• Viewpoint: Livestock influences on riparian zones and fish habitat: Literature classification

      Larsen, R. E.; Krueger, W. C.; George, M. R.; Barrington, M. R.; Buckhouse, J. C.; Johnson, D. E. (Society for Range Management, 1998-11-01)
      A key was used to classify articles about livestock influences on riparian zones and fish habitat into 3 classes: papers that contained original data, those that were commentary, and reports about methodology such as classification systems, policies, and monitoring criteria. Four hundred and twenty-eight of the total articles were directly related to grazing impacts on riparian zones and fish habitat. Only 89 of these grazing impact articles were classified as experimental, where treatments were replicated and results were statistically valid. This analysis revealed several limitations of riparian grazing studies that included: (1) inadequate description of grazing management practices or treatments, (2) weak study designs, and (3) lack of pre-treatment data. More long-term, replicated treatment studies are needed in the future.
    • Use of native plants on federal lands: Policy and practice

      Richards, R. T.; Chambers, J. C.; Ross, C. (Society for Range Management, 1998-11-01)
      Changing social values and advances in ecological knowledge determine native seed policy for revegetating range and forest lands. Natural resource managers are shifting from seeding introduced species for their widespread adaptability to reestablishing native species in order to maintain or restore the genetic and ecological integrity of naive ecosystems. Addressing the problems of reestablishing native plants on a site-specific basis has been increasingly recognized as an integral part of ecosystem management of large landscapes. We review the formation and implementation of native seed policy for fire rehabilitation and mining reclamation by the major federal land management agencies in the United States, the Department of Interior's Bureau of Land Management (BLM) and the Department of Agriculture's Forest Service. We then examine native seed policy implementation on specific land revegetation projects over the past 10 years for 4 BLM districts in the state of Nevada. We conclude with an analysis of native seed policy in principle versus practice and suggest implications for future policy review and implementation.
    • Seed weight and germination time affect growth of 2 shrubs

      Hou, J.; Romo, J. T. (Society for Range Management, 1998-11-01)
      The objective of this study was to investigate relationships between seed size, time of germination, and seedling growth in winterfat (Ceratoides lanata (Pursh) J. T. Howell) and silver sagebrush (Artemisia cana Pursh). Individual seeds of winterfat were placed into 6 weight classes ranging from 1.5-2.0 to > 4.0 mg seed-1 while silver sagebrush seeds were separated into 7 classes ranging from 0.53 to 0.83 mg seed-1. Seeds were incubated at 18 degree C, seedlings with radicles < 3.0 mm were removed at 1, 2, 3, 4-5 and 6-12 day intervals, grown 5 days in darkness at 18 degree C, and axial length measured. Total germination of winterfat increased 5.5% mg-1 increase in seed weight, but germination rate was similar among weight classes, averaging 53.1% day-1. Seed weight and time of germination interactively influenced growth of winterfat seedlings. Seedling length of winterfat was more than 2-fold greater in the > 4.0 than the 1.5-2.0 mg seed-1 class while lengths of seedlings in the > 2.0-2.5 through > 3.5 to 4.0 mg seed-1 weight classes were intermediate. Seedling length decreased 0.9 to 3.3 mm for each day that germination was delayed from 1 to 12 days with the least and greatest inductions occurring for lightest and medium weight seeds, respectively. Total germination for silver sagebrush initially increased with seed weight, but declined at weights greater than about 0.57 mg seed-1; germination rate was similar (57.1% day-1) among weight classes. Seedling length of silver sagebrush increased 0.3 mm mg-1 increase in seed weight whereas length decreased curvilinearly as time to germination was delayed. When winterfat is used for restoration, relatively heavy seeds should be used because they have the greatest germination and produce large seedlings. Because seedling length of silver sagebrush increased with increasing seed weight it is also desirable to select heavier seeds; however, reduced germination in heavier seeds may necessitate increasing seeding rates.
    • Observation: Life history of spotted knapweed

      Jacobs, J. S.; Sheley, R. L. (Society for Range Management, 1998-11-01)
      Spotted knapweed (Centaurea maculosa Lam.) is a non-indigenous weed infesting large areas of rangeland in western North America. Life history models have been used to identify key processes regulating weed population dynamics and may be valuable in developing and testing integrated weed management strategies. Our objective was to characterize the life history of spotted knapweed. Demographic attributes were monitored monthly during snow free periods beginning August 1994 through October 1996 on 2 sites. Data were arranged into life history tables, and sensitivity analysis was performed to determine key transition phases affecting seed output. Spotted knapweed seed production ranged from 998 to 7815 viable seeds/m2 at both sites during the study. Seeds reaching the soil averaged 41 and 50% of seed output at sites 1 and 2, respectively. Less than 6% of seeds reaching the soil germinated in the fall at both sites. Recruitment peaked in April at 36% and in June at 20% of seeds reaching the soil on sites 1 and 2, respectively. Spotted knapweed juvenile density peaked August 1995 and June 1996 at both sites. Peaks corresponded with the beginning of the summer dry period. Plants bolted beginning June 1995 and May 1996. Sensitivity analysis identified early-summer juvenile survivorship, late-summer adult survivorship, transition from juvenile to adult, and seeds produced per adult as critical stages for spotted knapweed seed output. Management strategies that reduce spotted knapweed populations at these stages are likely to have the greatest impact on spotted knapweed population growth and spread. A weed population dynamics model using the life history demographic data was developed and can be used to design and test integrated spotted knapweed strategies.
    • Microhistological analysis of sheep gastro-intestinal content to confirm poisonous plant ingestion

      Yagueddu, C.; Cid, M. S.; Lopez, T. (Society for Range Management, 1998-11-01)
      The epidermal remains of 4 poisonous plant species that produce acute intoxication in ruminants were quantified by microhistological analysis in the gastro-intestinal content of sheep experimentally poisoned. These species were 'romerillo' or 'mio mio' (Baccharis coridifolia DC); 'duraznillo negro' (Cestrum parqui L'Hérit.); 'poison hemlock' (Conium maculatum L.), and 'sunchillo' (Wedelia glauca (Ort.) Hoff.). All of these species produce important economic losses of livestock in the Flooding Pampa, Buenos Aires, Argentina. The plants used for intoxication were at the vegetative stage of growth. Results indicate that the microhistological technique can be used to confirm the diagnosis of ruminant intoxication by duraznillo negro, romerillo, and sunchillo, but not by poison hemlock because digestion degrades its fragments beyond recognition. It would be convenient to sample the final sections of the digestive tract to confirm romerillo and sunchillo ingestion, because their fragments tend to concentrate there. The uniformity of duraznillo negro fragment distribution would allow identification of this species from any section of the digestive tract. However, the considerable variability in fragment distribution found among animals poisoned with the same plant species makes it necessary to sample more than 1 digestive region if only 1 animal is available for necropsy.
    • Long term dynamics of 2 populations of Prosopis caldenia Burkart

      Dussart, E.; Lerner, P.; Peinetti, R. (Society for Range Management, 1998-11-01)
      Invasion of pristine grasslands by caldén (Prosopis caldenia Burkart), and increased densities of this species in savannas are well-known vegetation changes in the semiarid region of central Argentina although little is known about its rates and patterns. In this paper we studied the relationship between dynamics of 2 representative P. caldenia populations and factors that could control the invasion process such as range management, fire events and precipitation regimes. Rates of implantation and spatial patterns are quantified using the present age distribution and dendroecological techniques. The pristine landscapes of the 2 study sites were grassland plains with (Site 2) and without trees (Site 1). The present density of the caldén is 586 and 1,259 shrubs/ha in Site 1 and 2, respectively. No evidence of clustering was found at the spatial scale of the study (p = 0.52, Site 1 and p = 0.08, Site 2 for n = 112). The ages of sampled individuals ranged from 3 to 65 years in site 1 and 8 to 55 years in Site 2 (only trees with diameter lower than 30 cm were sampled in Site 2). The importance of cattle as an effective disperser of caldén seeds was confirmed, as changes in measured establishment rates coincided fairly well with changes in cattle management. Establishment rates during the period of sheep grazing were 0.99 plant/ha/yr (16 years) in Site 1 and 10 plant/ha/yr (15 years) in Site 2. However, 10 years after the introduction of cattle these values reach 12.7 plant/ha/yr and 48.5 plant/ha/yr, respectively. One fire event occurred at each site (1980 in Site 1 and 1964 at Site 2). This factor did not change the density trend at Site 1, and at Site 2 it coincided with cattle introduction and caused an impressive increase in tree establishment. Fire was not an effective means of controlling P. caldenia populations. No relationship was found between population dynamics and available precipitation data.
    • Integrating genetic concepts into planning rangeland seedings

      Jones, T. A.; Johnson, D. A. (Society for Range Management, 1998-11-01)
      Choice of plant materials is a fundamental component of any rangeland rehabilitation, reclamation, or restoration project. We describe here an integrated approach for such decision-making. This approach considers site potential, desired landscape, seeding objectives, conflicting land use philosophies, appropriate plant materials, weed invasion, community seral status, and economic limitations. Technical limitations are considered in generating a plan that has the greatest potential for success. Determining whether native-site plant material is best depends on objectives, heterogeneity of the site's environment, uniqueness of the site, plant population size, and biotic or abiotic site disturbance. Fixation of alien genes into a population is referred to both as introgression, which may ensure maintenance of genetic variation critical for adaptation to a changing environment, and as genetic pollution, with the potential for swamping native cross-pollinating annual or short-lived perennial gene pools. Precautionary procedures during seed increase minimize genetic shift, which may be reversible, but genetic drift could result in permanent loss of desirable genes. A variety of germplasm classes, ranging from site-specific to widely adapted and varying in degrees of heterozygosity and heterogeneity should be considered.. Material originating from multiple sites may increase the opportunity for natural selection. An understanding of the magnitude and nature of a species' genetic variation, its relationship to ecological adaptation, and its interaction with other ecosystem components contribute to informed decision-making. Though often unavailable, experience is the best guide for predicting performance of materials on non-native sites.
    • Implications of desert rodent seed preferences for range remediation

      Longland, W. S.; Bateman, S. L. (Society for Range Management, 1998-11-01)
      Many desert plants germinate and establish from seed caches made by granivorous rodents. As a preliminary test of the feasibility of using rodent seed-caching activities to enhance seedling emergence of native desert plants, we tested desert heteromyid rodents for preferential consumption and/or caching of native ("target") seeds versus a commercial ("decoy") seed. The target/decoy seed concept relies on rodents caching both seeds, and preferentially consuming an inexpensive decoy seed as a sacrifice to reduce consumption of less preferred target seeds. We used cafeteria-style, paired seed choice trials to test 2 potential target seeds known to germinate from rodent scatterhoard caches (Indian ricegrass, Achnatherum hymenoides [R. & S.] Barkworth, and four-wing saltbush, Atriplex canescens [Pursh] Nutt.) against a potential decoy seed (millet, Panicum miliaceum L.). Millet was highly preferred to saltbush, and may indeed be a useful decoy seed when saltbush is the target of range restoration. Also consistent with the target/decoy seed concept, more Indian ricegrass than millet seeds were cached in laboratory trials, and all seeds were cached in scatterhoards more than in larderhoards, where the probability of seedling emergence is negligible. However, millet seed may not always be a good candidate for a decoy seed, as it was not preferred to Indian ricegrass and was cached more frequently than saltbush. Overall, we find results of these choice tests to be encouraging for applying the target/decoy seed idea. We consider the relative merits of this idea versus traditional revegetation techniques.
    • Impact of leafy spurge on post-Conservation Reserve Program land

      Hirsch, S. A.; Leitch, J. A. (Society for Range Management, 1998-11-01)
      Leafy spurge (Euphorbia esula L.), a noxious weed infests some of the 1.2 million hectares of Conservation Reserve Program (CRP) land in North Dakota. Once established a leafy spurge monoculture will reduce expected CRP benefits and impact returns to some post-CRP land uses. The study estimated statewide direct economic impacts of about 351,000 on post-CRP land maintained in vegetative cover, 1.118 million on post-CRP grazing land, and negligible (assumed 0) on post-CRP cropland, for a total of 1.469 million. Total annual direct and secondary economic impacts to North Dakota's economy were estimated to be 4.665 million, which would support about 57 jobs.
    • Heterogeneity in tall fescue pastures created and sustained by cattle grazing

      Cid, M. S.; Brizuela, M. A. (Society for Range Management, 1998-11-01)
      In continuous grazing systems the amount of herbage and its quality is a matter of primary concern. However, at moderate stocking, cattle grazing usually leads to the generation of patches differing in forage quality and quantity even in virtually monospecific pastures. This patchiness influences subsequent vegetation and animal responses. We analyzed the heterogeneity created and sustained by cattle grazing in a tall fescue (Festuca arundinacea Schreb.) pasture at stocking densities ranging from 1.8 to 4.1 animals ha-1 over 2 years. Cattle grazing created and maintained a mosaic of areas with different degrees of utilization. Heavily utilized patches had less biomass per unit surface, but their live biomass was more dense and had a higher nitrogen concentration. Patch boundaries fluctuated throughout the year at all stocking densities. Patch locations were more stable at the lower stocking densities, where cattle repeatedly returned to heavily utilized patches even though they represented less than 30% of the total surface. This reinforces the idea that, at low and moderate stocking densities, cattle can obtain a nutritional benefit by patch grazing. The percentage of heavily utilized patches reached a maximum value at an instantaneous grazing pressure of approximately 0.0016 animal units kg forage-l. When this threshold is passed, animal selection between patches could be conditioned by the presence of feces or thistles, and pasture condition affected by overgrazing of the heavily utilized patches.
    • Herbaceous response to canopy removal in southwestern oak woodlands

      McPherson, G. R.; Weltzin, J. F. (Society for Range Management, 1998-11-01)
      An understanding of overstory-understory relationships in southwestern oak woodlands is important for predicting response of these systems to disturbance or manipulation. The objective of this study was to evaluate the response of herbaceous plants to removal of the overstory in evergreen oak woodlands. Overstory plants were removed from 30 X 30 m plots in January 1993 and January 1994, and the response of herbaceous plants in these plots was compared to untreated controls for 5 and 4 years, respectively. The C4 graminoid and total biomass increased after overstory removal to as much as 10 times greater than controls, remained at elevated levels the second year, declined in subsequent years to 3-7 times the production of controls, and increased slightly during the final year of the study (1997). Overstory removal was necessary but not sufficient to affect herbaceous dicot biomass, which increased relative to controls during years with above-average winter precipitation. The C3 graminoids did not respond to overstory removal.
    • Grassland fire effects on barbed wire

      Engle, D. M.; Weir, J. R.; Gay, D. L.; Dugan, B. P. (Society for Range Management, 1998-11-01)
      Fire and its effects on rangeland plants, animals, soils, habitats, and watersheds has been studied extensively. Few studies have been devoted to fire effects on rangeland developments and no studies to our knowledge have been done on the effects of fire on barbed wire. From fire records and a known fence age at the Cross Timbers Experimental Range near Stillwater, Okla., we were able to determine the effect of varying fire frequencies on the breaking strength and zinc coating of traditional 2-point, double-stranded barbed wire. Samples from 4 burning frequency treatments, 8 locations each, of either 4 or 5-wire fencing were collected and stripped of their zinc coating for mass determination. Weight of zinc coating remaining on the wire was determined after being subjected to 0X, 1X, 2X, or 6X burn treatments over a 14-year period. A subset of 4 wires from 1X, 2X, and 6X burn treatments was tested for breaking strength. Photomicrographs and coating thickness measurements were also taken on samples from 1X, 2X, and 6X burn treatments. All tests were compared with unused wire of the same lot that had been in storage since fence installation. For the 6X burn treatment, breaking strength of 5,160 Newtons (N) and zinc coating thickness of 18.5 micrometer were equivalent to unused wire breaking strength and zinc coating (5,160 N, 16.6 micrometer respectively). It appeared that repeated fires did not adversely affect the corrosion resistance or breaking strength, and therefore service life of relatively new barbed wire fence.
    • Experimental evidence for sex-based palatability variation in fourwing saltbush

      Maywald, D.; McArthur, E. D.; Jorgensen, G. L.; Stevens, R.; Walker, S. C. (Society for Range Management, 1998-11-01)
      Two small-plot grazing trials were conducted in the spring of 1996 and the winter of 1997 to determine whether sheep would differentially graze fourwing saltbush [Atriplex canescens (Pursh) Nutt.] on the basis of shrub sex in a uniform garden. Consumption was determined using an Australian method of leaf tagging in conjunction with the Adelaide Technique of biomass estimation. The results confirmed anecdotal field observations that herbivores prefer to graze the male shrub during late spring. No sex based preference was apparent during winter. We suggest that differences in physiological vigor and/or chemistry may influence relative palatability of the sexes through time. Results of these experiments contrast with those for an Australian member of the genus (A. vesicaria Hew. ex Benth.), for which it was found that the female was the preferred phenotype throughout the year.
    • Estimating ashe juniper leaf area from tree and stem characteristics

      Hicks, R. A.; Dugas, W. A. (Society for Range Management, 1998-11-01)
      Ashe juniper (Juniperus ashei Buchh.) is increasing in density and areal coverage on the Edwards Plateau of Texas. This is causing changes in community leaf area that could impact the ecosystem water balance through increased transpiration and interception of precipitation. Our objective was to estimate leaf area of selected trees and shoots in a range of size and age classes using nondestructive methods. We harvested all leaf material from 9 trees ranging in height from 0.8 to 4.8 m and recorded tree height and canopy diameter. We divided each tree into 6 sections based on 3 horizontal strata and 2 vertical hemispheres. Projected leaf area of subsamples, collected from each section, was multiplied by pi to give full cylinder leaf area which was used to calculate specific leaf area (cm2 g-1). Dried leaf biomass for each stratum, hemisphere, and tree was multiplied by the specific leaf area to determine the leaf area. We harvested leaf biomass from shoots and measured in situ stem diameters, dried the leaf biomass, and multiplied it by the specfic leaf area to determine shoot leaf area. There was no significant effect of stratum or hemisphere on specific leaf area or of hemisphere on leaf area. The middle stratum had a significantly greater percentage of total leaf area (52%). Total tree leaf area was best predicted (r2 = 0.97) by canopy area. Shoot leaf area was best predicted (r2 = 0.93) by stem area. Canopy and stem area measurements are rapid, nondestructive means of accurately estimating Ashe juniper tree and shoot leaf area, respectively.
    • Effect of ground squirrel burrows on plant productivity in a cool desert environment

      Laundre, J. W. (Society for Range Management, 1998-11-01)
      Previous work demonstrated that burrows of Townsend's ground squirrels (Spermophilus townsendii Merriam) in cool deserts increased the amount of spring recharge of soil moisture compared to areas without burrows. The objective of this study was to test the hypothesis that this additional soil moisture would enhance plant productivity. I compared productivity of western wheatgrass (Agropyron smithii Rydb.) and big sagebrush (Artemisia tridentata Nutt.) plants adjacent to burrows to plants in areas lacking burrows. Grass productivity was estimated within an experimental grid containing cells of either 0, 2, 4, or 6 artificial burrows and was based on measures of annual above ground biomass production and number of seed heads produced. For sagebrush, productivity was estimated from bushes without burrows (controls) and ones having a natural burrow near their base. Sagebrush productivity was based on average length of new annual terminal growth of vegetative stems. The mean annual estimates of grass biomass (50.0 g m-2 year-1, SE = 11.8) was significantly higher in test grid cells with the highest number of artificial burrows than controls (42.6 g m-2 year-1, SE = 11.4). The mean of annual estimates of sagebrush stem growth for bushes adjacent to burrows was a significant 0.6 cm (SE = 0.11) longer than bushes without burrows. I conclude that the added moisture from spring recharge at ground squirrel burrows can increase plant productivity in a cool desert environment.
    • Developmental stages of winterfat germinants related to survival after freezing

      Bai, Y.; Booth, D. T.; Romo, J. T. (Society for Range Management, 1998-11-01)
      Diaspores of winterfat (Eurotia lanata (Pursh) Moq.) collected from 2 locations in the USA and 1 in Canada were imbibed at 10 degree C and grown to 4 different developmental stages (2, 3, 6, and 14 days of incubation), then subjected to cooling temperatures as low as -30 degree C. Differential thermal analysis was used to detect exotherms associated with ice crystal formation in germinants. The temperature at which exotherms occurred was recorded, and the subsequent growth and mortality of germinants were determined. Only 1 exotherm was observed, and that occurred in the low-temperature exotherm range (usually < -10 degree C). Changes in the freezing tolerance of germinants from seed to seedling was a gradual process as indicated by increases in exothermic temperature and mortality with increasing developmental stage. Whether the exotherm indicated a lethal event depended on the developmental stage of the germinant. Germinant survival was also affected by cooling below the exotherm temperature.
    • Cold-hardiness of silver sagebrush seedlings

      Hou, J.; Romo, J. T. (Society for Range Management, 1998-11-01)
      Silver sagebrush (Artemisia cana Pursh), a common shrub on Northern Mixed Prairie in Canada, is an excellent species to consider for ecological restoration. On the Canadian Prairies, freezing temperature can occur during April and early May, months when most silver sagebrush seedlings emerge. Decreasing temperatures in autumn or exposure to freezing temperature through winter may also be lethal to seedlings of this long-lived shrub. The purpose of this study was to characterize freezing tolerance in silver sagebrush seedlings because low temperatures may reduce establishment. Seedlings were grown from 1 week to 1 full growing season, exposed to freezing temperatures under controlled conditions, and lethal temperatures for 50 and 95% mortality (LT50 and LT95) were determined. Averaged across 1- to 6-week-old seedlings, LT50 and LT95 were -7.7 and -11.1 degree C, respectively. Changes in mortality with temperature variations were more gradual in younger than older seedlings, and mean LT95 was 2.8 degree C lower in 1- and 2-week than 4- and 6-week-old seedlings. Within age groups, death after freezing was greater in non-acclimated than acclimated seedlings. Virtually no non-acclimated seedlings survived -14 degree C, while mortality of acclimated seedlings was nearly nil in most cases. Only 6.9% (SE = 5.5) of seedlings grown under field conditions died in November after exposure to -39 degree C. Freezing tolerance of field-grown seedlings remained high over winter; seedling mortality after exposure to -39 and -45 degree C averaged 5.6% (SE = 4.1) in March. No seedlings survived temperatures lower than -15 degree C in April, and predicted LT50 and LT95 averaged -15.6 and -19.3 degree C, respectively. Increased mortality after freezing in April indicates seedlings de-acclimated as temperatures rose and day length increased in spring. Since the potential of developing freezing tolerance is greater in older than younger seedlings, silver sagebrush seedlings that germinate early in growing season may survive the winter better than those germinating later. Under normal circumstances, temperatures on the Canadian Prairies should not threaten survival of silver sagebrush seedlings during their first winter.
    • Clipping and Japanese brome reduce western wheatgrass standing crop

      Haferkamp, M. R.; Heitschmidt, R. K.; Karl, M. G. (Society for Range Management, 1998-11-01)
      Japanese brome (Bromus japonicus Thunb.) and downy brome (Bromus tectorum L.), introduced annuals, have invaded many northern mixed-prairie plant communities. This study determined the effect of removing Japanese brome and clipping western wheatgrass [Pascopyrum smithii Rydb. (Love)] on aboveground forage production of a western wheatgrass dominated northern mixed-prairie community. During early spring 1993, a wet year, and 1995, a drier year, western wheatgrass tillers were clipped to ground level in May or June and Japanese brome seedlings were left undisturbed or removed in circular, 1-m2 plots on a clay-pan field site. Western wheatgrass standing crop and tiller densities were estimated by clipping and counting in May and June, and these plus community standing crops were estimated in all plots after Japanese brome matured in mid July. Year effects were significant for standing crop and tiller density due to annual variation in amount and distribution of fall, spring, and early summer precipitation. Conditions were most favorable for tiller initiation of western wheatgrass and germination of annual brome seed in fall 1994 and for herbage production in 1993. Clipping western wheatgrass tillers reduced accumulated standing crop 230 to 350 kg ha-1 and reduced tiller weight by 17 to 58%. Standing crop of western wheatgrass was increased 102 kg ha-1 with removal of Japanese brome, while total standing crop was reduced 284 kg ha-1 with brome removal. Increased standing crop of western wheatgrass appeared to result from increased tiller density rather than increased tiller weight. Removal of Japanese brome from northern mixed prairie plant communities may increase production of associated perennial grasses, but managers should also expect a short-term decrease in total standing crop.
    • Changes in reproductive habitat of gray partridge after burning

      Novoa, C.; Dumas, S.; Prodon, R. (Society for Range Management, 1998-11-01)
      We investigated the effects of winter controlled burning on the breeding habitat of the Pyrenean gray partridge (Perdix perdix hispaniensis Reich.). Floristic composition and vegetation structure were sampled on 198 sites, including 64 recently used by hens for nesting or rearing of broods, 90 within partridge habitat burned under dry conditions, and 44 burned under wet conditions. During the early breeding season, birds selected shrublands of broom (Cytisus purgans (L.) Boiss.) with an average canopy coverage of 60% and an average height of 0.50 m. Birds avoided sites where shrub cover was more than 80% or less than 20%. The most critical effect of burning on gray partridge brood habitat was the reduction of the cover in the 2 vegetation layers providing protection against predators (0.05-0.25 m, and 0.25-0.50 m). In the case of "dry burning", the recovery of suitable habitat took more than 8 years for nesting hens and flightless chicks, but only 5-6 years for broods older than 3 weeks. Data obtained by radio-tracking eight broods indicated that "wet burns" (mean size = 4 ha) were better than "dry burns" (mean size = 15 ha) for maintaining good brood habitat. For the "dry burns", we recommend that burned patches be equal to or less than 5 ha and separated by unburned patches of 10-15 ha. In both cases, the frequency of fires should not exceed once every 15-20 years.