124 Desert 3131 Autumn 1981

Introduction Vegetation of the Gila Lands along the Gila River and its tributaries imme- diately northeast of Safford, (Fig. 1), are gener- River Resource Area, ally considered a broad transition between Chihuahuan Desert to the east and to the west. How- ever, streams flow from highlands that are intimately asso- Eastern Arizona ciated with the-Rocky Mountain and Plains provinces to the north and northeast, respectively, and to the south are W. L. Minckley and "sky islands," disjunct mountain ranges that merge with the northwestern Mexican Plateau (Brown and Lowe Thomas 0. Clark 1978). Along these routes, unique biological communities Department of Zoology and and individual species have gained access to the region. Center For Environmental Studies Many kinds of plants and animals occupy the area, but Arizona State University, Tempe, Arizona some species seem to have been blocked by past or present conditions from moving into or dispersing through it. The region is one of the better studied parts of Arizona from the standpoints of geology and water resources, because of important mining and agricultural interests, but biological information is scattered and sparse. This report provides descriptive information on vegeta- tion of the area based upon field work conducted from October 1976 through September 1979. Intent of the study was to obtain baseline information prior to potential de- velopments, and it was accomplished under Contract No. YA-512-CT6-216 from the U.S. Department of Interior, Bureau of Land Management, to Arizona State University.

Description of the Area The Gila River Resource Area lies mostly below 1,500 m elevation (with the notable exceptions of the tops of Tur- tle Mountain and Guthrie Peak; Fig. 1). Topography is characterized by relatively short, parallel mountain ranges, surrounded by broad, gently sloping alluvial val- leys, and is part of the Basin and Range Physiographic Province of Fenneman (1931). Major strUctural features in- clude the Pinalerio (Graham), Gila, and Peloncillo moun-

Figure L Gila River Resource Area, eastern Arizona (within dashed lines), showing some place names men- tioned in text, general topography, and drainages. Con- tours are 1,000, 1,500, and 2,000 m, left to right, and width of the sketch map is about 32 km. Minckley and Clark Gila River Resource Area 125

tarns, rising above alluvial basins—the Safford and Dun- The study area is generally included in the Southeastern can valleys—which were filled during the same orogenic Section of Arizona of Green and Sellers (1964) with regard periods that produced the adjacent mountain ranges to climate. Precipitation is strongly bi-seasonal, charac- (Clarkson 1979). Perennial streams and many ephemeral terized by winter rainfall (snow rarely occurs), spring watercourses are deeply incised in conglomerates and drought, and summer-autumn monsoons (Lowe 1967). An other strata, and flow through alternating canyons and example of this variation for the period 1893-1965 at Clif- open valleys. Soils of upland areas are locally derived from ton, Arizona, is as follows (Burkham 1970): winter 12.3 -± heterogenous strata (Martin and Fletcher 1943), and gener- 7.4 cm (± one standard deviation); spring 1.9 ± 1.8; and ally are Gray Desert types developed from granites, gneis- summer-autumn 18.2 ± 5.8 cm (mean annual precipita- ses, schists, and limestones (U.S. Soil Conservation Ser- tion 32.3 ± 9.9 cm). Temperatures along the Gila River are vice 1975). Alluvial soils are of more distant origins, and mild to warm throughout the year. Below freezing tem- may be thick, of heavier texture, and sometimes contain peratures are rare at Safford, even in the coldest winters. zones of salt deposition (Lapham and Neill 1904). Essen- At Duncan, Arizona, freezing temperatures are not un- tially all soils are thermic in nature (mean annual temper- common in the period December through February. Day- ature at a depth of ca. 50 cm of 15-22° C), and are classed as time temperatures in colder months typically reach 20-25° semi-arid (average annual precipitation 25.4-38.1 cm) (U.S. C, and ranges of 15-20° C in a single day are not uncom- Bureau of Land Management 1978). mon. Summer temperatures often reach 32-38° C in mid- afternoon. Low humidities are the rule, and evaporation rates from water surfaces fall between 7 and 10 times the Table 1. Legal descriptions of selected transect origins and directions of 1964). transect lines in degrees from true north, Gila River Resource Area, average annual precipitation (Green and Sellers eastern Arizona. Transect lengths were 1,050 m unless otherwise noted. Methods and Materials Permanent transects (Table 1) were arbitrarily estab- Tran- Legal Descriptions and lished early in the study for sampling of vegetation and for sects Transect Directions Habitats census of mammalian and avian populations (to be pre- GL-1 NW1/4,NE1/4,T6S,R28E,S29. 260° Bosque, recent alluvium sented elsewhere). Vegetational data developed for GL-2 NE1/4,NE1/4,T6S,R28E,S28. 100° Bosque, recent alluvium selected transects include frequency of occurrence (Zim- (840 m) GL-3 NW1/4,NE1/4,T6S,R28E,S29. 000° Dry side canyon merman 1969) and tree counts (Anderson et al. 1977a). GL-4 SE1/4,SEV.,T6S,R28E,S20. 000° Wet side canyon Vegetation description on the basis of spot sampling GL-5 NE1/4,SEV4,T6S,R28E,S21. 180° Terrace, valley fill (Zimmerman 1969) was accomplished by noting species' (735 m) 2 GL-6 NE1/4,NE1/4,T6S,R28E,S21. 165° Terrace, valley fill presence in a circular, 250-m plot at ca. one-km intervals GL-7 NE1/4,SWV4,T6S,R28E,S21. 350° Upland, dissected throughout different parts of the study area and its envi- valley fill rons. This provided information on frequency of occur- B -1 NE1/4,SWV.,T6S,R28E,S16. 340° Broad-leafed riparian, alluvium rence of major woody species, plus Cactaceae and B -2 SE1/4,NWV.,T5S,R27E,S36. 330° Broad-leafed riparian, Agavaceae, and their distributions along various environ- alluvium mental gradients. Smaller forms, such as hedgehog and B -3 NWV.,SWV4,T5S,R27E,S10. 340° Broad-leafed riparian, alluvium fishook cacti (Echinocereus spp. and Mammillaria spp.), B -4 SW1/4,SW1/4,T6S,R28E,S16. 330° Upland, dissected plus young ocotillo (Fouquieria splendens), may be missed valley fill by this technique, especially when they grew within beds B -5 SWV4,SWV4,T5S,R27E,S36. 110° Upland, dissected valley fill of Devil's cholla (Opuntia stanlyi), so estimates for those B -6 SWV4,NE1/4,T5S,R27E,S10. 270° Upland, valley taxa are minimal. fill/volcanics E -1 SW1/4,SE1/4,T5S,R29E,S30. 340° Broad-leafed riparian, Direct tree counts, including for present purposes alluvium species of cacti and agaves, were made in belts 15 m wide E -2 SE1/4,SWY4,T5S,R29E,S18. 010° Broad-leafed riparian, on each side of transect lines. Numbers of individuals of alluvium E -3 NWN,NE1/4,T4S,R28E,S35. 270° Wet side canyon each species and size classes of individuals were recorded E -4 SW1/4,SEV4,T5S,R29E,S30. 280° Upland, dissected for each 30 m advance along the transect. Data could valley fill therefore be interpreted as numbers of plants per 150 m2 E -5 SE1/4,NWV.,T5S,R29E,S18. 270° Upland, valley fill/volcanics plot, or as numbers along the line of transect. E -6 NWV4,NWV4,T4S,R28E,S26. 270° Upland, dissected Devil's cholla and certain growth forms of mesquite valley fill (Prosopis juliflora)1 presented special sampling problems. SF-1 SW1/4,NW1/4,T5S,R29E,S14. 010° Broad-leafed riparian, alluvium (525 m) The former reproduces vegetatively, resulting in 2 SF-2 NW1/4,NWV4,T5S,R29E,S24. 010° Upland, dissected large clonal stands that exceeded 100 m . At other times, valley fill when reproduction results from fragmentation of larger GU-1 NE1/4,NW1/4,T5S,R29E,S27. 270° Bosque, recent alluvium 2 GU-2' NEV4,SEV.,T5S,R30E,S35. 110° Bosque, recent alluvium plants, a scattering of small (0.5-5.0 m ) clumps was pre- (735 m) sent. The average size of clumps of this cactus was ca. 7.5 GU-4 NE1/4,SEN,T5S,R29E,S25. 000° Dry side canyon m2, and that value was divided into the estimated ground GU-5 SE1/4,NE1/4,T5S,R29E,S27. 270° Terrace, valley fill GU-6 SE 1/4 ,SE1/4 ,T5S,R30E,S31.120° Terrace, valley fill (840 m) 'The area studied is in a region of range overlap between Pro- GU-7 NE1/4,SW1/4,T5S,R29E,S27. 220° Upland, dissected valley fill sopis juliflora var. velutina and P. julifora var. torreyana (Kear- GU-8 NW1/4,SW1/4,T5S,R30E,S30. 000° Upland, dissected ney and Peebles et al. 1973), and they both were probably pres- valley fill (840 m) ent on plots. Lehr (1978) lists these not as varieties of P. juli- No transect GU-3 was established flora, but as P. velutina and P. glandulosa var. torreyana. 126 Desert Plants 3(3) Autumn 1981

Figure 2. Semi-desert Grassland near the summit of Solomon Pass, Graham County, Arizona, ca. 1,380 m elevation; summer 1977. Principal grasses are gramma, brome, and cotton-top among a scattering of prickly pear, low mesquite, and yucca.

covered by this species within each 150 m2 plot to be re- Results and Discussion. Variations in the Vegetation. corded as "number of plants." Mesquite also grew in low, Vegetational variation in uplands of the Gila River Re- multi-stemmed clumps in areas of impervious soils or source Area may be attributed to altitude, topography (in- otherwise inaccessible ground water (Lacey et al. 1975), cluding slope exposure), substrate, and land use. Uplands and under those conditions each clump also was recorded are defined, for purposes of this report, as montane areas or as a single plant in tree-count data. deeply dissected valley fills, with shallow soils and rela- 6 1 Foliage v.. 6 e was obtained by measuring the distance tively high percentages of exposed rock. The same factors at which /0 of a board was obscured by leaves at various influence vegetation on terraces along the major streams, heiglir intervals at a series of pre-determined points on and lowland habitats (riparian, washes, and side canyons), each side of a transect. Leaf volume is computed from with the additional controlling factor in lowlands being these data by use of the formula: depth to the water table. Our development of information on vegetation of the = 12EL K Gila River Resource Area is geared toward description and D' presentation of data on the native biota. The U.S. Bureau where K is foliage volume in m2 of foliage per m3 of space, of Land Management (USBLM 1978) produced substantial and D is the measured distance. Plant communities in the amounts of data on phenology of major plant species, Gila River Resource Area were assumed to be at local trends in amounts of forage available for grazing, and vege- climax condition, so foliage densities were measured once tation types relative to land use, and should be consulted at a time of maximal development, between May and July for additional information. (see Anderson et al. 1977). Plant identifications were by Upland Habitats. Topographic diversity of the Gila McGill (1979), with nomenclature generally following River Complex, producing a mosaic of substrates, hydrol- Lehr (1978) (see, however, Footnote 1). ogy, and climatic regimes resulted in a similar mosaic of Minckley and Clark Gila River Resource Area 127

Aidif 0

$4.;111 -410* • * d. • dr • o, • • t • -6 vt drso-d.,* • .„ ." 416--

4 ,

Figure 3. Scattering of junipers on slopes along Bonita Creek near the northern boundary of the study area, Graham County, Arizona; autumn 1977. Mesquite and hackberry in foreground are on a dry terrace of the creek. Pirion pine occurs in this area on rocky ledges and in narrow canyons, in the latter with various oaks. vegetational cover on upland areas. Vegetation on gentle Species characteristic of Interior Chaparral Associations slopes between 1,250 and 1,750 m was appropriately were scattered through upper elevations of Semi-desert mapped by Brown and Lowe (1978) as Semi-desert Grass- Grasslands. Included were scrub live-oak (Quercus tur- lands. Pure grass landscapes were, however, rare, and most binella), buckthom (Rhamnus crocea and R. califomica), had a high percentage of cacti and (Fig. 2). A veg- squawbush (Rh us trilobata), silktassel (Garrya wrightii), etation map of the area produced by remote sensing tech- mountain-mahogany (Cercocarpus montanus), and rare niques (Treadwell et al. 1976) reflects this phenomenon. manzanita (Arctostaphylos pungens). Jojoba (Simmondsia Invasion of grasslands by shrubs in the last century, in chinensis) was a major along dry cliffs of Bonita and part a function of overgrazing, has occurred through- Eagle creeks, and especially on open slopes of Solomon out the region (Lowe 1967). Grama grasses (Bouteloua Pass, west of the study area. spp.), brome (Bromus arizonica), cotton-top (Trichach- Fringes of Madrean Evergreen Woodlands penetrated ne califomica), bristle-grass (Setaria sp.), and many others downslope through boundaries of the study area, co- were seasonally abundant in areas heavily used by domes- existing with small stands of Interior Chaparral within tic livestock, and formed a perennial cover where larger areas of Semi-desert Grasslands Formation (Fig. 3). lightly grazed. Largest stands of relatively pure grasses A small, open Madrean Evergreen Forest on Guthrie Peak were present on Guthrie Peak and Turtle Mountain. The was not mapped by Brown and Lowe (1978) or by Tread- uppermost part of Bonita Creek, north of the study area well et al. (1976). Other pockets of that formation were in on the San Carlos Indian Reservation, flows through a canyons that enter Bonita and Eagle creeks near the north- well-developed Plains Grassland Association (Brown ern edges of the study area. Scattered oaks (Quercus and Lowe 1978), and the Eagle Creek drainage also has emoryi and Q. arizonica) and junipers (Juniperus mono- such areas, although smaller in size, in its middle sperma and J. deppeana) were concentrated on steep, reaches (downstream from Honeymoon). northerly-facing slopes in that area, and were joined there Table 2. Occurrences of various perennial plants on spot sampling plots from southwest to northeast over the northwestern flank of Guthrie Peak, Greenlee County, Arizona, autumn 1978. Symbols: X = present with a plot, and + = presents adjacent to a plot.

Taxa Sampling Plots 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 X + X X X + X X + X X X Fouquieria splendens + X X X X X + Acacia app. X + + X X X X X X X X X X X Cercidium floridum X + + X X X Prosopis iuliflora X + X X X X X X X X X X X X Zizyphus obtusifolius X X Lycium app. X X 1 Atriplex sp. X X X Opuntia spp.' X+ XX XX + X X XX XX XX X + X X X X fasciculata X + Yucca app. X -i- + X X X X + X + Dasylirion wheeleri X + X X X + X funiperus spp. + X X X + X X X + Quercus turbinella + X X X X X + X X Pinus edulis + + X Rhus trilobata + + + + + Nolina microcarpa X + X X X X X Unknown chenopods XX X X X X X X X 'Principally 0. phaeacantha, some 0. violacea.

Table 3. Occurrences of various perennial plants on consecutive plots along transects in upland habitats, Gila River Complex, eastern Arizona, summer 1978. Symbol: X = present on a plot

Taxa Transect B-6 Acacia spp. XXXXXXXXX XX XX XXX XXX XX XX • XXX Fouquieria splendens X XXXXXXXX X Ephedra fasciculata X XXX X Ferocactus app. X X X Echinocereus app. XX Opuntia spp.' XXXXXXXXXXXXXXXXXXXXXXXXXXXXXX Prosopis juliflora XXXXXXXXXXXXXXXXXXXXX XXXXXXXXXXX X X X Lycium spp. XXX XX XXXX XX XXXXXXXX XXXX XXXX X XX spinosa X X X XX X XXX Zizyphus obtusifolius X X X X X X X X X X Opuntia spinosior X X X X X X X X X Dasylirion wheeleri X X X Yucca app. X funiperus spp. XX XXXXXXXXXXXXX XXX X Quercus app. X XXXXXXXXX Rhus trilobata XXXXXXXXX Pinus edulis X Transect B-5 Larrea tridentata XXXXXXXXXXX XXXXX X XXXXXX XXXXXXXXXXXXXXXXX Opuntia spp. XX XXXX XXXX XX X X XXX XXXXXX XX Opuntia leptocaulus X X X X O. spinosior X X X Prosopis juliflora XX XXXXXXXX Lycium app. X X XX Ephedra fasciculata X Echinocereus app. X X Ferocactus app. X X X Opuntia stanlyi X X XXXX XX XXXXXXX XXXX Fouquieria splendeps X XX XXX XXXXXXXXXXXXXXXXXXXX Tina Transect GL-7 Larrea tridentata X XXXXX XXXX XX XX XX XXX X XXXX X X XXXXXX XXXXX XXX Opuntia spp. XXXXXXX XXX XX XXX XXXXXXXXXXXXXXXX O. stanlyi X X XX X XXXXXXXXX X O. leptocaulus X X X X X X X X X XXXX O. spinosior XX X X X X X X Echinocereus app. X XX X Mammillaria app. X Ferocactus app. X XX X Cereus greggi X X Ephedra fasciculata X X Lycium app. X X X Acacia app. X X X X X X X Fouquieria splendens X XX X X X X XX XXXX XXXX X XXXXXXXXXXXXXX Prosopis juliflora X XXXXXXXXXXXXXX Cercidium floridum XXXX X X X X 'Principally 0. phaeacantha, some 0. viola cea. Minckley and Clark Gila River Resource Area 29

Figure 4. Undissected valley fill with relatively dense Desertscrub association, Transect GL-7, Graham County, Arizona; summer 1977. Major plants are creosote-bush and ocotillo. by Colorado pifion pine (Pinus edulis) on the more pro- tween plots 18 (1,375 m) and 27 (1,030 m, on the Gila tected, rocky outcroppings. Small, isolated stands of Mad- River). Taxa characteristic of local Semi-desert Grass- rean Forest near Morenci included relict populations of lands, Yucca sp. and Sotol (Dasylirion wheeleri), plus Arizona cypress (Cupressus arizonica) (Lowe 1967). bear-grass (Nolina microcarpa), appeared between sam- Between 1,250 and 1,400 m (locally as low as 1,100 m) pling plots 4 (1,310 m) and 8, and disappeared abruptly Semi-desert Grasslands gave way to Desertscrub Associa- near plot 18 on the northeast. Chaparral species (scrub tions (Figs. 4-5), most of which were dominated by live-oak and squawbush) and Woodland elements (junipers creosote-bush (Larrea tridentata) and mesquite. and pition pine) were concentrated between sampling Quantitative data on such a complex of vegetation is point 8 and point 16 (1,470 m). Distribution of mesquite difficult to generate and to interpret and present. Occur- up to plot 16 on the northeast slope of Guthrie Peak rences of various perennial plants on spot sampling plots may result from transport of seed in feces of cattle (Glen- along routes through the study area, and on transects that dening and Paulson 1955). The major cattle trail to bisected more than one substrate, altitudinal zone, or watering tanks on the mountain followed the same path slope exposure, appeared most instructive, and examples as used for sampling plots. are presented in Tables 2 through 6. Similar treatment of data for plots along Transect B-6 The tendency toward Evergreen Forest Associations at (Table 3) demonstrated dispersion of taxa relative to higher elevations adjacent to the Gila River Resource Area changing slope exposure. The transect passed southeast to was demonstrated by data from the northwest slopes of northwest over a low ridge. The first half of the transect Guthrie Peak (Table 2). From sampling plot 1 (1,164 m had a strong representation of Desertscrub and Semi- elevation) through 8 (1,530 m), on the southwest approach desert Grassland species. Within the last (northwest) half, to the mountain, Desertscrub (including Sonoran ele- gradual change occurred toward Evergreen Woodland ele- ments such as blue palo-verde [Cercidium floridurn]) pre- ments, until at the extreme end, Woodland was present. dominated, and it reappeared northeast of the peak be- Tree counts for this transect demonstrated absolute differ- 130 Desert Plants 3(3) AI-M.1171111981 ence even more spectacularly (Table 4). Presence of tions in this feature of vegetation were minimal through- crucifixion-thom (Koberlinia spinosa) on Transect B-6 out most Grasslands and Desertscrub; only extremes are reflected an outcropping of sandy lacustrine sediments in illustrated in Figure 6. valley fill of that area (Heindl 1960). The plant seems Terraces. The term terrace was applied to flat surfaces of characteristic of that type of substrate along the Gila River valley fills, generally adjacent to the mainstream Gila and its major tributaries. River, and less commonly along the San Francisco River Immediately west of the study area, a series of sampling and Eagle and Bonita creeks. These areas enjoy a variably- plots arranged over Solomon Pass (Table 5) demonstrated, thickened surface deposit of finer soils derived from through the occurrence of perennial associates of Grass- adjacent watercourses through wind action and from lands (junipers, sotol, and blue yucca [Yucca baccata]), a surrounding uplands by sheet flow of water during well developed, Semi-desert Grassland (Fig. 2) at higher heavy rainfall. elevations (plots 10-14). This graded downslope into mon- Terrace habitats often were sparsely vegetated (Fig. 7), in tane Desertscrub on both the south and northeast sides. part as a result of their proximity to surface water and Creosote-bush, blue palo verde, and Devil's cholla seemed resultant heavy use by domestic livestock. Livestock use to be sensitive indicators of Desertscrub in that area, dis- may also have enhanced development of extensive growth appearing at about 1,280 m on the south side and reappear- of low, brushy mesquite (Transect GU-6; Table 7), both ing at 1,220 m on the northeastern slope. through providing abundant seed and through browsing, Desertscrub Associations at lower elevations on the the latter of which stimulates development of multiple Gila River Complex (below ca. 1,250 m) were rather uni- stems. formly dominated by creosote-bush, ocotillo, and prickly Water availability on these flat terraces may also effect pear (Opuntia spp.). Sub-dominants included mesquite control over ocurrence of some plants. After heavy rain- and blue palo-verde on slopes in areas of dissected valley fall, water stood in wind-deflated areas for many hours fill and within desert washes. Devil's cholla and Christ- (and sometimes days), presumably as a result of underly- mas cholla (Opuntia leptocaulis) were sub-dominants on ing caliche, cemented conglomerate, or perhaps sealing by more level surfaces (Tables 3 and 6). Various other cacti, clays (see also Zimmerman 1969). Lack of drainage likely Acacia spp., and wolfberry (Lycium spp.) were locally excluded ocotillo, some cacti, and perhaps creosote-bush. common, but rarely dominated. In periods of drought, desert conditions were amplified by Overall foliage volume in upland habitats ranged from resistance of underlying materials to root penetration. relatively high in "piiion-juniper" areas (Fig. 3; Transect Growth form of mesquite on the habitats indicated the B-6), to very low in Desertscrub (Transect GL-5). Varia- latter condition. When rocky outliers of uplands were pre- sent within terraces (Transect GL-6), or where man had disturbed the surface and created a mound or rock - pile, upland Desertscrub species were fax more abundant Table 4. Numbers and percentages of perennial plants per hectare on Transect B-6, Gila River Complex, eastern Arizona.

Transect Segments Table 5. Occurrences of various perennial plants on spot sampling Taxa 0-526 m 527-860m 861-1,050 m Totals plots from south to northeast over Solomon Pass, Graham County, No. % No. % No. "k No. % Arizona, autumn 1978. Symbols: X = present within a plot, and + = present adjacent to a plot. Acacia greggi 86.4 14.25 - - - 43.2 9.68 A. constricta 10.8 1.78 - - 5.4 1.21 Sampling Plots Fouquieria 43.8 7.22 - - - - 21.9 4.91 splendens Taxa 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 Ephedra 5.1 0.84 - - - - 2.5 0.57 Larrea XX XXXXX X XX fasciculata tridentata Perocactus 1.2 0.20 - - - - 0.6 0.14 Prosopis X + X + X spp. juliflora Echinocereus 1.2 0.20 - - - - 0.6 0.14 Acacia spp. X X X X spp. , Cercidium XXXXXX Opuntia spp. 121.3 20.00 61.4 19.74 - - 79.0 17.74 floridum Prosopis 220.3 36.33 106.9 34.36 6.4 2.49 143.5 32.19 Opuntia stanlyi + XXX juliflora Opuntia X X X Lycium spp. 92.1 15.19 31.7 10.19 41.3 16.15 61.6 13.82 spinosior Koeberlinia 11.4 1.88 27.5 8.84 1.6 0.63 14.0 3.13 Ferocactus spp. X X spinosa Simmondsia + X X Zizyphus 3.2 0.53 10.6 3.41 1.6 0.63 5.4 1.21 chinensis obtusifolius Fouquieria + XXXXX Opuntia 1.2 0.20 6.3 2.03 4.8 1.88 3.5 0.78 splendens spinosior Opuntia spp. X X X X X X Dasylirion 1.2 0.20 - - 1.6 0.63 1.3 0.29 + x+ wheeleri luniperus sp. X X Yucca spp. 3.2 1.03 - 1.0 0.22 Lycium spp. Dasylirion X X X X funiperus spp. 5.7 0.94 56.1 18.03 27.0 10.56 25.4 5.70 wheeleri Quercus spp. 3.2 1.03 73.0 28.55 15.6 3.49 Yucca baccata X X Rhus trilobata 4.2 1.35 96.8 37.86 20.6 4.63 Ephedra X Pious edulis - - 1.6 0.63 0.6 0.14 fasciculata Unknown 1.2 0.20 - - - - 0.6 0.14 Opuntia taxon leptocaulus 'Principally 0. phaeacantha 'Principally 0. phaeacantha, some 0. violacea. Minckley and Clark Gila River Resource Area 131

Figure 5. Dissected valley fill with upland Desertscrub, near Transect GU-8, Greenlee County, Arizona; summer 1977. An almost-pure stand of creosote-bush, with some tall yucca on slopes to the right. (Table 6). Overall foliage volume was generally low (Fig. of the study area (see above). Growth forms of mesquite 6), quite comparable to that in most upland Desertscrub along Transect GU-4 included trees and shrubs, rather Associations. than prostrate clumps, increasing foliage volume sub- Desert Washes and Side Canyons. Vegetation of desert stantially over that of upland and terrace Desertscrub washes on the Gila River Complex ranged from a simple Associations (Fig. 6). enhancement of Desertscrub species along broader, Dry Canyon (Transect GL-3) was perhaps more repre- ephemeral watercourses, to development of mixtures of sentative of side canyon habitats of the study area (Table mesquite bosque and broad-leafed riparian communities 7), with greater foliage volume (Fig 6) generally due to in larger, narrower, and typically lower elevation canyons contributions by hackberry (Celtis spp.) and to a lesser where permanent water existed (such as Hot Springs Can- extent, blue palo-verde. yon on Eagle Creek [Fig. 8] and Spring Canyon Transect GL-4 was located in Spring Canyon which en- downstream from Bonita Creek). joyed a short reach of perennial water 1 to 2 km upstream Transect GU-4 represented the first-mentioned extreme from its confluence with the Gila River mainstream. (Table 7). Plants characteristic of upland and terrace Des- Broad-leafed riparian species occurred on the transect; wil- ertscrub Associations were dominant (creosote-bush, low (Salix spp.), walnut (fuglans major), and box elder Acacia spp., and mesquite), and smaller shrubs and cacti (Acer negundo), and a few cottonwood (Populus fremontii) were conspicuously absent or rare in the erosive environ- were present (Table 7). A relatively high overall foliage ment. It is notable that the last smaller species were volume (Fig. 6) resulted from presence of these trees, and common along precipitous margins of this, and other also from abundant hackberry and mesquite, the latter washes and canyons, although they were uncommon on growing on small terraces in bosque formations. Species of transects along canyon floors. Also conspicuous in its ab- true riparian affinities, such as seepwillow (Baccharis sence was blue palo-verde, a Sonoran Desert species gen- salicifolia) and burro-bush (Hymenoclea spp.), that erally absent east of Turtle Mountain (see also Lowe 1955), characteristically inhabit streamside habitats, were rare in but a common upland and canyon species on western parts side canyons with perennial water. As with smaller desert 132 Desert Plants 3(3) Autumn 1981

Table 6. Numbers and percentages of perennial plants per hectare on transects in Desertscrub Associations, Gila River Complex, eastern Arizona. Upland Transects Terrace Transects Taxa B-5 GL-7 GL-6 GU-6 No No No No Larrea tridentata 940.3 88.86 1,206.3 84.50 203.2 36.53 1.0 0.59 Opuntia spp.' 20.0 1.89 27.6 1.93 1.0 0.59 4.8 0.86 0. stanlyi 8.9 0.84 11.4 0.80 65.7 11.80 0. leptocaulus 1.9 0.18 10.5 0.74 2.5 0.45 0.3 0.90 0. spinosior 1.0 0.09 3.8 0.27 3.2 1.90 Prosopis juliflora 6.7 0.63 9.2 0.64 143.8 25.85 111.1 65.90 Lycium spp. 3.8 0.36 1.6 0.11 66.0 11.87 2.5 1.48 Ce/tis spp. 1.0 0.59 Ephedra fasciculata 0.3 0.03 1.0 0.07 1.0 0.59 Echinocereus spp. 0.3 0.03 1.6 0.11 6.0 1.08 Ferocactus spp. 1.0 0.09 3.5 0.24 1.9 0.34 0.6 0.36 Cereus greggi 0.6 0.04 Mammillaria spp. 0.3 0.02 0.3 0.05 Fouquieria splendens 74.0 6.99 132.4 9.27 33.3 5.99 Acacia spp. 7.3 0.51 13.0 2.34 46.6 27.65 Zizyphus obtusifolius 1.3 0.77 Cercidium floridum 10.1 0.71 14.6 2.62 'Principally 0. phaeacantha, some 0. violacea.

Table 7. Numbers and percentages of perennial plants per hectare in riparian vegetatin was mesquite. Within the present study side-canyon habitats of the Gila River Complex, eastern Arizona. area, this rose to 56% of 637 ha, and upstream (from Gut- hrie to the New Mexico Boundary), mesquite comprised Transects 43% of 780 ha. About 66% (234 ha) of mesquite stands Taxa GU-4 GL-3 GL-4 delineated within the Gila River Resource Area had No. % No. % No. % greater than 50% ground cover (36% > 75% ; Lacey et al. Prosopis juliflora 38.6 9.27 20.6 12.70 48.0 43.63 1975). Celtis spp. 61.7 28.10 16.6 15.06 Bosque habitats included trees 2 to ca. 12 m in height Acacia greggi 60.3 14.49 29.1 18.00 16.9 15.32 A. constricta 68.6 16.47 - - - - (Figs. 9-10). Reproduction (as indicated by trees < 2 m Cercidium floridum 23.4 14.50 4.0 3.64 high) was present on all bosques sampled (Table 9), al- Salix spp. - - 0.6 0.52 though variation in the proportions of various-sized trees fuglans major - - 0.9 0.78 Acer negundo was considerable. - - - - 0.3 0.26 bosque understory was variable and Larrea tridentata 232.6 55.90 22.3 13.80 13.4 12.21 Within habitat, Lycium spp. 12.9 3.09 - - 7.1 6.49 often seasonally absent as a result of grazing by domestic Ephedra fasciculata 3.1 0.70 4.0 2.50 - - livestock. A dense ground cover of annual forbs (princi- Opuntia spp.' - - 0.8 0.10 2.3 2.08 pally Cruciferae) and grasses may be present after pro- Yucca elata 0.3 0.07 - - - - longed or extensive rainfall. Those bosques which were 'Principally 0. phaeacantha, some 0. violacea. protected from livestock by cut banks or other obstacles shrubs and cacti, the erosive nature of the environment had understories of thick grasses (Sporobolus sp. and oth- subject to severe scour during flash floods, was too harsh ers) in open areas and along their margins. for establishment of permanent stands. Temporary popu- Photographs published by Olmstead (1919), information lations often were present, however, increasing diversity on flooding (Minckley 1979), and data evaluated by Turner in these habitats for periods of months or years. (1974), all indicate that much of the riparian vegetation Mesquite llosques. Mesquite made up 95 to 99% of total along the Gila River, including mesquite bosques, was de- trees counted on four of the permanent transects along the stroyed in the period 1888-1917. Stands of mesquite along mainstream Gila River (Table 8), comprising the major the mainstream Gila River rarely contained trees greater plant community on recent alluvial deposits along the than 40 cm in diameter, and most were less than 35 cm. stream. Bosques also occurred on fine-grained soils along Gavin (1973) conservatively estimated growth to ca. 25 cm the San Francisco River, Eagle and Bonita creeks, and in diameter in 60 years on floodplain sites in Arizona. The the larger side canyons. Minor components of this associa- degree of development of bosque vegetation, and sizes of tion were hackberry (principally Celtis reticulata, with trees present, correspond well to an age of 60 to 90 years lesser amounts of C. pallida), and wolfberry. Other woody for the habitat. plants collectively made up less than 1.0% of the totals Leaf volume was generally comparable on all mesquite recorded on transects. By frequency of occurrence in 124, bosques sampled (Fig. 6), and was invariably greatest a 150- to 200-m2 sampling plots on transects, mesquite was distance above the ground surface. Height of maximum at 100%, Lycium spp. 14.5%, Celtis spp. 12.1%, and all leaf volumes related well to relative age of the various other species occurred at less than 5%. stands, with Transects GL-2 and GU-2 having maxima at Lacey et al. (1975) produced maps of natural riparian 4 to 6 m (less than 10% trees less than 2 m in height; Table vegetation along the Gila River that also clearly demon- 9), and GL-1 and GU-1 below 4 m (more than 19% small strate dominance of mesquite. From Solomon to near the trees). mouth of Bonita Creek, about 26% of 465 ha of natural The sharp change from Desertscrub Associations of dis- Terrace - Upland 2 D4 0 - k Mesquite I 20 , Basque 20 - I P non-Juniper Broad-I eaf ed GU1 R i par' an I ' I

1 0 1 0 D1

Si de Canyon • ' GU2 HEIGHT ABOVE GROUND, m GL1

0 0 .66 33 66 0 .33 0 .33 LEAF VOLUME, m2 / m-3

Figure 6. Overall foliage volume in various habitats of the Gila River Resource Area, eastern Arizona, summer 1978; see text for further explanation and discussion.

Figure 7. Terrace habitat in foreground, mostly barren but with scattered wolfberry, mesquite, and creosotebush, Transect GL-5, Graham County, Arizona; summer 1977. The low ground cover near the lower part of the slope (center foreground) is Devil's cholla.

Table 8. Numbers and percentages of perennial plants per hectare in mesquite bosque habitats of the Gila River Complex, eastern Arizona.

Transects Taxa GU-1 GU-2 GL-I GL-2 Totals No. No. No. No. No. Prosopis iuliflora 158.9 96.53 119.1 94.13 503.2 98.82 243.3 96.08 256.1 97.19 Celtis spp. 1.7 1.04 0.6 0.45 2.2 0.44 7.1 2.82 2.9 1.10 Acacia greggii 0.3 0.17 0.3 0.06 0.2 0.07 Cercidium floridum 0.3 0.06 2.0 0.78 0.6 0.23 Populus fremontii 3.1 1.91 0.8 0.30 Salix spp. 1.7 1.35 0.4 0.15 Platanus wrightii 0.6 0.12 0.2 0.07 Lycium spp. 0.6 0.35 5.1 4.06 2.5 0.50 0.8 0.31 2.3 0.87 sected valley fill to mesquite bosque (Fig 10) was equalled Table 9. Numbers per hectare and percentages in parentheses) of in many segments by a similarly abrupt streamside change different-sized mesquite in riparian bosques along the Gila River in vegetation (Fig. 11). Dense thickets of almost-pure mainstream, Gila River Complex, eastern Arizona, summer 1978. seepwillow were present on low, saturated sandbars im- Height in m GL-1 GL-2 GU-1 Gu-2 mediately along the river, often in association with thick <2 95.87 16.67 65.14 4.28 stands of Bermuda grass (Cynodon dactylon). These plants ( 19.05) ( 6.85) (41.01) ( 3.60) are remarkably resistant to scour during flooding 2 - 9 406.03 218.25 92.00 98.00 (Campbell and Green 1968, Fisher and Minckley 1978), ( 80.69) ( 89.72) (57.91) (82.95) and perform an efficient armoring function, protecting the >9 1.27 8.33 1.71 16.86 soft alluvium of the bosque community. Seepwillow ( 0.25) ( 3.43) ( 1.08) (14.15) rapidly colonized barren sandbars along the river, forming borders that tended to protect scoured areas, enhanced Table 10. Occurrences of various perennial plants on paired spot sam- accumulation of windblown and water-borne sands, pling plots through broad-leafed riparian habitat of Bonita Creek (downstream to upstream, from the mouth), Graham County, Arizona, and promoted germination of other, more xeric-adapted autumn 1978. Symbols: L= occurrence in a plot on left side of tran- plants such as mesquite. sect; R occurrence in plot on right side of transect ; X = present in It is interesting that species such as arrowweed (Tessaria both plots. sericea), and especially the introduced salt-cedar (Tamarix Sampling Plots chinensis), were absent or very rare on the Gila River Taxa 1 23 Complex. The first species is an important immediate- Prosopis XXXXXXXXXXXXXRXLXXRXXXL riparian plant through much of the lower Colorado River juliflora basin (Turner 1974, Ohmart et al. 1977) that was not found Celtis spp. R R X LRXXXXRLLRXXXXRRLRL within the study area, and the latter has replaced much of Hymenoclea spp. X X XX LXRXRX XRRRXXLXXLX the natural riparian vegetation along southwestern Baccharis X X XXLL RR XXRX XXX R salicifolia streams (Robinson 1965, Haase 1972, Turner 1974, Horton Lycium spp. R RRX L RXLLLRXX 1977), but has yet to become established other than as Acacia greggii R RX L LL L LLR R individual plants in the reach of Gila River between the Salix spp. L R San Francisco River and the mouth of Bonita Creek. An Platanus wrightii XRLLLLLRR RR LXXRX Populus XRR RL abundant seed source is available on the San Francisco fremontii River near Clifton, and on the Gila River upstream from luglans major LXR L Guthrie, and germinating seedlings have been found on Zizyphus L L X R obtusifolius sand bars within the study area. Seedlings that we ob- Cercidium R R served were invariably destroyed by flooding. The species floridum has not invaded Bonita Creek, Eagle Creek, or other per- Acacia constricta Vitis arizonica ennial canyon tributaries (Hot Springs and Spring canyons) Larrea tridentata R X LXR as other than individual plants, and is rare along the San Nicotiana spp. LRRX Francisco River downstream from Clifton. Perhaps can- Opuntia spp.' R X LXLLXLLL R yons of the Gila River Complex are too erosive to allow Simmondsia XLX LLX chinensis establishment of this pest (see e.g. Turner and Karpiscak Fouquieria P. LL L 1980). It is notable that Robinson (1965) earlier recorded splen dens Dasylirion LL LLL absence of salt-cedar from these areas, and from other wheeleri canyon-bound streams such as the Salt River and Aravaipa Ephedra Creek, Arizona, canyons of the San Francisco River above fasciculata Chilopsis XRR Clifton, Arizona, and in New Mexico, and the Gila River linearis Canyon near Redrock, New Mexico. Ferocactus spp. P. Broad-leafed Riparian. Riparian vegetation within the Acer negundo R LL spp. L X Gila River Complex, in composite, resembled that de- funiperus Opuntia R L X scribed by the Interagency Task Force (1974) along the spinosior mainstream Gila River from the Gila Wilderness Area to Carya sp. P. the New Mexico-Arizona line. Species diversity was (domestic) Rhus sp. greatest at higher elevations, and declined along the lower Echinocereus P. reach. Near the Gila Wilderness, riparian communities spp. were dominated by large broad-leafed species such as cot- 'Mostly 0. phaeacantha, some 0. violacea. Figure 8. Side-canyon habitat, Hot Springs Canyon, tribu- tary to Eagle Creek, Greenlee County, Arizona; summer 1977. Plants in foreground are burro-bush, with mesquite along canyon walls.

Figure 9. Mesquite bosque, Transect GL-1, Graham County, Arizona; autumn 1978. Aggradation of this ter- race along the Gila River buried trunks of these mesquite with as much as 2.5 in of fine sand and silt. 136 Desert Plants 3(3) Autumn 1981

Table 11. Numbers and percentages of perennial plants per hectare in broad-leafed riparian habitats of the Gila River Complex, eastern Arizona.

Transects Taxa B-1 B-2 B-3 Totals No. No. No No. Prosopis juliflora 281.6 76.73 53.7 35.47 147.6 78.55 161.0 68.45 Celtis spp. 42.9 11.68 66.6 43.96 10.2 5.41 39.9 16.96 Acacia greggii 2.9 1.89 - - 1.0 0.43 Cercidium floridum 1.3 0.35 - - 0.4 0.17 fuglans major 3.2 0.87 3.1 2.08 1.9 1.01 2.7 1.15 Populus fremontii 1.6 0.43 0.9 0.57 0.3 0.17 0.9 0.38 Salix spp. 4.1 1.12 2.9 1.89 2.5 1.35 3.2 1.36 Platanus wrightii 5.7 1.56 4.1 2.00 9.7 6.42 6.5 2.76 Zizyphus obtusifolius 13.7 3.72 16.2 8.61 10.0 4.25 Chilopsis linearis 0.3 0.17 0.1 0.04 Lycium spp. 11.4 3.11 11.7 7.74 2.5 1.35 8.5 3.61 funiper monosperma 1.3 0.68 0.4 0.17 Fendlera rupicola 0.3 0.17 0.1 0.04 Ficus sp. (domestic) 1.6 0.43 - - 0.5 0.21

Table 12. Some indicators of the Sonoran and Chihuahuan deserts of precluded access to much of this type of habitat, and al- North America (Lowe 1967), and their occurrence in the Gila River tered or destroyed a large percentage of it along Eagle Complex, eastern Arizona. Creek and the San Francisco River. Impact of flooding on Sonoran Gila River Chihuahuan riparian communities along the mainstream Gila River Species Desert Complex Desert were equally severe, with many bosques simply being re- Ambrosia deltoidea moved, but none of these events have been other than Cereus giganteus qualitatively assessed. Olneya tesota Plant communities along Bonita Creek were diversified Cercidium microphyllum (Fig. 12). With a total of 29 taxa being identified in 46 Cercidium floridum holacantha quadrats arranged at 1-km intervals from the mouth to the Encelia farinosa San Carlos Apache Indian Reservation (Table 10). Thirteen Simmondsia chinensis kinds of plants occurred at more than 15% frequency. Acacia constricta Mesquite dominated throughout the transect (91.3% fre- Fouquieria splendens quency of occurrence), followed in decreasing order of Larrea tridentata Opuntia phaeacantha abundance by burro-brush, hackberry, and seepwillow. Prosopis juliflora Large, broad-leafed tree species were dominated by syca- Koeberlinia spinosa , more (Platanus wrightii), which occurred in 43.5% of the var. spinosa Rhus microphylla 46 plots, followed by cottonwood and walnut (15.2% Agave lechugilla each), willow (10.9%), and box elder (6.5%). Flourensia cernua 3 A relatively high incidence of plants characteristic of 3 Mortonia scabrella Desertscrub and Desert Grassland (Table 10; cacti, creo- 'May extend slightly into the eastern edge of the Sonoran Desert. A sote- bush, sotol, etc.), was in part a reflection of proximity Sonoran Desert variety (tenuispina) is found far west of the eastern of canyon walls to the stream. Arid-land species did, how- edge of the Sonoran Desert (Kearney and Peebles et al. 1973). ever, occupy drier parts of terraces, especially where dis- 2Indicator species not present in Arizona's Chihuahuan Desert. turbed by removal of large, riparian trees. 3Indicator species for Arizona's Chihuahuan Desert. Zonation of species relative to perennial flow was ap- parent both laterally and longitudinally within the system tonwood and sycamore, as along upper Bonita Creek in our (see also Zimmermann 1969). Mesquite, wolfberry, and study area, supplemented by stands of hackberry, walnut, Acacia spp., and hackberry, walnut, and box elder, for box elder, and willow. Along the middle portion of the example, were present on small terraces (first three), or Gila River reach, cottonwood and willows were sub- along canyon walls or near inflows of ephemeral water- dominant, and a substantial increase in seepwillow was courses (last three). Sycamore was generally on terraces or recorded. This condition was present in much of the Eagle alluvial fans from side canyons. Burr-brush and desert wil- Creek Canyon within the Gila River Resource Area. Gal- low (Chilopsis linearis) were on open, boulder-strewn leries of cottonwoods, such as described by the Inter- areas where scour was severe during floods, but flow was agency Task Force near Cliff, Gila, and Redrock, New subsurface in periods of drought (center of the channel or Mexico, were locally present along the San Francisco mid-way in the reach). More hygric species, seepwillow, River, and less so along the Gila River mainstream (in cottonwood, and willows, occurred generally only where part, Lacey et al. 1975) in Arizona. Mesquite bosques permanent surface waters were present (see especially formed the principal vegetation along the Gila River seepwillow, which became rare in the center of the reach within the study area (Fig. 10). [Table 10] where water flowed beneath the substrate dur- Quantative data on broad-leafed riparian communities ing drought). were obtained from Bonita Creek, and were supplemented Tree counts on three transects within Bonita Creek by observational information. Severe flooding in 1978-79 Canyon provided similar data (Table 11). Mesquite made Figure 10. Mesquite bosque along the Gila River, includ- ing Transect GL-1, Graham County, Arizona, near the western margin of the study area; summer 1977. Much of this bosque was removed by flooding in winter 1978-79.

Figure 11. Gila River just upstream from the old Safford- Morenci highway bridge, Greenlee County, Arizona, in- cluding Transect GU-2; summer 1977. Seepwillow forms a dense border for thick mesquite along bluffs bordering the river and on low terraces in the mouths of side can- yons. Figure 12. Bonita Creek, Graham County, Arizona, near Transect B-2; summer 1977. Seepwillow borders the erosive channel, backed by tall cottonwood with an understory of mesquite, hack , and other lower trees. A few sycamore are present in right background.

Figure 13. Cottonwood gallery with mesquite understory at a wide place along Eagle Creek Canyon, Greenlee County, Arizona, about 1.6 km north of Bat Cave; summer 1979. Minckley and Clark Gila River Resource Area

Figure 14. San Francisco River, Greenlee County, Arizona, at Evan's Point; summer 1977. Note girdling of cotton- . wood in foreground by beaver, seedling seepwillow on sandbar across stream, and xeric association of mesquite on rocky terrace in background. up an average of 68.5% of all trees present, followed by often exceeded 30 m in height, presumably in response to hackberry. Large, broad-leaved trees other than hackberry the narrow gorge, but also reflecting a greater maturity made up only 0.9% (cottonwood) to 6.5% (sycamore), of (less frequent scour) of parts of the community. Reproduc- all trees counted on the transects, but produced substan- tion by large broad-leafed trees (cottonwood and sycamore) tial volumes of foliage higher than 10 m above the ground is rarely successful in Gila River complex (USBLM 1978), (Fig. 6). Presence of large numbers of mesquite, hackberry, largely as a result of concentrated grazing in riparian zones and other intermediate-sized trees produced substantial by domestic livestock. leaf volume concentrated between 2.5 and 10 m, however, Riparian communities along the San Francisco River and small shrubs such as gray-thom (Zizyphus obtusifolia) were far more sparse than those of Eagle and Bonita creeks, and wolfberry created a dense understory. with cottonwood occurring as single trees and relatively Gallery forests of cottonwood were more fully de- small gallery stands, and the principal vegetation being veloped along Eagle Creek than elsewhere in the Gila mesquite and more xeric plants on low terraces (Fig. 14). River Resource Area (Fig. 13), and associated species were Along the Gila River, cottonwood-willow associations comparable in diversity to that described for Bonita Creek comprised only a small proportion of total area (5.7% ) of Canyon. Eagle Creek Canyon is far larger than that of riparian communities mapped by Lacey et al. (1975) in the Bonita Creek, deeper and with a broader floor commensu- reach from Solomon, Arizona, to the New Mexico— rate with its larger drainage area, so that riparian vegeta- Arizona Boundary. We consider the total of 107.2 ha indi- tion was well developed to include marginal bosques of cated to be of this vegetation type as too great; since those mesquite in wider places and in the mouths of side can- authors did not include areas of less than 0.405 ha in size yons. Flooding in 1978-79 removed many of these bos- in their compilations, and many such places were found in ques, scouring the canyon from wall to wall. Cotton- our ground survey to support only one to a few, large indi- woods, sycamores, and other trees in Eagle Creek Canyon vidual cottonwoods or willows. 140 Desert Plants 3(3) Autumn 1981

Summary Horton, J. S. 1977. The development and perpetuation of the Perennial vegetation of the Gila River Resource Area is permanent tamarisk type in the phreatophyte zone of the at present dominated by Desertscrub associations. Transi- southwest. In, R. R. Johnson and D. A. Jones, Tech. Coord., tion between the Sonoran and Chihuahuan deserts is Importance, preservation, and management of riparian demonstrated in the region by the presence of a number of habitat: a symposium. USDA For. Serv. Gen. Tech. Rept. taxa shared by both, side by side with some mutually ex- RM-43: 124-127. Survey of the Gila River, New clusive elements (Table 12). Semi-desert grasslands clothe Interagency Task Force. 1974. Mexico. New Mexico Game and Fish Department, Santa Fe, higher slopes, with a scattering of interior chaparral ele- New Mexico. Unpubl. ments growing locally in disjunct stands. Smaller chapar- Kearney, T. H., and R. H. Peebles, and collaborators. 1973. ral species include California buckthorn, silk-tassel, Arizona Flora. 2nd Ed. with Suppl. by J. T. Howell, E. McClin- squaw-bush, and mountain-mahogany, to name a few (see tock, and collaborators. Univ. Calif. Press, Berkeley and Los McGill 1979). Angeles (4th printing). Deciduous riparian woodlands form a conspicuous and Lacey, J. R., P. R. Ogden, and K. E. Foster. 1975. Southern Arizona important mesic environment through arid areas, acting riparian habitat: spatial distribution and analysis. Off. Arid as connecting links among and between points of higher Lands Stud. Univ. Ariz., Bull. 8: i-viii, 1-148. elevation, and from such places to permanent lowland Lapham, M. H., and N. P. Neill. 1904. Soil Survey of the Sol- omonsville Area, Arizona. Field Oper. Bur. Soils, 1903, USDA, streams. Steep, narrow, side canyons leading to mesic Washington, D. C. gorges of larger watercourses provide further diversity by Lehr, J. H. 1978. A catalogue of the Flora of Arizona. Desert supporting riparian vegetational elements intermixed Botanical Garden, Phoenix, Arizona. with woody plants more characteristic of terrace and up- Lowe, C. H. 1955. The eastern limit of the Sonoran Desert in the land habitats (including some chaparral elements). Higher United States, with additions to the known herpetofauna of altitude forms as alder (Alms oblongifolia), box elder, and New Mexico. Ecology 36: 343-345. others, penetrate far into desert along narrower canyons, (ed ) 1967. The Vertebrates of Arizona. Univ. Ariz. sharing habitats with more characteristic desert ripar- Press, Tucson (Second Printing). ian species as cottonwood and Arizona sycamore. The McGill, L. A. 1979. Vascular flora. In, W. L. Minckley (ed.). Re- superficial monotony of desertscrub is thus marked source Inventory for the Gila River Complex, Eastern Bur. Land Mgmt., Phoenix, Ariz. with distinct, thin threads of larger riparian vegetation, Arizona. Final Rept. USDI Pp 56-85. which increase diversity greatly out of proportion to Martin, W. P., and J. E. Fletcher. 1943. Vertical zonation of great their physical areas. soil groups of Mt. Graham, Arizona, as correlated with cli- Literature Cited mate, vegetation, and profile characteristics. Univ. Ariz., Anderson, B. W., R. W. Engel-Wilson, D. Wells, and R. D. Ohmart. Agric. Exp. Sta. Tech. Bull. 99: 89-153. 1977. Ecological study of southwestern riparian habitat: tech- Minckley, W. L. 1979. Hydrology and Climate. In, W. L. Minckley niques and data applicability. In, R. R. Johnson and D. A. (ed.), Resource Inventory For the Gila River Complex, Eastern Jones, Tech. Coord., Importance, Preservation and Manage- Arizona. Final Rept. USDI Bur. Land Mgmt., Phoenix, Ariz. ment of Riparian Habitat: A Symposium. USDA For. Serv. Pp. 17-37. Gen. Tech. Rept. Rm-43: 146-155. Ohmart, R. D., W. 0. Deason, and C. Burke. 1977. A riparian case Brown, D. E., and C. H. Lowe. 1978. Biotic Communities of the history: the Colorado River. In, R. R. Johnson and D. A. Jones, Southwest. USDA For. Serv. Gen. Tech. Rept. RM-41: map. Tech. Coord., Importance, Preservation and Management of Burkham, D. E. 1970. Precipitation, streamflow, and major floods Riparian Habitat: A Symposium. USDA For. Serv. Gen. Tech. at selected sites in the Gila River drainage above Coolidge Rept. RN1-43: 35-47. Dam, Arizona. U.S. Geol. Surv. Prof. Pap. 655-B: 1-33. Olmstead, F. H. 1919. Gila River flood control-a report on flood Campbell, C. J., and W. Green. 1968. Perpetual succession of control of the Gila River in Graham County, 65th Congr., 3rd stream-channel vegetation in a semiarid region. 1. Ariz. Acad. sess., Sen. Doc. 436: 1-94. Sci. 5: 86-98. Robinson, T. W. 1965. Introduction, spread, and areal extent of Clarkson, R. W. 1979. Geology, topography, and drainage de- salt-cedar (Tamarix) in the western states. U.S. Geol. Surv. velopment. In, W. L. Minckley (ed.), Resource Inventory For Prof. Pap. 491-A: 1-12. the Gila River Complex, Eastern Arizona. Final Rept. USDI Treadwell, B. D., D. A. Mouat, J. S. Conn, and C. M. Sawtelle. Bur. Land Mgmt., Phoenix, Ariz. Pp 7-17. 1976. Vegetation Type Map of the Upper Gila-San Simon Dis- Fenneman, N. M. 1931. Physiography of the Western United trict. Office of Arid Lands Studies, Univ. Ariz., Tucson. States. McGraw-Hill Book Company, New York. Turner, R. M. 1974. Quantitative and historical evidence for vege- Fisher, S. G., and W. L. Minckley. 1978. Chemical characteristics tation changes along the upper Gila River, Arizona. U.S. Geol. of a desert stream in flash flood. I. Arid Environ. 1: 25-33. Surv. Prof. Pap. 655-H: 1-20. Gavin, T. E. 1973. An Ecological Survey of a Mesquite Bosque. Turner, R. M., and M. M. Karpiscak. 1980. Recent vegetation Unpubl. M.S. Thesis, Univ. Ariz., Tucson. changes along the Colorado River between Glen Canyon Dam Glendening, G. E., and H. A. Paulsen, Jr. 1955. Reproduction and and Lake Mead, Arizona. U.S. Geol. Surv. Prof. Pap. 113: establishment of velvet mesquite as related to invasion of 21-125. semi-desert grasslands. USDA For. Serv. Tech. Bull. 1127: U.S. Bureau of Land Management. 1978. Upper Gila-,San Simon 1-50. Grazing Environmental Statement. USDI Bur. Land Mgmt., ' Green, C. R., and W. D. Sellers (eds.), 1964. Arizona Climate. Phoenix, Ariz. Univ. Ariz. Press, Tucson. U.S. Soil Conservation Service. 1975. Supplement to Arizona Haase, E. F. 1972. Survey of floodplain vegetation along the lower General Soil Map. USDA Soil Cons. Serv., Phoenix, Ariz. Gila River in southwestern Arizona. I. Ariz. Acad. Sci. 7: Unpubl. 75-81. Zimmerman, R. C. 1969. Plant ecology of an arid basin. Tres Heindl, L. A. 1960. Geology of the lower Bonita Creek area (a Alamos-Redington area, southeastern Arizona. U.S. Geol. preliminary report). Ariz. Geol. Soc. Digest 3: 35-39. Surv. Prof. Pap. 485-D: 1-51.