The Effect of Removal of Creosotebushes,
fromLarrea-dominated Shrub Grassland

Robert M. Chew
PO Box 16306, Portal AZ 85632

© 2002 RM Chew
Permission for use granted. Please cite the author.

 

    Abstract. All creosotebushes were removed from seven 10 m x 10 m experimental plots, and the coverages of subshrubs and grasses were measured in paired experimental and contiguous control plots over 24 years. There was no significant difference in the increase in cover of the subshrubs Parthenium incanum and Zinnia pumila between control and removal plots. This in spite of the observation that Parthenium seedlings were primarily found under cover of other plants, including Larrea, and that Zinnia seedlings were predominantly found away from shrubs and subshrubs. There was a clear difference in the response of grasses. Bush muhly, Muhlenbergia Porteri, declined in removal plots but increased in controls, whereas Aristida spp. increased in the greater open spaces of removal plots than in controls. Interpretation of vegetational trajectories was confounded by heterogeneity of the plots in terms of soil depth and initial coverages of subshrubs, and by interannual variation of rainfall.

Introduction

    In 1958 a 9.3 ha cattle exclosure was established in open range of the San Simon Valley 8 km north of Portal, Cochise County, Arizona, for purposes of studying the ecology of a shrubland dominated by Larrea tridentata. Growth rings of a sample of Larrea, and historical accounts, suggest that the site was previously a black grama (Bouteloua eriopoda) grassland that had been invaded by creosotebushes beginnig about 1890 during and after a period of prolonged drought and overgrazing. By 1958 Larrea canopy coverage was 17.4 %, which was 88%of the cover of woody species (Chew and Chew 1965). By 1976 two major vegetational changes were evident: a decline of Larrea cover and an increase in density and cover of subshrubs.  In 1977 creosotebushes were removed from seven 10 x 10 m quadrats to see if this would result in long term vegetation trajectories different from undisturbed contiguous control quadrats.

    There are various ways in which Larrea could affect associated vegetation. (1) “Islands of fertility” can develop under creosotebushes for a variety of reasons and favor some associated species (e.g. Reynolds et al. 1999; Cox et al. 1984). (2) Associated species can be favored under Larrea by protection from extremes of temperature and mammalian herbivory and by the accumulation of wind or water disseminated seeds (McAuliffe 1988). (3)Larrea can bring water up from deeper levels in the soil by hydraulic lift. The lateral spread of this water can possibly benefit associated species (Yoder et al. 1996; Dawson 1993). (4)Rainfall intercepted by the canopy goes partly (16-25%) to soil storage by stem flow and root channelization (Whitford et al. 1995). (5) Some species are associated with open space and hence limited by spread of Larrea (McAuliffe 1988), posssibly by active competition of roots (Fonteyn and Mahall 1981) and/or competition for water.

    Analysis of data following experimental removal of Larrea will test the hypotheses that: (1) There will be a reduction in coverage of some species that were favored by islands of fertility or another beneficial association with Larrea. (2) There will be an increase of coverage of species negatively affected by Larrea.

Materials and Methods

    Site. The site lies 1.5 km east of the foothills of the Chiricahua Mts. at 1370 m elevation on a gentle 1.8% slope of alluvium largely derived from limestone and volcanic tuff. Soil depth averages only 32.2 cm over hardened calcium carbonate (caliche); the soil surface is predominantly small gravel. The main plants are creosotebush (Larrea tridentata) and two subshrubs Parthenium incanum and Zinnia pumila. The main weed species, favored by grazing, are fluff grass (Tridens pulchellus) and snakeweed (Gutierreziae sarothrae). [Plant names herein are as in Kearney and Peebles, 1960.] The vegetation in 1958 is described in Chew and Chew (1965), and changes in herbaceous and subshrub species in Chew (1982). A summation of results of the long-term study is in progress.

    The 9.3 ha site was divided into a 10 x 10 grid of 30.5 m x 30.5 m (100’) sub plots. In 47 randomly chosen plots 30-meter permanent transects lines were established and beginning in 1985 plant cover, densities, and sizes of individual Larrea were measured at 10-year intervals. Soil depth over caliche was measured at 5-meter intervals along each transect.

    Vegetation manipulation. Seven pairs of 10 m x 10 m (0.01 ha) quadrats were randomly located in different parts of the exclosure that had different densities of. In one quadrat of a pair all Larrea were removed by digging out their root crowns, with as little disturbance of soil and other plants as possible. Retrospectively, disturbance could have been greatly reduced by clipping stems of Larrea and treating cut bases with a shrub inhibitor. If stems are simply cut and removed, new stems will develop (Havstad et al. 1999). Plants that were stem killed by prolonged below freezing weather in large areas near Portal, AZ in 1987 vigorously regrew stems in 1988 and thereafter (personal observations, unpublished). Only a few Larrea regrew from severed roots. The contiguous quadrat of a pair served as a control.
    Experimental and control plots were subdivided into 25 2 m x 2 m quadrats, and all perennial plants were mapped and measured. Length, width and average height of canopy was measured for shrubs and subshrubs. Basal diameter was measured for grasses. Seedlings of subshrubs were counted, and on a few occasions mapped.
    All seven pairs of plots were established in 1977. Three pairs were remapped in 1991 and two pairs were mapped again in 1998 and the third pair in 2001. The other pairs were not remapped until 1998, 1999 or 2001. All plots could not be remeasured in the same year because of other ongoing projects on the site. This obviously limits statistical interpretations.
    Size of shrubs and subshrubs was calculated as that of an inverted cone with an elliptical crown: Cover = Pi (L/2 x W/2). Coverage of each plot was summarized from the 2 x 2 m quadrats. Spearman Rank Correlations of species within the 2 x 2 m quadrats of each plot were calculated by the SigmaStat software programs (1992-1995). Regressions through the origin (as in Fig. 1 - 3 ) were calculated as in Snedecor and Cochran (1967:169-170). Importance values were calculated as:

    IV = Relative coverage of a species X Presence in 25 subquadrats.

    Because of heterogeneities of soil depth and plant densities, initial conditions in paired removal and control plots were often very different, which again limits statistical treatments.

Results

Larrea coverage through time

    In all seven control plots the coverage of Larrea declined through time, except one case with an increase from 1998 to 2001. The decline was at a rate of about 1.8% of the initial cover per year since first measurements (Fig. 1 ). There was consider able variability among plots with most points falling outside the 95% confidence limits of the value of b of the regression. A consistent decline of Larrea cover and volume was also observed in the 47 30-meter line transects measured at 10-year intervals beginning in 1958-59 (Chew, unpublished). No causal relationship has been made.

    The Larrea cover showed no relationship to soil depth.

Subshrubs

    Table 1 gives morphological data for the four species of subshrubs found in the seven paired plots. Gutierrezia sarothrae was also present but was highly variable in density and size through time. Parthenium incanum is the largest and most vigorous subshrub, increasing in the control and removal plots. As observed through time since 1979, it appears to be the subshrub least susceptible to death due to drought. Menodora is probably just as resistant to drought, but its low numbers make judgement uncertain. Dalea is least drought tolerant, losing stems and whole plants during drought seasons; Zinnia is intermediate. Parthenium and Zinnia have, in relative terms, densely leafed broad compact canopies; the canopy of Dalea is very open, with a small surface area of small compound leaves; that of Menodora is of linear leaves oriented vertically along stems. All four species are facultatively deciduous, retaining leaves through warm and moist to wet winters.

Parthenium incanum coverage through time

    Cover increased in both control and removal plots (Fig. 2 ) at rates of 20.3% per year (control) and 14.3% (removal). These rates are not significantly different since the 95% confidence limits of b are 0.082 - 0.342 for control plots and 0.020 - 0.266 for removal. However, Spearman rank correlations suggested that in eleven cases therewas a significant tendency of the cover of Parthinium and Larrea to increase together in the 25 subquadrats of plots. This was in 11 instances out of 28 sets of data for control plots through time and removal plots at initial time before removal of creosotebushes. Five instances were for 1977 measurements of plots and six were at later dates. There was no instance of a negative Spearman correlation coefficient.
    There are two things that confound interpretation of Fig. 2 . First, the initial cover of Parthenium, which ranged from 0.24 to 7.40 m2 per 0.01 ha plot, varied inversely with soil depth (Spearman rank correlation - 0.68, p = 0.074 for control plots and - 0.72, p = 0.055 for removal). On the 47 30-m transects, Parthenium density showed a highly significant inverse correlation with average soil depth of transects (p = <0.0001). This was consistent through the periodic 10-yr censuses, and pooled censuses. Second, the proportional increase in cover by year 22 in individual plots was highly significantly related to the initial cover (Spearman coefficient -0.813, p = 0.000). The plot with the lowest initial cover (0.24 m2) had an estimated proportional increase in cover of 16 times by year 22 (this outlier is not plotted in Fig. 2), whereas the plot with the highest initial cover (7.40 m2) had an increase of only 1.24.
    The number of seedling and small young Parthenium was counted in 14 plots in 1977 and seedlings were mapped in detail in four plots in 2001. Spearman rank correlation coefficients showed 15 significant positive relationships of seedling number withcoverage of Parthenium (n=6), Muhlenbergia Porteri (a grass, n = 4), Zinnia pumila (subshrub, n = 2), Larrea, Flourensia cernua (shrub), and total shrub coverage (each n = 1). There were three significant negative coefficients, once each with Zinnia, Parthenium and Menodora. Detailed mapping of location of seedlings in 2001 showed that they were predominantly under or near the canopy of other species and only infrequently away from other plants (Table 2 ). The associations of seedlings was very different between the two pairs of plots.

Zinnia pumila through time

    In 11 of 14 plots Zinnia cover increased by about 42% per year from 1977 to 1999 (Fig. 3 ); in general the paired control and removal plots were not significantly different. However, following an exceptionally dry winter and early spring from week 45 of 1999 through week 17 of 2000, there was a high mortality of mature plants.   Table 3 compares the incidence of dead individuals in 1998 and 1999 (as determined from maps) with that in August 2001. By 2001 cover in B7 control and removal had dropped to below initial values; plot G4 control, which started at a very low cover (0.249, not plotted in Fig. 3 ) was only 0.070 in 1999. In two plots, G4 removal and H8 removal, Zinnia showed no invasibility as there were no individuals present through 24 years.
    Unlike Parthenium, seedling and small Zinnia were predominantly found in the open, spaced away from other plants. Mature sized individuals in 10 plots (control and removal) in 1998 and 1999 were: (1) predominantly completely in the open or “loosely” associated with Gutierrezia and/or grass species that were found principally away from shrubs and subshrubs (Aristida spp., Bouteloua spp.) — mean 72.1% (range 59 - 85%), (2) under or next to, and possibly influenced by Larrea or other large woody species — mean 16.5% (9% - 26%), and (3) under or next to Parthenium — mean 11.4% (4% - 21%). The number of Zinnia categorized in each plot averaged 61 (range 0-162). There was no significant difference between control and removal plots. Zinnia had no frequent or consistent Spearman correlation coefficient with other species. This suggests that in the categorization above for Zinnia, their being under or next to shrubs and subshrubs did not involve any causal relationship, haphazard survival of seedlings.

There was no relationship of Zinnia cover to soil depth.

Other subshrubs

    Dalea formosa (found in 17 of the 35 plots censused) and Menodora scabra (5/35) never had the highest importance value for a subshrub in any plot census, due to their low relative cover and/or low presence. For Dalea median IV = 0.019 (range 0.001-0.18), median relative cover = 0.081 (0.010-0.43) and median presence 0.16 (0.04-0.56). Dalea had few significant Spearman correlations: negative with shrubs and Parthenium and grass species associated with them (n=5); positive with grass species typical of open locations (n=3). The incidence of Menodora was too low for analysis.

Grasses.

    Two categories of grasses were present in all plot censuses except for six in 2001: (a) Muhlenbergia Porteri (bush muhly), (b) Aristida spp. [Aristida longiseta, A. purpurea and A. ternipes were not separately distinguished in censuses] and Bouteloua eriopoda. Aristida spp. had the highest importance value in 18 of 34 censuses, M. Porteri in 10/34 and Bouteloua eriopoda in 6/34.

    M. Porteri had multiple positive Spearman rank correlations with shrubs and subshrubs: with Larrea (n = 9), Flourensia cernua (n = 8), Parthenium (n = 14), and Zinnia (n = 2). Aristida spp. had few significant correlations, but these were notably negative: with Larrea (n =1), Flourensia (n = 1), Parthenium (n = 5). The impression from maps is that Aristida “prefers” open ground away from woody vegetation, although it is often found next to Zinnia, which is also predominantly found away from larger woody species.
    Coverage of these grasses, and other uncommon species, varied with rainfall. Continuous weather records from the site since 1987 show extreme droughts in early summer 1994, late spring 1995, late summer 1998 and early spring 1999. These periods might have stressed grasses at one or more phenological stages. Before 1994 there were no extreme droughts and there were frequent wet, moist or normal biological seasons (Chew, unpublished). In the short term Muhlenbergia and Aristida increased in coverage through 1991, but had declined below initial coverages by 1998-1999 and 2001 (Table 4 ). The average values are suggestive: by 1991 Muhlenbergia, associated with woody species, increased more, by four fold, in control plots than in removal plots; its decline after 1991 was then the same in control and removal. In contrast, Aristida, associated with open areas, by 1991had increased much more in removal plots than control plots, and then disappeared by 2001. The data are limited, but suggestive, of different trends in these two grasses following experimental Larrea removal.
    Black grama, Bouteloua eriopoda, is of special interest since it was probably the dominant vegetation of the site prior to 1890. B. eriopoda was not present in any plot in 1977, but was present in 14 of the 21 censuses of 1991 and later, in both control and removal plots. It had the highest IV for grasses in six censuses. There were only four Spearman correlations, with no pattern.

Discussion

There is circumstantial evidence of considerable heterogeneity within the study site that is probably due to microtopography as it affects runoff and deposition of water and to variations of soil depth. The widespread annual, Gilia longiflora, consistently bloomed earlier along a 150-m transect that was slightly upslope of another transect 150 m transect. The most common cactus, Opuntia phaecantha var. major Engelm., showed stress of water shortage, by loss of surface chlorophyll, on the down-slope line before that of upslope. Several species, such as the perennial herb Croton corymbulosus and the annual grass, Bouteloua aristidoides, were found in only very limited spots and showed little or no invasibility in 20 yr or more.
    Soil depth over caliche varied among transects and within transects. For the 47 30-meter transects, soil depth averaged 32.2 cm (95% cl 29.57-34.79 cm; range 16.8-55.5 cm; coefficient of variation 0.288). Within transects, the average coefficient of variation was 0.245 (95% cl 0.255-0.319; range 0.101-0.796). Soil depth in the seven Larrea removal plots averaged 30.30 cm (24.73-35.85) and the control plots averaged 28.92 cm (23.22-34.62).

    There have been two major vegetational trends since the study site was established in 1958-59: (1) overall a consistent decline in coverage of Larrea with increasing number of dead stems and decreasing volume of individual shrubs (Chew, unpublished). (2) an increasing density and coverage of subshrubs, particularly Parthenium incanum. Two possible causes of the decline of creosotebushes are: (1) This invasive population was established since 1890, mostly in 1948-1958 (Chew and Chew 1965). It may be that the shrubs then grew to sizes they could not sustain on the resources of the shallow soil over densely cemented caliche, and then died back towards sustainable sizes.This is supported by the observation of Chew and Whitford (1992) that, in the present site, where creosotebushes are on and around the large abandoned mounds of banner-tailed kangaroo rats (Dipodomys spectabilis), where the soil is much deeper due to the digging of the rats, Larrea are dramatically larger and more robust than adjacent to the mounds. (2) The creosotebushes are affected by competition with increasing subshrubs and perennial grasses. In the shallow soil there is little or no opportunity for absorption of water by different levels of root systems (Dodd et al. 1998). New experimental plots have been established in which subshrubs and grasses have been removed to test whether this favors growth and/or stabilization of sizes of Larrea. From his study of creosotebush communities along the Rio Grande Valley in New Mexico, Gardner (1951) concluded that increase in subshrubs was an initial stage in the recovery of these desert shrub communities to grasslands.
    Havstad et al. (1999) studied the effects of periodic removal of Larrea over 50 years at a site in New Mexico which has vegetation similar to the present site and also shallow soil (25-30 cm) over indurated caliche. Larrea stems were cut and removed six times at intervals of 4 to 12 years (ave 8 y) from 1938 to 1995. They also observed a decline of Larrea in control plots.

    Present data do not support the hypothesis that Parthenium develops more coverage in control plots than in experimental plots without Larrea cover, in spite of the appearance that Parthenium are often clumped around Larrea. Certainly Parthenium did not decline after removal of creosotebushes, as observed by Havstad et al. (1999). If there is a benefit to Parthenium from Larrea - induced “islands of fertility”, the effects of these “islands” may persist for some time after removal of Larrea and / or Parthenium may be able to induce its own “islands” where it is already established. This would confound the development of any striking difference between paired control and experimental plots. The inverse relationship of Parthenium cover and density to soil depth is surprising. In a study in a semidesert range in New Mexico Herbel et al. (1972) found similarly, that cover and yield of grasses were inversely related to soil depth, and maximum at depths of 24-46 cm, and concluded that caliche under shallow soil holds moisture at a more readily available depth for grassses than can occur in deeper soil. Measurements by Chew (1982) support this conclusion for the present site; soil moisture was greater in the shallow soil of creosotebush dominated areas than in the much deeper soil of an adjacent community dominated by tarbush (Flourensia cernua), and that after rains of 15-38 mm the soil immediately above caliche was moister than soil 5 cm above the caliche. In shallow soil moisture may be held close enough to the surface to enhance the survival of shallow-rooted young Parthenium and their further growth. Survival of young Parthenium may be enhanced by cover of other species, but not particular species such as Larrea.

    The effects of intra- and interannual variability of rainfall also confound the occurrence of obviously different vegetational trajectories in experimental and control plots, particularly for grasses, which are so responsive to seasonal summer rains. Historically, Muhlenbergia Porteri “originally existed in extensive stands on the open range lands of southern Arizona but now occurs for the most part in the protection of shrubs and subshrubs” Gould (1973, p. 202). This grass showed no ability to increase in the period 1977-1991 after removal of Larrea, although it continued in association with Flourensia and Parthenium. Its decline in removal plots by 1991, although it increased in control plots (Table 3), suggests its sensitivity to loss of cover of woody plants. Conversely, the vigorous increase of Aristida spp. in removal plots suggest an exploitation of space freed of shrub cover. M. Porteri and Aristida have functional differences. Bush muhly has apical meristems and axillary buds, which persist after the growing season, allowing it to maintain food in elevated parts and to grow new leaves more quickly from above ground stems than does Aristida, which dies back to base in winter (Miller and Donant 1981). However, when it was experimentally drought stressed, bush muhly had greater reduction of specific leaf area than Aristida (Fernandez and Reynolds 2000). When field plots of Larrea shrubland were experimentally droughted from mid June to mid September for five years, all perennial grasses died (Whitford et al. 1995).
    The lack of clear cut differences between Larrea - removal plots and control plots is due to the confounding effects of hetergeneity of soil depth, weather and unknown factors. It might be possible to better test the hypotheses by matching control and experimental plot by average soil depth, rather that by placing them randomly contiguous to each other. This would possibly adjust for variation of Parthenium density with average soil depth Larrea density showed no significant relationship to average soil depth. However, because of the variation of soil depth within 10 x 10 m plots, the possibility for an effective design to test effects of plant removals is obscure.

Literature Cited

Chew, R. M. 1982. Changes in herbaceous and suffrutescent perennials in grazed and ungrazed grassland in southeastern Arizona, 1958-1978. American Midland Naturalist 108:159-169.

Chew, R. M., and A. E. Chew. 1965. The primary productivity of a desert shrub (Larrea tridentata) community. Ecological Monographs 35:355-375.

Chew, R. M., and W. G. Whitford. 1992. A long-term positive effect of kangaroo rats (Dipodomys spectabilis) on creosote-bushes (Larrea tridentata.)Journal of Arid Environments 2:375-386.

Cox, J. R., J. M. Parker, and J. L. Stroehlein. 1984. Soil properties in creosotebush communities and their relative effects on the growth of seeded range grasses. Soil Science Society of America Journal 48:1442-1445.

Dawson, T. E. 1993. Hydraulic lift and water use by plants: implications for water balance, performance and plant-plant interactions. Oecologia (Berl.) 59:565-574.

Dodd, M. B., W. K. Lauenroth, and J. M. Welker. 1998. Differential water resource use by herbaceous and woody plant life-forms in a shortgrass steppe community. Oecologia (Berl.) 117:504-512.

Fernandez, R. J., and J. F. Reynolds. 2000. Potential growth in drought tolerance of eight desert grasses: lack of a trade-off? Oecologia (Berl.) 123:90-98.

Fonteyn, P. J., and B. E. Mahall. 1981. An experimental analysis of structure in a desert plant community. Journal of Ecology 69:883-896.

Gardner, J. L. 1951. Vegetation of the creosote bush area of the Rio Grande Valley in New Mexico. Ecological Monographs 21:379-403.

Gould, F. W. 1973. Grasses of southwestern United States. University of Arizona Press, Tucson, Arizona, USA. 352 p.

Havstad, K. M., R. P. Gibbens, C. A. Knorr, and L. W. Murray. 1999. Long-term influences of shrub removal and lagomorph exclusion on Chihuahuan Desert vegetation dynamics. Journal of Arid Environments 42:155-166.

Herbel, C. H., F. N. Ares, and R. A. Wright. 1972. Drought effects on a semidesert grassland range. Ecology 53:1084-1093.

Kearney, T. H., and R. H. Peebles. 1960. Arizona Flora, Second Edition with supplement. University of California Press, Berkeley, USA. 1085 p.

McAuliffe, J. R. 1988. Markovian dynamics of simple and complex desert plant communities. American Naturalist 131: 459-490.

Miller, R. F., and G. S. Donant. 1981. Response of Muhlenbergia porteri Scribn. to season of defoliation. Journal of Range Management 34:91-94.

Reynolds, J. F., R. A. Virginia, P. R. Kemp, A. G. de Soyza, and D. C. Tremmel. 1999. Impact of drought on desert shrubs: effects of seasonality and degree of resource island development. Ecological Monographs 69:69-106.

SigmaStat Statistical Software, Jandel Corporation 1992-1995.

Snedecor, G. W., and W. G. Cochran, Statistical Methods. Sixth Edition. 1967. Iowa State University Press, Ames, Iowa, 593 p.

Whitford, W. G., G. Martinez-Turanzas, and E. Martinez-Meza. 1995. Persistence of desertified ecosystems: explanations and implications. Environmental Monitoring and Assessment 37:319-332.

Yoder, C. L., and R. S. Nowak. 1996. Spatial and temporal patterns of hydraulic lift among native plant species in the Mojave Desert. Bulletin Ecological Society of America 77:496.

---

Table 1. General characteristics of subshrubs. Means and (95% confidence limits) are as determined from log10 transformed data. nm = not measured; * two values to illustrate the difference that can occur through time, low value 1991, high 1998. Values based on 14 plots initially and one pair of plots measured 1977, 1991, 1998.

Fresh weight

Species	Mean volume, cm3	Total wt, gm	leaf:stem	Initial cover, m2/0.01	ha
Parthenium	11,200 (8820-14,280)*	135 (100-181)	22.5 :77.5	1.93 (1.12-2.74)	n=13a
	19,560 (15,700-24,400)*
Zinnia	1508 (1004-2256)	114 (68-194)	22.9:77.1	0.49 (0.27-0.71)	n=11b
Dalea	5348 (2705-10,570)	nm	nm	0.66 (0.03-1.29)	n=8c
Menodora	854 (695-1049)	42.3 (15-123)	nm	1.21 (0-2.59)	n=5d

a- one extreme outlier was omitted from among 14 plots

b- Zinnia was absent in two plots, and one 1977 outlier omitted

c- Dalea was present in only 5 plots 1977 and 3 more later

d- Menodora was not present in 1977, and only on 5 plots later

 

---

Table 2. Distribution of seedling and very young Parthenium incanum, Oct. 2001, in four plots, compared in terms of percentage of total number per plot.

	Gr Control	G4 Removal	B7 Control	B7 Removal
Under Parthenium	54.0	37.9	6.4	11.9
Under dead Zinnia	16.4	0	34.9	82.3
Under Larrea	12.6	0	26.6	0
Under, next to or among live grass clumps	4.9	21.9	1.1	0
Under, next to or among dead grasses	7.7	27.9	0.2	0
Under Opuntia	2.2	1.1	0.5	2.5
Under other shrubs, subshrubs	1.1	4.2	27.0	0.5
In open away from above	1.1	7.0	3.4	2.8
Total # seedlings per 0.01 has	183	627	563	570

---

Table 3. Incidence of dead mature-sized Zinnia in 1998-1999 and 2001. * These plots were not mapped 1998, 1999.

% individuals dead

Plot	n	1998-1999	n	2001
B4 CTRL	77	11.7	51	84.3
B4 RMVL	29	10.3	23	56.5
C4 CTRL	76	4.0	60	68.3
C4 RMVL	154	10.4	132	84.1
D0 CTRL	67	9.0	34	100
D0 RMVL	61	8.2	46	95.7
F6 CTRL	24	0	27	18.5
F6 RMVL	108	0	93	39.8
H8 CTRL	25	8.0	22	81.8
H8 RMVL	0		0
B7 CTRL*			194	93.4
B7 RMVL*			160	90.0
G4 CTRL*			17	82.4
G4 RMVL*			2	0

---

Table 4. Mean percentage change of basal coverage from initial value through time fo Muhlenbergia Porteri and Aristida spp.   * = total basal coverage as cm2 in 10 m x 10 m census plot.

1977 coverage*

Mean percentage change     from 1977

	media (range)	1991	1998-1999	2001
M. Porteri
Control Plots	927 (180-3310)	+ 162	- 2	- 96
Removal Plots	633 (170-14,500)	+ 39	- 45	- 100
Aristida
Control Plots	2590 (710-7860)	+ 12	- 4.5	- 100
Removal Plots	1340 (210-8060)	+ 509	- 2.8	- 100

 

---

Fig. 1.  Declining cover of Larrea in the seven control plots over 24 years, 1977-2001.  Broken lines show 95% confidence limits of regression: b = 0.0115 and b = 0.0245.

 

---

Fig. 2.  Increase of cover of Parthenium by year 22 as measured then or extrapolated from the nearest measurement; b = 0.203 for Control Plots, b = 0.143 for Removal Plots.