Great Basin Naturalist
Volume 58 Number 1 Article 3
1-30-1998
Bitterbrush (Purshia tridentata Pursh) growth in relation to browsing
Carl L. Wambolt Montana State University, Bozeman, Montana
W. Wyatt Fraas Montana State University, Bozeman, Montana
Michael R. Frisina Montana Department of Fish, Widlife, and Parks, Butte, Montana
Follow this and additional works at: https://scholarsarchive.byu.edu/gbn
Recommended Citation Wambolt, Carl L.; Fraas, W. Wyatt; and Frisina, Michael R. (1998) "Bitterbrush (Purshia tridentata Pursh) growth in relation to browsing," Great Basin Naturalist: Vol. 58 : No. 1 , Article 3. Available at: https://scholarsarchive.byu.edu/gbn/vol58/iss1/3
This Article is brought to you for free and open access by the Western North American Naturalist Publications at BYU ScholarsArchive. It has been accepted for inclusion in Great Basin Naturalist by an authorized editor of BYU ScholarsArchive. For more information, please contact [email protected], [email protected]. Great Basin Naturalist 58(1), © 1998, pp. 28-37
BITfERBRUSH (PURSHIA TRIDENTATA PURSH) GROWTH IN RELATION TO BROWSING
Carl L. Wamboltl , W. Wyatt Fraasl , and Michael R. Frisina2
ABSTRAG'1:-The objectives of this study were to compare vegetative and reproductive growth characters of bitter brush (Purshia tridentata Pursh) stands as they relate to browsing levels. Growth characters were measured on 10 eco logically diverse stands in southwestern Montana on which browsing ranged from 0% to 60% of all current annual long shoot (LS) growth. Bitterbrush plants exhibited both twig-level and plant-level responses to browsing. Total bud density per plant was similar for browsed and unbrowsed sites, but differed (P < 0.01) between browsed and unbrowsed twigs. Browsed twigs produced one-half the leaf cluster density produced by unbrowsed twigs. No significant (P < 0.05) rela tionship between browsing levels on browsed plants and bud densities was found. Length ofold-growth twigs per plant was shorter (P < 0.001) on browsed sites than on unbrowsed sites. Burning at 2 environmentally paired sites reduced flower bud density (P < 0.001) 9 and 10 growing seasons later although LS length was not affected. Growth of LS showed a site-by-year intef'd.ction (P < 0.05). Our data suggest that long-term unbrowsed plants should not be used as a standard for comparison with normally browsed plants.
Key words: Purshia tridentata, bitterbrush, browsing, shoot growth, bud development, Montana.
Antelope bitterbrush (Purshia /ridentata Guenther (1989) studied the environmental Pursh), well documented as a valuable food relationships ofbitterbrush stands on Montana source for big-game animals (Kufeld 1973, Fish, Wildlife, and Parks' Mount Haggin Wild Kufeld et a!. 1973), is highly palatable, moder life Management Area (MHWMA) in south ately nutritious, and common on many big western Montana and noted the wide range of game winter ranges (Giunta et a!. 1978), al habitats and stand growth, Guenther (1989) though it seems to be declining in some areas also found a high level of browsing on bitter (Winward and Findley 1983), Bitterbrush is brush plants and little successful reproduction found in a wide range ofhabitats (Franklin and during the previous decade, Wambolt et al, Dyrness 1973) and is useful as a ground stabi (1996) compared some of the same MHWMA lizer on exposed soils (Nord 1959), Therefore, sites with 5 other southwestern Montana loca land managers are interested in its propaga tions and found differences in crude protein tion, growth, and management to improve de content by site. graded wildlife habitat. The specific objectives ofthis study were to Known for its variability in habitat, mor~ compare vegetative and reproductive growth phology, and physiology, bitterbrush ranges cbaracters of 10 bitterbrush stands on and from prostrate forms only 10 em high to colum near the MHWMA and to relate them to nar forms over 3 m tall (Winward and Findley browsing levels, We hypothesized that plants 1983), Color, shape, and size of leaves, stems, from nearby bitterbrush stands are not uni and seeds vary between and within popula form in their growth characteristics. tions (Alderfer 1977), Mowing and burning result in responses that range from death to METHODS vigorous sprouting (Clark et a!. 1982), While Study Sites these adaptations enable bitterbrush to inhabit widely divergent habitats in western North We chose 10 study sites primarily to repre America, they can also make management of sent bitterbrush stands from a range of envi the species more difficult unless the response ronmental conditions (Table 1), Included were a oOocal populations is known, burned site and sites protected from browsing.
lOepartment ofAnimal and Runge Science$. Montana State University, Bozeman. MT 59717. 2Montaml Department ofFish, Wildlife, and Parks, Butte, MT59701.
28 1998] BITTERBRL'SH GROWTH 29
All sites are located within a radius of 14.5 km 'fAm,E 1. Topographic characteristics ofthe 10 study sites. near Butte and Anaconda in southwestern Data from the last 4 sites were obtained from Guenther (1989). Montana. Long-term climatic records were available for the general study area from the Elevation Slope Aspect Site (m) (%) (') Anaconda weather station at 1700 ill elevation. Annual precipitation at Anaconda averages Buttc 1730 26 234 340 mm, with 47% received between April Deer exclosure 1830 12 225 Cattle exc10sure 1830 16 188 and July (NOAA 1991). Cattle + Decr 1820 10 190 Vegetation types at all but 3 sites (Burn, Burn 2010 22 200 Unburn, and High Rye) are seral stages of the Unhurn 2010 22 200 bitterbrush-bluebuneh wheatgrass (Agropyron Puwerline 1640 16 85 spicatum Pursh) habitat type (Mueggler and Willow Creek 1780 31 llO Railroad Gulch 1650 32 ll5 Stewart 1980). The dominant shrub is bitter High Rye 1940 38 120 brush, but understory vegetation is regressed primarily from grazing (Fraas et a!. 1992) on the other 7 sites from the described potential climax composition (Youtie et a!. 1988). We selected the Butte site at Maude S Can stock use on both sites resumed 15 September yon, near Butte, Montana, because it receives 1982. no ungulate browsing. The plant community Four sites were located on the MHWMA, consists of bitterbrush, Centaurea rnaculosa owoed and managed by Montana Fish, Wildlife, Lam. (spotted knapweed), Ribes cereum Doug!. and Parks. The Powerline site is on a slope 50 (squaw currant), and Rosa woodsii Lindl. m above a perennial stream on the northeast (Woods rose). edge of the MHWMA big-game winter range. At Dry Cottonwood Creek in the Deerlodge The plant community consists of bitterbrush district of the Deerlodge National Forest, we and spotted knapweed. The Willow Creek site, studied a 2-part exclosure. One portion, known near the top of a grassy ridge 150 m above as the Deer exclosure, was game proof. The Willow Creek, supports a relatively large other half allowed deer use but served as a amount of Elymus cinereus Scribn. & MelT. livestock exclosure and thus was known as the (basin wild rye), along with other perennial Cattle exclosure. Near the exclosure, we studied grasses and bitterbrush. This area was used as a bitterbrush stand kno""" as the Cattle + Deer winter range by mule deer, elk, and moose. site because it sustained both cattle and mule The Railroad Gulch site is also on the deer deer browsing. These 3 sites have a scattered and elk winter range. This site occupies a mid overstory of Pseudotsuga menziesii [Mirb.] slope position 30 m above an intermittent Franco (Douglas-fir). A large number ofnative stream, where the plant community consists of perennial [orbs occur in the understory on bitterbrush and spotted knapweed. The High these sites. Rye site, 1500 m higher in elevation than the To gauge the impacts ofburning bitterbrush other MHWMA sites, appears to receive the in southwestern Montana, we selected 2 sites. greatest snowpack The plant community on These sites (Burn, Unburn) were environmen the High Rye site is typical of the bitter tally paired on both sides of a burn line on the brush-rough fescue (Festuca scahrella Torrey south flank of Steep Mountain, 8 km northwest ex Hook) habitat type (Mueggler and Stewart of Butte, in the Butte district of the Deerlodge 1980), with those species currently dominant. National Forest. The plant community on these Guenther (1989) found the least amount of 2 sites is a bitterbrush-mountain big sagebrush big-game use at this location among the 4 (Artemisia tridentata Nutt. ssp. vaseyana [Rydb.] MHWMA sites. The MHWMA study sites re Beetle)-bluebunch wheatgrass association inter ceived insignificant levels oflivestock grazing. mediate to the big sagebrush-bluebunch wheat Sampling and Analyses grass and bitterbrush-bluebunch wheatgrass habitat types of Mueggler and Stewart (1980). Study sites typical oftheir communities were The prescribed burn was conducted 3 Novem delineated by five 15-m transect lines placed ber 1981 after a year's rest from livestock graz perpendicular to the slope at 3-m intervals, ing on both sites to increase fuel loads. Live- thus comprising a study plot of 15 x 12 m. We 30 GREAT BASIN NATURALIST [Volume 58
recorded topographic information at each site, ered, regardless of availahility to browsers, to determining aspect by taking a compass bear determine plant response to removal ofa per ing from the major slope, measuring slope centage oftotal annual growth to relate to pre with a clinometer. and ascertaining elevation viously recommended use levels (Hormay 1943, from USGS topographic maps. The informa Garrison 1953, Martinsen 1960, Lay 1965, tion from MHWMA sites was taken from Urness and Jensen 1983). Browsing-level analy Guenther (1989). ses were conducted by comparing the number We used the following definitions during of browsed and unbrowsed live LS on each the study: bitterbrush plant-a single stem or plant in the manner detailed by Wambolt group of stems with a single point of origin; (1996). Browsed and unbrowsed twigs on a leaf cluster-a bud which had produced a plant were each pooled across branches for group of leaves and which had not elongated comparison of browsing response on a plant «7 mm in length); long shoot (LS)-a bud level. By combining plant averages we then structure that had elongated (>7 mm in length) created averages for the 10 sites. in the current growing season and consisted of Occurrence of unequal variances for com a stem and attached leaf clusters; flower-a parisons, as experienced by Bilbrough (1990) bud which had produced a flower; flowers with similar data, required use ofnonparamet grew only on l-yr-old or older stems. ric statistical tests (Sokal and Rohlf 1981): a Two bitterbrush plants rooted within 1 m of Wilcoxon signed-rank test (Snedecor and each transect line were randomly selected for Cochran 1989) for comparison of paired mea measurements (10 plants per site). We ran sures (such as the same plants between years), domly chose 4 branches on each plant using a and a Mann-Whitney rank-sum test (Snedecor frame with lO-cm grids placed on top of the and Cochran 1989) for comparison of group plant. Random numbers identified grid inter means, both at P < 0.05. Interactions between sections. The closest live branch to a plumb years, sites, and boeatments were analyzed with line dropped through the grid was sampled. a multi-factor analysis of variance (Snedecar On each sampled branch we recorded the fol and Cochran 1989). Correlation was used to lowing: age and length of each stem segment, measure the relationship between some vari length of LS, number of flowers, leaf clusters, ables without a dependence relationship and LS. Apical bud status of each terminal LS (Snedecor and Cochran 1989). Comparisons segment was recorded as browsed (within the between sites were based on least significant past year), unbrowsed, or dead. Flowers were difference (LSD; Snedecor and Cochran 1989) counted in early July, and leaf clusters and LS at P < 0.05. Least significant differences were were counted and measured in earJy Septem calculated as part of the analysis of variance ber. We compared measurements only from for pairs ofmeans, such as site-to-site or year branch (LS) segments <3 yr old, as little bud to-year comparisons. All statistical tests were activity occurred on older portions of the programs of the MSUSTAT statistical program branches. To determine age, we examined (Lund 1991). annual growth scars after an initial trial of comparing growth scars with growth rings. RESULTS AND DISCUSSION These measurements were summarized across Browsing Effects all 4 branches sampled per plant to create a plant average for each category. Overall aver At the 8 browsed sites the browsing level ages resulted from averaging the 10 plant ranged from 23% to 60% removal of all cur averages for each of the 10 study sites (100 rent annual L8 (Table 2). This range was with plants). in previously recommended levels for long We observed each sampled branch for term health and maintenance of slands (Har browsing use during the previous winter. may 1943, Garrison 1953, Steinhoff 1959, Mar Guenther (1989) found a high correlation (r = tinsen 1960, Lay 1965, Shepherd 1971). Only 0.94, P < 0.0001) in measuring percent bitter 2 sites had less browsing the 2nd winter (P < brush utilization by determining either per 0.05), while the other 6 were browsed at centage of LS browsed or length of LS re nearly the same level both years. During 1990 moved; thus, we determined the percentage of the 8 sites were equally browsed, but in 1991 total LS browsed. All branches were consid- some variation in browsing levels occurred 1998] BITTERBRUSH GROwrn 31
TABLE 2. Browsing level (percent) for 1990 and 1991 at Total hud density per plant, expressed as the study sites, based on number of total long shoots (LS) the sum of the number of flowers, leaf clus removed. ters, and LS per unit length ofstem, was simi Site 1990 1991 lar for browsed and unbrowsed sites (P > Butte O" 1x2 on 0.10; Fig. 1). However, total bud density did Deer exclosure on 0" differ at the twig level (P < 0.01) between Cattle exclosure 45'w 39aw~y browsed and unbrowsed twigs (Fig. 1). Browsed Cattle + Deer 55'w 54'" Burn 48aw 5on, plants had a lower flower bud density (P < Unburn 5law 61 ay 0.001) and higher LS bud density (P < 0.01) Powerline 5law 40awJt}' than unbrowsed plants (Fig. 1). Howevel; at Willow Creek 52aw 23bw the twig level (Fig. 1), flower or LS denSities Railroad Gulch 53'w 37a",)(y High Rye 6O'w 3()bwx were similar between browsed and unbrowsed twigs pooled for all browsed sites. Unbrowsed IRow entries ....ith slmiw lowercase lellers (ab) are not significantly different ~o ten, P < 0.05). twigs from browsed plants had lower flower (P !column entne5 with !imi!:r.I·lo-..ca~ letter's (WX)'Z) IIlre not significantly dif· < 0.01) and higher LS (P < 0.001) bud densi rer~t (LSD, P < 0.(5). ties than twigs from unbrowsed plants. This suggests that browsing affects bnth browsed among sites. Evaluation of browsing effects and unbrowsed twigs on browsed plants, which should consider that post-browsing LS length is a plant-level response. Further, density of represents the sum ofeach year's growth minus any of the 3 types of buds did not appear to the cumulative reduction by brOWSing. In depend on actual level of browsing per plant, addition to the direct effect of removing twig as 0% to 100% ofterminal twigs were browsed material, browsing migh t also affect length by on sampled branches on plants exposed to changing the potential for growth. Growth herbivores, with r = 0.07 (P > 0.22) between potential might be affected by the ability of bud density and percentage browsed. This the whole plant to grow or by the number or suggests that any degree of browsing affects type of buds available, either for the whole flower and LS production on the whole branch plant or for individual twigs. and probably on the whole plant.
0.16,------,------, • o Browsed 0.14 .Unbrowsed
0.12 • .&: "C 0.1 ~ E 0.08 ~ -g 0.06 m 0.04
0.02 • • o-l-= Rowen Leaf LS All Bud. Rowers Lear LS AI BudI 01..... Clu..... Twigs Plants
Fig. 1. Average number of buds per mm of branch (density) by type of bud structure (flowers, leaves, long shoots). Comparisons are between browsed and unbrowsed twigs 00 browsed plants (n = 8), and between plant averages from browsed (n = 7) and unbro\vsed (n = 2) sites. Pairs orbars with similar letters are not different (Mann-Whitney test, P < 0.05). 32 GREAT BASIN NATURALIST [Volume 58
Several researchers have attributed low Although browsing levels (Table 2) were growth rates to whole-plant effects on vigor statistically the same for the Burned and (Hormay 1943, Garrison 1953) or carbohy Unburned sites, flower bud density was lower drate reserves (Menke and Trlica 1983) and on the Burned site (Table 3) than on the have recommended moderate browse levels or Unburned site (P < 0.001). Leaf cluster and specific seasons of use. Tueller and Tower LS densities were similar between the two (1979) reported a lower growth rate in rested sites (P > 0.10), apparently uruillected 9 and or lightly used plants than in those that were 10 growing seasons after the fire. Fraas et al. heavily browsed, terming this a stagnation (1992) had earlier reported that bitterbrush on effect. Bilbrough (1990) found that clipped the Burned site was significantly lower in bitterbrush was able to mobilize inactive buds canopy cover (P > 0.01), flower production (P for elongation and hypothesized that this > 0.1), and seed production (P > 0.1) than on would eventually alter flower and LS ratios. the Unburned site. Because these 2 sites were Although we could detect differences in bud adjacent and environmentally the same (Table density (buds per unit length ofstem) between 1), including their management before and sites and treatments and could construct bud after the burn treatment, it is logical to assume frequencies from this information, we could that flower bud density was lowered by the not determine whether changes in frequency fire just as were the characteristics reported of flowers, leaf clusters, or LS were due to by Fraas et al. (1992). We could not find addi variable densities before browsing or to bud tional burned sites to include in OUf investiga differentiation after browsing. tion. Therefore, we are uncertain whether Leaf cluster bud density was 49% lower on similar results would be the rule, but our find browsed twigs (P < 0.05) than on unbrowsed ings indicate that a reduction in flower buds twigs (Fig. 1). This decrease did not appear should be anticipated. between browsed and unbrowsed plants or The Dry Cottonwood Cattle exclosure site between unbrowsed twigs from browsed plants had lower total bud densities than the un and twigs from unbrowsed plants, suggesting browsed Deer exclosure site (P < 0.01), where that this leaf bud response occurred only on as flower bud densities (Table 3) were lower browsed twigs. Possible mechanisms for this (P < 0.001) and LS bud densities were higher decline include increased mortality ofleafbuds (P < 0.001) in 1990, as were most other browsed either by physiological effects or by higher to unbrowsed comparisons. Although browse leafbud density at the distal (browsed) end of levels (Table 2) were not significantly different the twig. Physiological effects could include (P < 0.10) between the Dry Cottonwood Cat physical or chemical damage due to browsing tle exclosure and Cattle + Deer sites, in 1990 or a change in resource allocation patterns with the Cattle + Deer site had twice as many LS in the plant to maintain flower and LS bud buds per unit length of stem (Table 3). Other numbers at the expense ofleafbud numbers. bud densities did not differ (P < 0.05) between
TABLE 3. Bud density (buds per 100 mm stem) for flowers, leafclusters, and long shoots (LS) on all study sites in 1990 and 1991. The 2 unbrowsed study sites are at left. Bud Deer Cattle Cattle Power Willow RR High type Year Butte excl excl + Deer Bum Unburn Line Creek Gulch Rye Flower t990 7.5flv1 6.gcly OAay 0.7aby O.py 2.gb"" 1.6abcy 3.5cdy 2.2abcy 5.0dey t991 2.7cz 2.5cz O.Oay 0.7aby O.l:l.by 1.1bz O.laby O.3abz OAabz O.pbz
Leaves 1990 1O.7cy 7.3ay 8.1abcy 7.9aby 1O.3bcy 9.7abcy 8.9abcy 9.5abcy 1O.()bcy 10.5bcy 1991 6.3ahez 8.goy 6.7ahcy 5.6abz 4.1az 6.2abcz 8.2boy 3.5az 8.6bcy 4.5az
LS 1990 1.7aby 1.Iay 3.2cdy 6.1ey 4.6dy 3.2cdy 3.4cdy 3.1bcy 2.1abcy 3.0boy t991 1.0ay l.Oay 2.4ahy 3.Sbedz 2.3abz 2.6abcy 4.7dey 3.7edy 3.31x:z 6.oez
'Row entries with similar letters (abOOef) are not $ignificantly different (LSD, f > 0.05). 2Site entrie~ with similar letters (yz) for year pairs are not significantly different (Wilcoxon test. f > 0.05). 1998] BITIERBRUSH GROWTH 33 the 2 sites. Few differences for any type ofbuds ever, because these differences were not always were found among browsed unburned sites or at the same sites, the correlation between between the 2 unbrowsed sites (Table 3). number and average LS length was not signif icant (r = -.12, P > 0.60). As discussed ear Growth lier (Fig. 1), LS bud density was highest on We measured old-growth branch length (3-, browsed plants (P < 0.01). However, LS bud 2-, and l-yr-old segments), LS growth (annual numbers (Fig. 3) generally did not differ (P < growth), and leaf weights (leaf clusters). Total 0.10) between browsed and unbrowsed plants branch length of old-growth twigs per plant largely because of longer branches on un (Fig. 2) was considerably shorter on the 8 browsed plants (Fig. 2). browsed sites than on the unbrowsed Butte Total LS length (annual growth; Fig. 5) was and Deer exclosure sites (P < 0.001), reflect not significantly correlated to total branch ing the influence of browsing in modifying length (r = 0.45, P > 0.13) across all sites. branch length. Accordingly, at the Dry Cotton Although the unbrowsed Butte and Deer ex wood location the unbrowsed site had longer closure sites that had the longest branches hranches than the Cattle exclosure site (P < also had high total LS growth, this total length 0.01; with only deer browsing), whereas the was not significantly (P < 0.10) longer tban on Cattle + Deer site had the shortest hranches most browsed sites (Fig. 5). (P < 0.05). Total branch length per plant (Fig. Long shoot length per unit length ofbranch 2) at the Burned and Unburned Steep Moun varied between several sites and sometimes tain sites did not differ (P < 0.10), which indi between years (Fig. 6). This growth rate gen cates that the combination of growth and erally increased on the MHWMA sites (Power browsing (Table 2) was similar between these line, Willow Creek, Railroad Gulch, and High sites for the previous 3 yr. Rye) in 1991, although all other sites decreased. The number and length of LS produced Neither total LS length (Fig. 5) nor LS length varied by site and year, with 3 sites having per unit of branch (Fig. 6) differed between fewer LS (Fig. 3) and 3 sites having longer (P the Burned and Unburned sites, although the < 0.05) LS in 1991 than in 1990 (Fig. 4). How- Unburned site had significantly (P < 0.01) more
3000,------, * w 01990 .1991 2500 - § 2000 * x ~- -'" -'•" ~ 1500 0 * **** * E" b b b b c c c c b 1000 y "~ , d d c .... c d d , , , 500 ,
o .j...L Butte Deer Cattle Cattle & Burn Unbum Pwrline Willow RR' High Rye Excl Excl Deer Creek Gulch
Fig. 2. Average total branch length (mm) of 1-,2-, and 3-yr-old twig segments for plants (n = 10) at all study sites. The 2 unbrowsed sites are at left. Site~to-site, within~year differences (LSD, P < 0,05) are denoted by columns with unlike letters (abed = 1990, wXYZ = 1991). The asterik (*) denotes a site that had a year-to-year difference (Wilcoxon test, P < 0.05). 34 GREAT BASIN ATURAL\ST [Volume 58
40 ...------~----, * • 01990.1991 35 • * * • 30 . a • • • z , b "c , • e 25· x b x -" x y y !" ~ 20 '0 x • 11 15 e, z 10
5
o ./-L B_ Dw catUe c.we & 8U'n Unbum Pw11ine Willow RR High Rye Exe' Exct Deer Cr'Itn: Gulch
Fil4- 3. Avw:agc !lumber oflong shoots (1.S) per branch for plants (n = 10) at an study siles. The 2 unhrowsed sites are Ht left. Site-ta-site, ...\!ithin-year tlifferences (LSD, P < 0.05) are denoted by columns with unlike letters tab = 1990, xyz = 1991). Only thos~ sites denoted by an aslerik (*) showed a year-w·yt:ar, within-sHe di(ference (Wilcoxon test, P < 0.(5).
80.------, 01990.1991 * x c *
'0
o1-L Butte De.r Cattle Cattle & Bum Unbum Pwrline WiDow RR High Rye Exel Excl Deer Creek Gulch
Fi~. 4. Average (n = 10) long shoot (L8) length (total long shoot length divided by number oflong shoots) per branch for all study sites. The 2 unbrowsed sites are at left. Sitc-to-site, within-yenr differences (LSD, P < 0.(5) are deooted by columns with unlike letlcrs (abt:.'d = 1990, wxyz := 1991). Only those sites denoled by an asledk (*) showed a ycar-Io year, withilHite dilTerencc (Wilcoxon test, P < 0.05). 1998J BITfERBllUSn CnOWTH 35 2OOOr------,
1BOO 01990.1991 • • E 1600 x E - '400 ~ ~" '200 o ,g '000 tJl ",BOO "o ..J 600 ~ ~ '00
200
o J-.I-- Butte Deer Cattle Cattte & Bum Unbum Pwrllne WiUow RR High Rye Excl Exd Deer Creek Gulch
Fig. 5. Total long shoot lenW-h (mm) per branch (11 = 10) lor 1990 nnd 1991 for all study sites. The 2 unhrowsed sites are at left Site-lo-site, within-year difference.s (LSD, P < 0.05) nrc denoted by columns with unlike letters (ab = 1990, xy~ = 1991). Only those sites denoted by an asterik (*) showed a yc.lr·to-ye 4r------, 01990.1991 * * 3.5 , '5 3· !!" '"E25 E '!J 2 E E 1.5 , 0.5 o .t-J- Butt. De... cattle Cattle & Bum Unlxnl PwrIine WiRow RR High Rye Ex" Excl Deer Creek Gulch Fig. 6. Long shoot (LS) length per length of brdJIch (,"m/mUl) (n = 10) in 1990 and 1991 for aU study sites. The 2 unbrowsed sites are at left. Site-Io-site, \\ritbin-year differences (LSD, P < 0.05) are denoted by columns with unJike Id ters (abc = 1990, vwxY.l = 1991). Only those sites denoted by an asterik (*) showed a year-to-year, within-site difference:: (\Vilcoxon test, P < 0.05). 36 GREAT BASIN NATURALIST [Volume 58 bitterbrush (Fraas et al 1992). The fact tbat FRAAS, WW, c.L. WAMBOLT, AND M.R FRISNA. 1992. Pre both the growth rate and the browsing level scribed fire effects on a bitterbrush-mountain big sagebrush-bluebunch wheatgrass community. Pages (Table 2) were the same at the 2 sites suggests 212-216 in WP. Clary, E.D. McArthur, D. Bedunah, that browsers removed approximately the same and C.L. Wambolt, compilers, Proceedings of the amount of LS material from each branch at symposium on ecology and management of riparian each site. shrub communities. USDA Forest Service, General At the Dry Cottonwood exclosure site, LS Technical Report INT-289, Ogden, UT. FRANKLIN, ].F., AND CT. DYRNESS. 1973. Natural vegeta length per unit of branch (Fig. 6) was greater tion ofOregon and Washington. USDA Forest Service, on the Cattle + Deer site than the Cattle Pacific Northwest Forest and Range Experiment exclosure site in 1990, as was the Cattle exclo Station General Technical Report PNW-8, Portland, sure greater than the totally unbrowsed Deer OR. GAHRISON, G.A. 1953. Effects of clipping on some range exclosure that year (P < 0.01). This tendency shrubs. Journal ofRange Management 6:309-317. supports Tueller and Tower's (1979) stagnation GIUNTA, B.C., R. STEVENS, K.R. JORGENSEN, AND A.P. theory, which predicts relatively higher growth PLUMMER. 1978. Antelope bitterbrush: an important rates at higher browsing levels. Reiner and wildland shrub. Utah Division of Wildlife Research Urness (1982) also reported that livestock graz Publication 78-12. GUENTHER, G.E. 1989. Ecological relationships of bitter ing increased bitterbrush growth by reducing brush communities on the Mount Raggin Wildlife herbaceous competition during the growing Management Area. Unpublished master's thesis, season. Montana State University, Bozeman. Overall we found only relatively minor vari HORMAY, A.L. 1943. Bitterbrush in California. USDA For est Service, California Forest and Range Experiment ations in the characteristics measured among Station Research Note 34, Berkeley, CA. browsed unburned sites or between the 2 un KUFELD, RG 1973. Foods eaten by the Rocky Mountain browsed sites. However, our data indicate a elk. Journal of Range Management 26:106-113. fundamental difference in bud allocation pat KUFELD, Re., O.C. WALLMO, AND C. FEDDEMA. 1973. terns between browsed and unbrowsed bitter Foods ofthe Rocky Mountain mule deer. USDA For est Service, Rocky Mountain Forest and Range brush plants and suggest that plants protected Experiment Station Research Paper RM-ll1, Ft. from browsing for many years should not be Collins, CO. used as a standard for comparison with plants LAY, D.W 1965. Effects of periodic clipping on yield of exposed to normal browsing pressures. Our re some common browse species. Journal ofRange Man agement 18:181-184. sults should increase knowledge ofhow bitter LUND, R 1991. MSUSTAT statistical analysis package, brush responds to browsing. An understand microcomputer version 5.00. Montana State Univer ing of the relationships between bitterbrush sity, Bozeman. growth characters and management strategies MARTINSEN, C.F. 1960. The effects of summer utilization should improve management for bitterbrush of bitterbrush in northcentral Washington. Unpub lished master's thesis, University ofIdaho, Moscow. stands and the big-game winter ranges they MENKE, ].W, AND M.]. TRLICA. 1983. Effects ofsingle and often occupy. sequential defoliations on the carbohydrate reserves offour range species. Journal of Range Management ACKNOWLEDGMENTS 36,70-74. MUEGGLER, W, AND WL. STEWART. 1980. Grassland and shrubland habitat types ofwestern Montana. USDA We thank Dr. Richard E. Lund ofthe Depart Forest Service, Intermountain Forest and Range ment of Mathematical Sciences, Montana State Experiment Station, General Technical Report INT University, for assistance in conducting the 66, Ogden, UT. statistical analyses. NOAA. 1991. Climatological data, Montana, 84-94 (1-13). National Climatic Data Center, Asheville, NC. NORD, E.C. 1959. Bitterbrush ecology-some recent find LITERATURE CITED ings. USDA Forest Service, Pacific Southwest Forest and Range Experiment Station Research Note 148, ALDEHFEH, J.M. 1977. A taxonomic study of bitterbrush Berkeley, CA. (Purshia tridentata [Purshl DC.) in Oregon. Unpub REINER, RJ., AND P.]. URNESS. 1982. Effect of grazing lished master's thesis, Oregon State University, Cor horses managed as manipulators of big game winter vallis. range. Journal of Range Management 35:567-571. BrLuHouGH, C.]. 1990. Growth responses of sagebrush SHEPHERD, H.R. 1971. Effects of clipping on key browse and bitterbrush to simulated winter browsing. Unpub species in southwestern Colorado. Colorado Game, lished master's thesis, Utah State University, Logan. Fish, and Parks Division, GFP-R-T-28, Denver. CUHK, RG., C.M. BRITTON, AND EA. SNEVA. 1982. Mor SNEDECOR, G.W, AND WG. COCHRAN. 1989. Statistical tality of bitterbrush after burning and clipping in methods. Iowa State University Press, Ames. eastern Oregon. Journal of Range Management SOKAL, R.R, AND EJ. ROHLF. 1981. Biometry. WH. Free 35ml-714. man and Co., New York. 1998] BITfERBRUSH GROWTH 37 STEJ:.IHOFF, H.W 1959. Some eRects of clipping bilter WA.\lBOLT, C.L., Wow. FnAAS. AND M.R. FR1SI~". 1996. brush at different intensities. Pages 23--24 in Tntns Variation in bitterhrush (Purshia tridelltata Pursh) actions of the fourth :lnnual summer conference of crude protein in southwestern Montana. Creat Basin the Central Mountains and Plains Section of the Naturalist 56:20.'5-210. Wildlife Society. WINWAHD, A.H., ANn J.A. FINDLF.Y. 1983. Taxonomic v