Boer by Spanish kids were stocked at the rates of 0, 12, 24, and 36 goats/ha in 0.4 ha paddocks.  Paddocks 5 and 10 were control paddocks with 0 goats per acre.  Paddock 1 was the home paddock and paddock 6 was not used.



Figure 4:  Goats were removed from their respective paddock when plant material was 95% to 100% cleared. 




Goat stocking rate affected the number of days it took goats to clear the paddocks (Figure 5).  While goats stocked at 24 and 36 goats/ha had little difference in days to clear, they took less time (P=0.005) than those stocked at 12 goat/ha. Goats stocked at 12 goats/ha took an average of 48 days to clear their 0.4 ha paddock, while those stocked at 24 and 36 goats/ha required only 30 days, on average, to clear the same area (Figures 6 and 7). Differences (P=0.009) were also noted between the number of goats per hectare and the year of the study as well.  In 2006, when 308 mm rainfall fell from May to October, the trial area experienced a severe drought resulting in a shorter number of days, regardless of stocking rates, to clear brush compared to 2005 (when plants had accumulated reserves from previous years) or 2007 when 623 mm rainfall fell from May to October compared to the 452 mm long term rainfall.

Figure 5:  As number of goats per paddock increased, the number of days to 95% greenbriar leaf removal declined (P=0.005; LSD 7.1). 

Figure 6:  Twelve goats/ha (left) took 46 days to clear a 0.4 ha paddock.  Twenty-four goats per hectare (right) took just 30 days to clear the same area.

Figure 7:  Thirty-six goats/ha effectively cleared the paddock on the left in an average of 28 days.  The paddock on the right is a control paddock which contained no goats.


Greenbriar  as a percent of total herbage, did not show a decline over the 3-yr study period (P>0.05) (Figure 8).  Control paddocks that were not browsed did show an increase in greenbriar percentage (P<0.05), however.   Percent greenbriar tended to increase in 2006, likely a reflection of decreased contribution of the herbaceous component as a result of low rainfall (308 mm) from May to October compared to 623 mm during the same period in 2007.

Figure 8:  Percent of herbage mass up to 2 m height that greenbriar (Smilax spp.) contributed per paddock did not show a decline over the 3-yr study period (P>0.05).  Greenbriar percentage in paddocks that were not browsed increased (P<0.05; LSD 16.5).

Removing browse should allow more sunlight and growing space for herbaceous species.  However, no increase was observed over the study period (P>0.05; Figure 9). Control paddocks did show a decrease (P<0.05) in percent grass over the three year study period relative to paddocks browsed by goats.  There was a tendency for greenbriar percentage of herbage mass to increase relative to herbaceous species in 2006, a dry year, likely because deep roots and nutrient storage in the roots and vines gave the perennial greenbriar an advantage vis--vis the more shallow-rooted herbaceous grasses. When rainfall increased in 2007, this trend was reversed.


The cutting treatment produced no effect (P>0.05) in the decrease of leaf regrowth potential during the three years of the study.  In fact, by the end of the study many plants that had been cut to allow full access of leaf growth to browsing goats had grown vines above 2 m and beyond the reach of the animals in subsequent years. This indicates that short-duration exposure to browsing, even when all leaves are within browsing range, is insufficient to permanently suppress greenbriar.

Figure 9:  Percent grass in the paddocks did not show a decrease (P>0.05) over the 3-yr study period when browsed by goats at three stocking rates.  Control paddocks which were not browsed showed a decrease (P<0.05; LSD 15.2) in percent grass



Hons F M, Larson-Vollmer L A and Locke M A 1990 NH4OAc-EDTA-extractable phosphorus as a soil test procedure.  Soil Science 149:249-256.


Kiely T, Donaldson D and Grube A 2004 Pesticides Industry Sales and Usage: 2000 and 2001 Market Estimates. EPA's Biological and Economic Analysis Division, Office of Pesticide Programs, and Office of Prevention, Pesticides, and Toxic Substances.


Luginbuhl J-M, Harvey T E,  Green Jr. J T, Poore M H and Mueller J P 1999 Use of goats as biological control agents for the renovation of pastures in the Appalachian region of the United States.  Agroforestry Systems 44:241-252.


Luginbuhl J-M, Green J T, Mueller J P and Poore M H 1996 Meat goats in Land and forage Management. In: Proceedings of the Southeast Regional Meat Goat Production Symposium “Meat goat production in the Southeast-Today and Tomorrow.” Florida A &M University, Talahassee, Florida.


Magee A C 1957 Goats Pay for Clearing Grand Prairie Rangelands. Texas Agricultural Experiment Station Miscellaneous Publication 206. College Station, Texas.


Mellado M, Valdez R, Lara L M and Lopez R 2003 Stocking rate effects on goats:  a research observation.  Journal of Range Management 56:167-173.


Owens M K, Mackley J W and Carrol C J 1991 Foraging Behavior of Goats.  Risks in Ranching PR-4873. Texas Agricultural Experiment Station, Uvalde, Texas.


Packard C E, Muir J P, Wittie, R D and Harp R M 2007 Peanut stover and bermudagrass hay for wethers on summer hardwood range in north central Texas. Sheep & Goat Research Journal 22:7-14.


Pfister J A and Malechek J C 1986 Dietary Selection by goats and Sheep in a Deciduous Woodland of Northeastern Brazil.  Journal of Range Management 39:24-28.


Torranno L,  Holst P J and Stanley D F 1999 The effect of herbicide and goats on survival and seed production of Illyrian thistle (Onopordum illyricum L.) Plant Protection Quarterly 14:13-15.


USDA 2011 Plants profile: Smilax rotundifolia roundleaf greenbriar. United States Department of Agriculture Natural Resource Conservation Service, Washington D.C. USA.


Wilson A D 1969 A Review of Browse in the Nutrition of Grazing Animals. Journal of Range Management 22:23-28.

Received 8 February 2012; Accepted 26 March 2012; Published 7 May 2012

Go to top