Life-History Characteristics of the Creosotebush, Larrea tridentata

•-

' - /ffl ^'•s

•

w^m^^r^ Z&SBH&*Z^WSa&:" - '• *mi 'N-">&5A&S*INP\- J

NEW MEXICO SI All IMVl.Rsl |\ A AGRICULTURAL EXPERIMENT STATION BULI.l ll\ 526

CON I 1 VIS

\( knowlcdgincnts 2

Introduction 3

Methods 4

Description ol experimental areas 4 Fi nit production 5 Carpel fill 5 Viability 5 Influence ol site on (reosotebush establishment 5 Influence of soils on emergence and early growth of creosotebush seedlings 6 Influence of range condition on creosotebush establishment 6 Natural field establishment 6 Growth and maturation 7 Death loss and borei damage 7 Results and discussion 7 Fi uil production 7 Carpel fill 11 Seed viability and longevity 12 Influence of site on creosotebush seedling emergence 11 Influence of soil on emergence, damping-off, and growth of creosotebush seedlings 14 Influence of range condition on creosotebush establishment 16 Natural establishment ol < reosotebush as influenced 1>\ proximity and number of seed source plants and soil stability 21 Growth and Ii nit production ol naturally established creosotebush seedlings 24 Death, borer damage, and down (IKay in creosotebush 26 Summary and conclusions 29 Literature cited 31

October 1968 Las Cruces, New Mexico Acknowledgments

Work reported herein was done under a New Mexico State University Agri­ cultural Experiment Station project contributing to Western Regional Co-op­ erative Project W-25, "Ecology and Improvement of Brush Infested Range- lands." Financial support from the regional project is acknowledged. In addi­ tion, acknowledgment of aid in the work is given to the following co-workers at New Mexico State University: Miss Elizabeth McSwain, State Seed Analyst, for use of facilities; Dr. Eugene Staffeldt, Professor of Biology, for aid in evaluation of damping-off damage to seedlings; Dr. Thomas E. Smith, former Professor of Botany and Entomology, for microbiological examination of creosotebush crowns; Dr. Gerald L. Nielson, State Department of Agriculture, for aid in identification of insect larvae; Dr. M. D. Finkner, Professor of Biometrics, for aid with statistical analyses; and Dr. }. U. Anderson, Associate Professor of Agronomy, for identification of soils. Life-History Characteristics of the Creosotebush, Larrea tridentata

K. A. Valentine and J. B. Gerard1

Creosotebush {Larrea tridentata soil (Cooperridei and Hendricks, 6; (DC.) Coville) aggressively invades Gardner, 10; Leithead, 14). Jardine desert plains grassland in Aii/ona. and Hunt (12) reported extensive New Mexico, and western Texas. areas ol black grama (Bouteloua erio- Gardner (10) recorded invasion ol poda (I'orr.) Torr.) grassland on the grassland along the Rio Grande Val fornada Experimental Range, as ley in southern New Mexico and shown In range survey in 1915. These pointed out that it takes the form of areas were entirely replaced by creo­ frontal advances along the margins ol sotebush by 1934, a scant 20 years later. the largei creosotebush communities C.infield (I) recorded basal cover of and ol single plants and then small bl.uk giam.i ol 1,127 square centime­ outlier stands at varying distances ters pei square metei (11.27 percent) from the main bodies of the larger for the period 1925-27 on a remnani communities. Yang (21) recorded in­ of this grassland. Average herbage vasions of grassland on upland plains yield of black grama, clipped to stub­ adjacent to the Rio Grande Valley. ble heights ol one and two inches, Buffington and I In bel (2), using 1858 was 768 pounds pel acre foi the same General Land Office survey records years. Bv 1940, the last ol this remnant and recent observations, reported in­ stand had disappeared, leaving the crease ol creosotebush on the Jornada site dominated bv creosotebush with Experimental Range in southern New little or no herbaceous ground cover. Mexico from (il 1 a< res ol moderatel) I bus the shilt from productive and invaded range (creosotebush compris protective perennial grass cover to the ing 15 to 55 percenl ol the perennial creosotebush type may be rapid and vegetation) in 1858 to 12,388 acres of extreme. heavily invaded range (55 to 1 *>*> per­ Accompanying the reduction in cent ol the vegetation) in 1963. Dick- perennial grass cover is a great loss in Peddie (8) found from these earl) grazing capacity ol the range. Paulsen General Land Office survey records and Arcs (15) reported a decrease in that level and gently rolling land in capacit) from 7.1 to 3.7 animal-unit southern New Mexico generally was veais pei section ovei a 25-year period covered with grama grass and that in a creosotebush-tarbush (Flourensia creosotebush was restricted to foot­ cernua DC.) pasture on the Jornada hills and outlying well-drained knolls. Experimental Range. They attributed Damage to grassland range follow­ ing invasion by creosotebush takes the form of reduced perennial grass cover and yield and increased erosion of the loss in capacity to continuous en­ whether by grazing management or croachment of the brush. While this other range management practices, was a large loss in capacity, it does not and the choice of control methods represent the loss which generally ac­ also depend, or should depend, upon companies the full development of these same and other autecological the creosotebush community. This characteristics. The work reported commonly attained development here was designed to determine cer­ brings about the loss of practically tain of these characteristics which 100 percent of original grazing ca­ have a bearing on the invasion of pacity. grassland range by creosotebush and The capacity of any noxious plant on the control of creosotebush on in­ species to invade and dominate a vaded rangeland. More specifically, range and the time and spatial pat­ the objectives of the work were to de­ terns involved are in large part re­ termine fruit and seed production, lated to the autecological characteris­ viability, establishment, and growth tics of that species, especially its fruit­ characteristics of creosotebush. These ing, dissemination, establishment, are, of course, only a few of the many growth, and competitive characteris­ characteristics which constitute the tics. The susceptibility to control, complete autecology of the species.

Methods

June, and an average maximum of Description of Experimental Areas 56°F is reported for January. Ex­ tremes of —20° and 109°F have been The field work was done on or near recorded at the Jornada Experimental New Mexico State University's Agri­ Range headquarters (16, 17). Depar­ cultural Experiment Station Ranch tures from mean annual and mean in Dona Ana County. The ranch lies growing season precipitation are fre­ on the southern part of the Jornada quent and marked, and drought is Plain and in the bordering hills and common. mountains between the plain and the Soils of the southerly part of the Rio Grande River on the west. This plain on which the experiment sta­ part of the plain is nearly level to tion ranch is located are mainly sandy gently undulating and varies in eleva­ loams underlain by indurated calcium tion between 4300 and 4400 feet (20). carbonate at depths varying from a The climate of the area is arid. few inches to 30 inches and more. Paulsen and Ares (15) report the 38- The vegetation of the lower Jor­ year average annual precipitation at nada Plain consists of grass and brush the Jornada Experimental Range types. The grass types consist in the headquarters, representative of the main of the black grama-mesa drop- central southern part of the plain, to seed (Sporobosns flexuosus (Thurb.) be 9.02 inches. An average of 4.74 Rydb.) climax and various succes- inches, slightly over half the annual sional stages of this climax indirectly total, falls during July, August, and resulting from the extremely damag­ September, the main growing season ing drought of the early 1950's. To- on the range. An average maximum bosa (Hilaria mutica (Buckl.) Benth.)- temperature of 97°F is reported for burro-grass (Scleropogon brevifolius l'hil) types dominate the more re was collected neai the University cam­ stinted lowland swales on the plain. pus, about 20 miles south of the The brush types consist in the main ranch. Aftei collection, all material of separate stands of mesquite (Pro- was dusted with DDT to control in­ sopis juliflora (Swartz) DC. var. glan- sects and stored in the laboiatoiv in dulosa) and creosotebush, but ovei paper bags. Tests were conducted in limited areas these two species occur the New Mexico State Seed Labora- in mixed stands. tory, in 17 C constanl temperature, humidified germinators. Foui lots of Mi i II pels ea( h were used in ea( h test. Fruit Production I he carpels were placed in pen i dishes on standard germination blot- Fruit production was determined tei papei moistened with distilled on plants in the creosotebush vegeta­ water. 'I'he carpels were dusted with tion type and in grassland, both on Spergon to control fungus growth. the station ranch. Groups of 10 full- I he tests were i un foi 60-day pei iods. grown plants each were delineated at After each test, the carpels with no three locations in each of these vege­ germinated seed were opened to de­ tation types. Each fall from 1955 to termine whether any seed was present. 1962, except 1956, a typical branch Germination percentage was com­ was collected from each plant, the puted on the basis ol the filled carpels. fruits were removed and counted, and the branch was oven-dried and weighed. Fruit production is reported Influence of Site on Creosotebush as number per 100 grams of branch Establishment weight. Upon completion of these determinations, all leaf-bearing Ihis experiment was conducted to branches and portions of branches of determine the influence of site char­ each plant in the several groups were acteristic on establishment of creoso­ dipped, oven-dried, and weighed loi tebush. Three sites were included in an estimate of fruit production per the study: one a calcareous variant ol plant. l'alnia loamy sand with caliche out- dops at the top of a gently sloping ridge, occupied l>\ a fully developed Carpel Fill creosotebush community; the second, Simona sand) loam on an upland Carpel fill, or seed set, was deter­ plain, occupied In good condition mined by examination of carpels used black grama grass cover; and the in the germination tests after these third, a deep (lav loam in an inter­ tests were completed and bv direct mittently lightly flooded swale, occu­ examination of carpels of fruits col- pied by fair to good tobosa grass h (ted for the fruit production study. cover. Five one-yard-square grids were established on each site in June or early July in 1956 to 1959, inclusive. Two carpels were planted at depths Viability of 14 to % inch at each of 100 points Viability was determined through (10 tows of 10 points each) on each germination tests of material collected grid. The planted carpels were the on the ranch in November, each year heavier ones remaining after light from 1955 to 1961. In 1954, material carpels were separated by flotation on water, (.rids were examined for seed­ expanded to include two additional ling emergence and survival each fall locations. At each location in each from 1956 to 1963, except 1961. year, 10 transects were established- five in very poor condition range and five in adjacent good condition range. Influence of Soils on Emergence Condition was determined by basal and Early Growth of Creosotebush cover of desirable perennial grasses— 0.15 percent on the very poor condi­ Seedlings tion range and 1.71 percent on the This was a pot test conducted in good condition range. Each transect the greenhouse. Soils from six sites on consisted of a 200-foot line marked by and in the vicinity of the station stakes at each end. Seed planting ranch were used. These soils are rep­ spots were located at two-foot inter­ resentative of sites which are presently vals along the transect, from 2 to 200 occupied by creosotebush or subject feet. Spots were marked with nails for to invasion by creosotebush. future observation. One hundred carpels from a collec­ In June or July each year, four tion averaging 43.3 percent fill and carpels were planted at each spot at 91.1 percent viability were spot-seeded depths of 14 to 3/8 inch. Only the in a gridiron pattern in each of two heavier carpels separated by water five-gallon buckets of each soil. Damp- flotation were planted. ing-off of emerged seedlings reached At the time of planting, perennial very high levels, all seedlings being grass cover values were obtained by affected in some buckets. This neces­ observation of the surface condition sitated replanting, which was done for presence of perennial grass under after drenching the soil in each bucket each foot mark along the transects with a 1:10,000 water solution of from 1 to 200 feet. After the summer Panogen (methyl mercury dicyandi- growing season each year, emergence mide, 2.2 percent). Thus, five seed­ and survival of all creosotebush seed­ lings were obtained in each bucket, lings on the transects were recorded. one centrally and four peripherally located. After eight months' growth, tin loin peripheral seedlings in each Natural Field Establishment bucket were clipped at soil level, dried, and weighed. (Three of the This experiment was designed to four peripheral seedlings in one determine how the number and prox­ bucket of stable sandy loam soil died imity of seed-bearing plants and the before reaching the age of eight soil stability affect the rate of natural months. The cause was not apparent.) establishment. In 1956, plots repre­ At the end of 12 months, the central senting three variations of these con­ seedling in each bucket was clipped ditions were established, and in 1962 at soil level, dried, and weighed. plots representing a fourth variation were added. Circular plots 400 feet in diameter were used where creosote­ Influence of Range Condition on bush plants were scattered. In mod­ Creosotebush Establishment erate to heavy stands of creosotebush, 200-foot diameter plots were used. All From 1956 to 1958, this experiment creosotebush plants on the plots were was conducted at one location on the located from a central hub by means ranch, and from 1959 to 1961, it was of azimuth and radial tape readings. Repeal observations were made in 1957 to I9(i.r) ai the Jornada location. 1962 and 1965. Keiold ol It nil piodm lion was made lol tile planls as the) i .mie into ll nil ing. Growth and Maturation

This experiment had as its objec Death toss and Borer Damage live the determination ol the rate ol growth and maturation ol creosote I his study was made in mature bush seedlings. Three groups of 20 stands. Temporary plots were laid pi.mis each ol n.mu.illv established oui on the experiment station ranch seedlings were delineated at each ol .mil at seveial outlying locations in two locations in the vicinity of the Siena, Luna, Dona Ana, and Otero experiment station ranch. One loca­ counties. Dead creosotebush plants tion was on the adjacent Jornada Ex­ weie counted on these plots. Insect perimental Range and the other was larvae wen observed to be responsi­ on the A. L. Winder ranch, also adja­ ble lor damage lo some ol the dead cent to the experiment station ranch. plants, so additional dead plants and Height measurements were recorded unthrifty and thrift) live planls were lor each plant each fall from 1956 to dug, and the (towns were split and 1965 at the Winder location and from examined loi larva] < flannelling.

Results and Discussion

Fruit Production flowering and fruiting were greatly reduced in heavily damaged (reosote­ Results of the fruit production bush stands. study are presented in table 1. The I li< se effects of cold were similar unusually low production in 1962 is io those repotted by Fosberg (9) for attributed to the extremely low tem­ (reosotebush near St. George, Utah perature which oecuned on January in 1937. Here damage to stems was 11 of that year. The Jornada Experi­ heavy, and all plants exhibited greatly mental Range headquarters recorded reduced fruit production, but no a near-record low of —19°F for this plants were killed. date (18). Other extreme lows were In 195C), fruit production on that recorded at numerous stations in pai i o| the ranch where the fruit pro­ south-central New Mexico on this duction study groups were located date. Following this, the leaves of was exceedingly low. Although rain­ creosotebush plants over wide areas fall was low from January through on and near the experimental ranch September, it did not appeal to have turned brown and dropped, and been low enough to account for the many plants exhibited splil stems and extreme deficiency of production. sloughed bark, few il am plants were Low temperature may have been par- killed, however, and most plants pro- tiallv responsible. A minimum of 5°F dincd vigorous basal sprouts. Flower was recorded for February 3, 1956 at buds are thought to have been se­ the Jornada Experimental Range verely damaged by the cold, because headquarters. No stem, twig, or leaf damage was observed following this in table I. Plants in the Powerline low temperature, however. and Pastille 2 grassland groups pro­ Variance analysis of the data of duced much less fruit than those in table I, excluding that for 1962, re­ the creosotebush vegetation type. The vealed significant differences between greater size of the plants in the Selden plants in the creosotebush type and group compensated largely for the those in the grassland type and be­ lower production per unit of branch tween those on certain of the loca- weight, with the result that plants in tions in the grassland type. Fruit pro­ this grassland group produced almost duction in the creosotebush type was as much fruit as the smaller plants in between two and eight times that in the creosotebush type. open grassland. The difference in The difference in production be­ fruit set in the two vegetation types tween the Selden group and the ether was easily seen in the field. It ap­ grassland groups was indirectly a re­ peared to be related in large part to sult of more numerous creosotebush difference in clipping and trampling plants at the Selden group location. damage. The greater number of plants had Isolated plants were commonly the effect of reducing clipping and (lipped by small rodents and rabbits trampling each plant, and this al­ and trampled and horned by cattle lowed the plants to reach a larger size and other larger animals, presumably and more mature condition. This, in antelope, in the study area. Rabbits turn, caused further reduction in dipped the stems of isolated plants at clipping and trampling. The produc­ heights ranging from 4 to 18 inches tion of the plants in the Powerline above ground level, apparently at ir­ and Pasture 2 groups was more repre­ regular intervals. Cattle commonly sentative of plants in lightly invaded trampled and horned the larger iso­ grassland. lated plants, breaking the stems back The similarity in production be­ to one- or two-foot stubs. Rabbits ap­ tween the Selden group and the creo- pealed to be responsible for the sotebush-type groups was also caused greater part of the total damage. Few by the larger size of plants in the Sel­ mature isolated plants escaped this den group, which compensated for damage: in fact, close observations the lower production per unit weight commonly revealed a series or se- of branch in this group. The larger quence of damage events on oldei si/e of plants in the Selden group as plants. Apparently, as a result of this compared with the creosotebush-type clipping and trampling, the shrubs plants probably results from the more were kept in a vigorous, vegetative, favorable site conditions in the grass­ non-fruiting condition. Typical rab­ land and the much lower interplant bit clipping and the low-fruiting con­ competition. dition are contrasted with the un­ Census figures on jackrabbits in the dipped, high-fruiting plants in fig­ study areas in recent years reflect low ure 1. populations. These have in the main The clipping and trampling of not damaged the desirable forage grassland plants also had some influ­ grasses. The cattle stocking rate has ence on their size, and this, in turn, been relatively light, averaging 3.77 affected fruit production per plant. animal unit years per section year­ Size of plants, expressed as dry weight long before the drought of the early of branches, and calculated average I950's and somewhat lower since. fruit production per plant are shown Thus, relatively low numbers of both

Fig. 1. Upper left. Normal creosotebush plant which was protected from clipping and trampling. Upper right. Plant in a highly vegetative, low fruiting condition, because of repeated clipping by rabbits. Lower left. Branch from undipped plant showing abun­ dant fruit production. Lower right. Branch from clipped plant. cattle and jackrabbits produced the tation for the preceding January-Sep­ degree of trampling and clipping tember period suggests a trend in the damage observed. direction of reduced production with The correlation of fruit production increased precipitation at the .10 in the creosotebush type and precipi­ level of probability (figure 2). This is "e> 500 I I

50-yeaI r mean (7.04I inches)

£ 400

6 300 Y = 432.02 - 19.38 X r = .77 (P 2.10 )

£ 200

z

o ,0°

~o

• ---1,. -

Fig. 2. Relationship between January-to-September precipitation and creosotebush fruit production, Agricultural Experiment Station ranch, 1955 and 1957-61.

consistent with Dalton's (7) observa­ later in the year. Chew and Chew (5) tions that the least watered plants reported spring and fall seed produc­ produced the most flowers and that tion near Portal, Arizona to be ap­ frequent watering induced flower proximately proportionate to rainfall diop. Over a long period, fruit pro­ before blooming in spring and fall, duction might be a little higher than with fall production being between 9 that observed in the present study, and 10 times spring production. Gen­ since the long-time average January- eral observations indicate that a simi­ September precipitation as recorded lar response to spring and main grow­ for the Jornada Experimental Range ing-season precipitation prevails in headquarters (19) is 1.48 inch lower the area of this study, and that the than the average for this period ob­ fall fruit production is several times served in this study. Annual fruit pro- as great as the early season produc- dud ion may also exceed the values found in the current study because of biseasonal fruiting. Creosotebush plants commonly carry nearly mature Carpel Fill fruit early in the summer preceding the main growing season. It was not Fill of carpels is presented in table determined in this study whether 2. Fill varied from 12.50 to 61.75 per­ such plants produce additional fruit cent and averaged 35.26 percent. This average is much less than the 60.6 per­ cent (3.03 mature seed per fruit of five carpels) reported by Chew and Chew (5). Their value was for the year 1958 only, however, and may simply repre­ sent fill in a single favorable year. Variance analysis revealed significant differences among years and between the two collection sources in the creo­ sotebush site. Correlation analysis failed to reveal a relationship between seed set and January-September pre­ cipitation. No difference in fill was found between the creosotebush vege­ tation type and the grassland.

Seed Viability and Longevity Results of germination tests with seed collected each year are presented in table 3. Germination percentages varied from 54.8 to 89.6 and averaged 73.9. Variance analysis revealed sig­ nificant differences among years. Cor­ relation analysis failed to reveal a re­ lationship between viability and pre­ cipitation for the January-September period preceding collection. Germination values for the 1954, •1=1 1955, and 1956 collections from col­ lection to January, 1962 are presented in table 4. A number of significant differences are indicated within each collection. The November-to-january increase in germination percentage of the 1954 crop may reflect an after- ripening process in this lot. The lower values for the 1956 crop in the latter part of the test period may reflect a decline in viability. Other differences are thought to have resulted from in­ advertent variations in water, light, and fungal infection during the tests (Gerard, 11). The average germina­ tion percentages varied little for two years and then tended to decline. Six of the last nine values in the last col­ umn of table 4 fell below the overall average more than the margin of tol- TABLE 3. Average germination percentage of creosotebush seed collected on the Agricultural Experiment Station Ranch, 1954-61 *- 2

1954 1955 1956 1957 1958 1959 1960 1961 Avg. 82.5 89.6 75.9 79.7 69.4 66.3 73.2 54.8 73.9 cd d bed bed be ab be •.ilMiilirnntly .linYroit at the .05 pr f 13 tests run from November IS ti

TABLE 4. Average germination of creosotebush seed from 1954, 1955, and 1956 crops, from dates of collection to January, 1962

1954 Crop1 1955 Crop1 1956 Crop1 Seed Germinatioi percent percent percent yr. mo. percent (1954) (1955) (1956) 63.7 a 85.1 c 77.9 efg 75.5 72.4 c 87.9 de 87.5 i 0 1 82.5 79.9 d 87.1 cd 68.5 c 78.4 (1955) (1956) (1957) 84.9 fgh 91.0 fgh 60.6 b 0 2 78.8 77.8 d 97.3 i 80.1 gh 85.0 88.5 ijk 88.1 de 63.5 b 0 3 80.0 87.0 ghijk 82.6 b 75.3 def 81.6 72.2 c 89.6 defg 83.8 hi 0 4 81.8 87.4 hijk 89.7 efg 73.7 de 83.5 85.0 fgh 95.0 i 83.7 hi 0 5 87.8 93.9 I 88.6 def 74.4 def 85.5 96.6 I 92.0 gh 77.0 defg 0 6 88.4 84.0 fg 91.4 gh 80.7 gh 85.3 94.3 I 92.7 h 50.2 a 0 7 79.0 93.6 I 87.3 cde 74.4 def 0 8 85.0 85.5 fghi 92.8 h 74.8 def 0 9 84.3 83.9 fg 91.6 gh 76.6 defg 0 10 83.9 90.0 k 81.5 b 87.9 i 0 11 86.4 84.4 fgh 96.2 i 77.0 defg 0 12 85.8 (1956) (1957) (1958) 85.0 fgh 78.7 fg 13 81.8 85.0 fgh 87.6 de 1 6 86.3 94.0 I 72.9 d 1 9 83.4 80.5 de 97.3 i 2 0 88.9 (1957) (1958) (1959) 89.4 jk 63.4 b 2 5 76.4 86.8 ghijk 3 1 86.8 (1958) (1959) (1960) 67.4 b 77.8 efg 3 7 72.6 85.2 c 59.9 b 4 1 72.6 (1962) 61.5 b 4 2 61.5

82.9

61.0 crance allowed in official seed testing low in 1956 because of very low rain­ (1). Indications are that viability fall. At the swale site, emergence was holds up fairly well for four or five highest in 1956 because of reduced in­ years, and the 1954 data suggests that terference from excessive moisture and it may, in some instances, continue to flooding. At the ridge top and upland remain up for as long as seven years. sites in 1958, growing-season rainfall Viability under natural conditions, was about three times the long-time involving exposure on or near the soil average; however this was not reflected surface, cannot be projected from the in increased emergence. No seedlings present study. survived at any site in 1963. All plant­ II germination of seed in the carpels ing grids were located in plant cover brought about a reduction in germ­ normal to the sites. The extent to ination percentage, it was not nearly which the biotic factors of microcli­ as great as that reported by Knipe mate and competition influenced the and Herbel (13)—90.5 percent with outcome is unknown. It cannot be the seed removed from the carpels concluded from the results obtained and 34 percent with seed remaining over the short period of study that in the carpels. there are no differences in seedling emergence and establishment among the sites. Influence of Site on Creosotebush Seedling Emergence Influence of Soil on Emergence, Analysis of variance revealed no Damping-off, and Growth of significant differences in average seed­ Creosotebush Seedlings ling emergence among sites or years (table 5). Some differences within sites Emergence, damping-off and and years were revealed. At the ridge- growth data for the creosotebush seed­ top and upland sites, emergence was lings grown in the several soils are

TABLE 5. Average emergence of creosotebush seedlings at three sites on the Agricultural Experiment Station Ranch, 1956-591

1957 1958 1959 Average

ridgetop

Sandy loam; upland Black gra 4.84ab 13.71b 8.05b 6.00ab 8.15

11.11b 3.05a .76a 2.14a

5.93 9.08 4.70 6.38 shown in table 6. Emergence values ex­ in this category in the one deep hibited significant differences among sandy loam soil. Emergence and dam soils. age in soils from the creosotebush Emerged seedlings showed varying types were not different from emer­ degrees of infection by damping-off gence and damage in soils from the organisms, and most were either dead other vegetation types. or heavily damaged. The low emer­ Growth, as measured by weight, ex­ gence probably reflected highly dam­ hibited erratic values, and differences aging infection in early germination among soils were not significant for stages. Death and damage values, seedlings clipped at either 8 or 12 which also differed significantly among months of age (figure 3). All seedlings soils, were very likely influenced by were dark green and thrifty. These emergence values; that is, the greater results indicate that the soils tested the number of emerged seedlings, the are about equally capable of support­ greater the number to exhibit death ing growth of creosotebush seedlings, and damage. The number of seedlings but they may differ some with re­ sustaining only light damage, from spect to emergence and establishment. which they might have recovered un­ In one bucket of the sandy loam from der favorable conditions, exhibited the black grama grassland, three of considerable variation. No seedlings the four peripheral seedlings were fell in this category in two of the soils, lost before they were eight months and an average of six per bucket fell old.

TABLE 6. Average emergence, damping-off damage, and growth of creosotebush seed­ lings in greenhouse tests with soils from various sites and vegetation types on Agricultural Experiment Station Ranch Dead and Weight of Oven-Dry Site and Moderate Light Seedlings*' Soil Vegetation Type gence' 8 mo. 12 mo. percent of viable seed planted gm gm

Clay loam Swale 31.6a 31.6mn 0.0 1.0 5.8 tobosa grass

Eroded Simona Plain 8.9a 8.9m 0.0 1.5 8.1 sandy loam Snakeweed-fluff-

Shallow stoney Hillside 32.9a 30.4mn 2.5 1.1 3.9 Creosotebush

Shallow loamy Low ridge 63.3b 57.0no 6.3 1.8 5.8 sand Creosotebush

Deep sandy loam Plain 82.3b 67. lo 15.2 1.4 4.1 Snakeweed-drop- seed

Stable Simona Plain 72.2b 65.8o 6.3 1.0 4.1 sandy loam Black grama grass Fig. 3. Eight-month-old creosotebush seedlings grown from various sites on and near the Agricultural Experiment Station ranch. 1. Clay loam. 2. Eroded Simona sandy loam. 3. Shallow stoney silt loam. 4. Shallow loamy sand. 5. Deep sandy loam. 6. Stable Simona sandy loam.

Influence of Range Condition on factors, first, microclimate of range Creosotebush Establishment in this condition class appears to be more favorable. Campbell (3), work­ Results of this experiment are pre­ ing under similar climatic conditions sented in table 7. Variance analysis on the Jornada Experimental Range of the six-year emergence values at with Livingston atmometers, observed the Pasture 15 location revealed signi­ that the average daily evaporation ficant differences between the two rate from a tobosa grass vegetation range condition classes and among cover was 19 percent below that from years. Analysis of the 1959-61 results barren soil. Second, a greater degree at the three locations revealed a dif­ of soil stability prevailed within the ference between the condition classes good condition cover. As a result, the at the 10 percent level of significance carpels remained in the soil about as and highly significant differences they had been planted. On the very among years. The slight drop in signif­ poor condition range, wind scoured icance level between condition classes the surface and the carpels were resulting from the three-year location blown out at some spots along the analysis appears to be due to the fact transects and deeply buried at others. that two of the three years were quite Actual establishment results were abnormal with respect to rainfall. In very erratic, but indications are that 1960 the growing-season rainfall was there is little if any difference between 46 percent of average and in 1961 it the two condition classes. It is possible was 213 percent of average, so that that competition of the better peren­ the differences due to condition classes nial grass cover acted to eliminate in more nearly average years may many of the emerged seedlings in the have been largely eliminated. good condition range. The artificial The greater emergence in the good planting procedure used in the ex­ condition range is attributed to two periment may have had some influ- o a || E

cn u « <

0 -O I a 0 (2 <

0 8 i -aCT> o o s o -o o o *o o s D- s

P

o o 2

0 g 0 "5 1/5 -D s E " g CL " S s 5 a o < | 5 § 01 2 p < < • 2 TJ 0 0 5) 9) c r fe_ s t_ s 11 1

Eu 0) a.-5^ . 5 -9) r<< •| * 18 < ence on the outcome, but the direc­ mated 3,570 viable seeds planted un­ tion and magnitude of this influence, der each condition produced estab­ if any, is unknown. The results are lished seedlings. In Pasture 4 and 5, consistent with general observations planting of the same number of seeds on and in the vicinity of the experi­ under each class resulted in the es­ mental ranch; no evidence has been tablishment of one plant. The overall seen of a relationship between range rate of establishment at the three condition and establishment of creo­ locations was about .37 seedling plant sotebush. The plants seem about as per thousand viable seed planted. likely to appear on good condition When the surviving seedlings were range as on poor condition range, counted in the fall of 1961, the causes provided that a comparable source of of stress and damage to surviving seed is available. Neither the time seedlings and of death of seedlings covered by the present study nor the were observed (table 8). Drought total number of seed involved was caused most of the stress and loss, al­ great. Results covering a considerably though damping-off was responsible longer period and involving many for a small proportion of the loss. It is times as many seed, all under fully possible that loss was actually brought natural conditions, might be different. about by a combination of damping- In the good condition range, two off damage and late-season drought. of the six creosotebush seedlings be­ The soil influence study revealed that came established in openings in the creosotebush seedlings may become perennial grass cover. One of these severely infected by damping-off or­ was six or eight feet across, and the ganisms. Exceptionally good rainfall other was considerably larger—50 or of the early and middle part of the 60 feet across and centered on an ac­ 1961 growing season may have been tive kangaroo rat den. Such openings quite favorable for infection in the as these, although representing a small field, which may have produced un­ proportion of the total area, are not detected losses (losses before emer­ at all uncommon in good condition gence). The slightly greater percent­ black grama range (figure 4). age of drought-stressed seedlings in The five seedlings which became the good condition range may reflect established in the very poor condition the greater competitive strength of the range all emerged in 1961 when the perennial grasses. excellent growing-season rainfall may The influence of the year on seed­ have tended to offset the microcli­ ling emergence is chiefly a matter of matic adversity. Five of the six seed­ grcwing-season rainfall, figure 5. This lings which became established in the relationship was curvilinear; average good condition range emerged in and below-average rainfall brought years other than 1961. about little emergence. Average emer­ The very low rate of establishment gence (about 10 percent of the viable seems noteworthy in view of the fact seed planted) was reached only when that creosotebush, when mature, is growing-season rainfall reached about so well adapted to xeric conditions. eight inches, nearly 60 percent above From the 7,740 viable seed (number the long-time growing-season rainfall. estimated from fill and germination According to the 50-year record (1916- data) planted in each of the two con­ 65) at the Jornada Experimental dition-class ranges in Pasture 15, only Range headquarters, eight inches or five seedlings became established in more of growing-season rainfall was each. In Pasture 3, none of the esti­ received in only seven years. Thus, Very poor condition range. Eroding soil and harsh micro­ climate are adverse to seed­ ling establishment.

Barren opening in good condition black grama range adjacent to a kangaroo rat den. Such openings are not uncommon in generally good condition range. Although microclimate is harsh, rodent- induced soil instability may be favorable for planting of seed.

Good condition black grama range. Here relative stability of soil may be un­ favorable for planting of seed, but microclimate in­ duced by the grass cover ap­ pears to favor germination and emergence.

Fig. 4. Effect of range condition and microclimate on creosotebush seedling estab­ lishment. TABLE 8. Stress, damage, and death of creosotebush seedlings under poor and good range condition, average all locations planted in 1961, Agricultural Experiment Station Ranch

Range Condition of Condition Seedlings tive Seedlings Dead Seedlings Total -percen

Drought-stressed 71.0 20.1 91.1 Damped-off .4 3.1 3.5 Trampled .2 .3 Shaded out .0 .2 .2 Unknown .0 .2 .2 Thrifty 4.7 4.7 100.0

Good Drought-stressed 57.6 40.1 97.7 Damped-off .0 1.2 1.2 Thrifty 1.1 100.0

9'= -.101 + .14379 r = .95** •/

*<

-A/

Growing Season Rainfall (July - Sept.) - Inches

Fig. 5. Relationship between growing-season rainfall and emergence of planted creosotebush seedlings, Agricultural Experiment Station ranch. emergence ol any considerable pro­ increase, figure 6. From 1956 to 1965, portion of seed planted would occur the average annual rate of increase in relatively few years. The Jornada loi the five plots was .06 plants pei headquarters record also reveals er­ acre. Annual observations ol liuit ratic 01 iiugiilai spacing of these production on the single mature rains. 1 his rainfall characteristic and plants on the plots were not made. the actual emergence values observed On the basis ol the calculated fruit in the present stiuh suggest an irreg­ production from the Powerline and ular time pattern ol seedling estab­ Pastiite 2 grassland groups, which lishment, rather than the gradual were neai these plots, and the average increase pattern postulated by Chew carpel fill and viability value, it is and Chew (5). Gardner's findings estimated that these mature plants (personal communication) of no seed­ produced 30 to 70 viable seeds per lings on any of the observation plots plant per year. From this low fruit in his suivev of 172 miles (one plot production and the low rate of es­ pei mile surveyed) in the (reosote­ tablishment observed in the experi­ bush type in 1947, 1948, and 1949 ment with range condition, one would and numerous seedlings along about expect a low rate of establishment hall the distance surveyed in 1951 around single mature plants. are similar to the results of this study. The plots with single seed-bearing plants and unstable soil appear to represent a critic al point in creosote­ bush invasion of range, figure 6. Al­ Natural Establishment of though these plots were actuall) Creosotebush as Influenced by established in 1962, the size of the Proximity and Number of Seed young plants and the size-age rela­ Source Plants and Soil Stability tionship brought out in the growth and fruit production study indicate Results of this study are presented that they were about five or six years in table 9. These results are erratic, old at this time. They probably were but they reveal a general pattern of five years old, starting in 1957, because change. 1956 was a very dry year on the ex­ The plots without creosotebush at perimental ranch. If they started in the outset remained free of the species 1957, the average annual rate of in­ through the period of observation. crease on these plots was about one This is probably chiefly a result of the plant per acre over the eight-year low rate of transport of carpels to period. The soil instability of these any particular location far removed plots may have contributed much to from seed-bearing plants and the low the establishment of seedlings by rate of establishment of plants from planting the seed. The similar sizes of carpels actually appearing at a par­ the young plants on these plots gives ticular location. Chew and Chew (5) further support to the view of irreg­ pointed out that seed dispersion is ular occasional establishment of seed­ relatively limited in creosotebush. The lings rather than regular or steady crescent-shaped carpels tend to collect rate of establishment. in depressions in the soil surface not The loss of young plants on the far from the seed-producing plants. plots adjacent to the creosotebush The plots with a single mature type is noteworthy, although no cause plant in open grassland and stable can be suggested to account for it. At soil varied from no change to slight the time of the 1962 observations, the II 8 I «s |o *J. -o 8. The single plant here was repeatedly clipped by rabbits and trampled by livestock. It is in a thrifty vegetative con­ dition and produces little fruit. The soil, although ex­ hibiting much open surface, is relatively stable. Two young plants became established near this plant over a nine- year period.

Here the single mature plant was subject to less clip­ ping and trampling. Soil is moderately stable. Forty-six young plants became estab­ lished in the vicinity of this plant over a period of 8 to 10 years.

Former grassland which was invaded by large num­ bers of young creosotebush plants. The young plants here are 11 or 12 years old and number up to 20,000 or 25,000. Soil surface is open and unstable. Seed source for this stand is an older creosotebush type immedi­ ately adjacent.

Fig. 6. Natural invasion of creosotebush into rangeland representing varying conditions with respect to seed availability and soil stability. plants had disappeared, leaving no croachment of creosotebush into grass­ indication as to cause. It is possible land range. No contrary observations that merging of enlarging plants may have been made. account for some reduction in appar­ ent plant numbers, but this process probably does not account for more Growth and Fruit Production of than a small fraction of the missing Naturally Established plants. The loss however, had no real Creosotebush Seedlings effect on the character of the plant cover, the plots remained lightly to Height growth and fruit produc­ heavily invaded by vigorous, growing tion data from the naturally estab­ young creosotebush plants. While lished seedlings are presented in table there are no evaluations of the amount 10. Analysis of the height growth and of seed coming to the brush margin age data revealed highly significant plots or degree of soil instability, gen­ correlations of height with age for eral observations indicated that both seedlings from both locations for the seed supply and soil movement pre­ period of observation (figure 7). It is vailed at high levels. Both probably to be expected that as the plants ap­ contributed greatly to the production proach full size the relationship will of the young shrub stand observed in become curvilinear. 1956 and the reproduction observed Plants at the two locations exhibited in the course of the study (figure 6). minor differences in height growth. The type of soil instability which The essentially static heights of 9- and appeared to favor seedling establish­ 10-year-old plants at the Jornada lo­ ment was deposition on the surface. cation probably was due in part to The soil may have originated only a increased competition among the cre­ few feet from the area of deposit or at osotebush plants themselves, because a distance of several hundred feet or density of the young plants at this lo­ more. Establishment took place at a cation was much above average. The very low rate on wind-scoured eroding decrease in height in the eighth year surfaces. at the Winder location is a result of Practically all seedlings near isolat­ freeze damage in January, 1962. Most ed pioneer plants in grassland were stems were killed back to or near to the leeward and within a few hun­ ground level, and the height growth dred feet of the pioneer plants. This measured for the eighth year was relatively short distance conforms with made for the most part during the the short dispersal distance reported 1962 growing season. This single by Chew and Chew (5). Extension of year's growth, although from an es­ young plants from older stands was tablished root system, was not quite also downwind, and again for only sufficient to restore the full pre-freeze a relatively short distance from the height. The Jornada location plants, older stands. Thus, control efforts can situated on a gentle slope which pro­ be confined to relatively restricted motes drainage of cold air, sustained areas. some freeze damage to leaves and Although this study did not involve twigs, but only light damage to stems. systematic sampling of extensive areas, If the Winder location plants had not the results obtained are believed to been damaged, their height growth apply to such areas. They are consis­ would probably have been very simi­ tent with Gardner's (10) and Yang's lar to that of the Jornada plants. The (21) observation of continuing en­ 10- and 11-year-old plants of these two Jornada Experimental Range <> ? < < Y .14 + 1.76 X .97" .'.' /S

_*/ Winers Ranch 9 = 1.43£ + 1.27 .?< r = .94* <

Fig. 7. Relationship between height and age of naturally established creosotebush plants at two locations.

groups attained 50 to 56 percent of the average height reported by Gard- nei (10) for mixed-age plants of ma­ un < stands on upland topography. In addition to damage from freez­ ing, a number of plants at both loca­ tions were subjected to varying de­ grees of rodent and rabbit clipping. Values for all such plants were in­ cluded in computation of height av­ erages, because such damage appears to be normal to all naturally occur­ ring young creosotebush plants. Fruit production began at a very low level in the fourth and fifth years at the fornada and Winder locations, respectively. By the tenth and elev­ enth years, the plants were fruiting abundantly for their size. The de­ crease in production at the Winder I = location in the eighth and ninth years was caused by the freezeback of stems of 1962. Almost all stems in 1962 were from new basal sprouts, which were very thrifty. This thrifty, highly vege- tative condition carried over into stands consisted entirely of young 1963. plants not in competition with an es­ Chew and Chew (5) found no fruit tablished community of older plants; production on plants estimated to be the observations of Chew and Chew younger than 13 years near Portal, were within established stands of Arizona. Site differences may account older plants. Results obtained in the in part for the difference in time re­ present study appear to be applicable quired to attain fruiting size or con­ to other comparable sites in the area dition. Soil at the Winder location is of the study. a deep sandy loam with almost opti­ mum water-supply characteristics un­ der arid climatic conditions. Topog­ Death, Borer Damage, and raphy is very gently sloping, with an Crown Decay in Creosotebush east exposure. Soil at the Jornada lo­ cation is a deep gravelly sandy loam. Results of this study are presented Although the site is gently sloping, in table 11. Dead creosotebush plants inducing some runoff, the seedlings are rarely conspicuous in the creo­ themselves were in a very shallow sotebush type, but a few were found slowly-draining trench. This micro- at each of the locations (figure 8). topography may have operated to off­ Most of these dead plants were medi­ set normal runoff. It is possible, too, um-size to full-grown, and most had that differences in age at first fruiting been dead for several years, judging as found in the present study and as from the degree of weathering and found by Chew and Chew are ac­ termite attack. Smaller plants may counted for by differences in observed undergo some death loss, but they stands. In the present study, the may decay and disappear more rap-

Death and borer channelling of creosotebush plants from the creosotebush vegetation type and from open grassland Degree of Borer Channelling in-

percent of plants examined

Pastures 2 and 14 01 Past. 1 and Powerline 0 Selden 0

irifty plants were found in the grassland type. Fig. 8. Dead creosotebush plants (indicated by markers) in the creosotebush type. The area occupied by these dead plants is approximately two-fifths of an acre. Such plants, although generally inconspicuous, are consistently present in the creosotebush type.

idly than the larger, woodier plants, losses; however, it is possible that the leaving no evidence of death in then population at the Nunn location rep- size class. Dead plants generally repre­ resented a stage prior to one in which sent only a small proportion of the stress was causing losses. Plant counts total population on any given area. at the Nunn location averaged 2,010 Counts of creosotebush plants at the per acre, many of which were small four locations listed first in the table plants. Tarbush was also present in averaged 1,122 pei acre. The average greater density at the Nunn location number of dead plants indicated a than the average for the four loca­ low percentage of loss. The annual tions, 1,048 and 361 plants per acre, death rate is much lower. Such losses respectively. This also indicates rela- do not aliet the vegetation type of the tive recency of creosotebush invasion ranges in this study. according to Gardner (10) who con- Gardner (personal communication) (hided that early invasion stages of states that he observed no dead creo­ creosotebush were evidenced by the sotebush plants in the extensive sam­ presence of tarbush, the earlier in­ pling of the type over the four-yeai vader of the sites. period of his study. Dead plants ob­ The present study was made only a served in the present study consider­ few years after the very severe drought ably exceed the one plant per hectare period which prevailed in south-cen- (.4 plant per acre) reported by Chew tial New Mexico in the 1950's. It is and Chew (5). possible that this period of drought I he scarcity of dead plants at the contributed to the death loss ob­ Niinii location may be related in part served, and may account in part for to age of stand. Chew and Chew (5) the differences between Gardner's and concluded that high mortality was Chew and Chew's observations and associated with young populations those of the present study. which had reached a stage when den­ No dead plants were observed sity-dependent factors were causing among the isolated plants, even though many of these were large and ers. Fhe degree of channelling was evidently old. Also, all of the isolated generally much less in the thrifty plants exhibited a relatively high de­ plants. This was especially true of iso­ gree of thrift. lated plants; channels, when present, As shown in the table, borer chan­ were usually single and unbranched nels were generally present in plants (figure 9). Most of the channelling ob­ occurring in the creosotebush type, served was old rather than recent, and but crowns of isolated plants exhib­ very little current channelling was ited little channelling. It appears too, noted. that in the creosotebush type, the bor­ Very few larvae were found in the ers must have attacked fairly young plant crowns. Of those which were in plants, because the thrifty plants, condition to be submitted for identi­ most of which were a little less than fication, three were buprestids (Ac- lull grown, exhibited as high a pro­ maeodera sp.), one was a cerambycid portion of channels as the unthrifty (Cerambycinae—not determined to and dead plants. Unthrifty and dead genus) and one was the larva of a leaf plants commonly exhibited a number cutting bee (Megachilidae—Ashmea- of channels, suggesting that they are diella sp.). The beetle larvae are mem­ subjected to repeated attack by bor­ bers of families well known as destruc-

Fig. 9. Crowns of live creosotebush plants taken from open grassland (upper row) and from the creosotebush type (lower row). Crowns of open grassland plants generally exhibit little channelling by beetle larvae and are relatively free from decay. One borer channel, surrounded by sound wood, is pointed by arrow in second crown from left. Crowns of plants from the creosotebush type generally exhibit more channelling and death, and decay of wood is more extensive.

28 tive wood borers of both cultivated which affected the sapwood as well and native plants. The bee larva is as heartwood, would accentuate wa­ considered to have been present be­ in stiess in the tops ol the plants dur­ cause ol oviposition in an exposed ing drought. Ibis might materially pan ol a channel made by a wood- increase drought loss in affected pop­ boring species. ulations. Crown specimens from a number Another cause of death of creo­ ol unthrift) creosotebush plants were sotebush is (lipping of roots by po< ket taken lor determination of fungal in­ gophers; however, very few plants fection. About hall weie infected by have been observed to have been common saprophytes associated with killed in this manner, even under con- soil and decaying vegetative material. dilions in which damage to ground The unthrifty condition did not ap­ covei by gopher mounding has been pear to be clue to attack by virulent \ei\ heavy. Generally the few plants fungi. Crowns of the borer chan­ clipped were so heavily damaged that nelled plants of the creosotebush type the) died (|ui(kl\ and were subject to were generally conspicuously decayed; removal In wind because ol total loss while comparatively little decay was of anchorage. Gophers are not gener­ associated with the channelled crowns ally present in the creosotebush type, of isolated plants (figure 9). however, and damage from this source It seems probable that the exten­ is infrequent and highly localized, sive diannelling and decay of crowns, when it does occur.

Summary and Conclusions

Several characteristics in the aute­ plants in the creosotebush type, plants cological life history of creosotebush occurring as scattered individuals in were studied on the New Mexico State grassland produced from 10 to 111 University Agricultural Experiment fruits per 100 grams of branch and Station ranch in Dona Ana County averaged 49 fruits. This lower fruit and at other locations in south-central production ol the scattered plants was New Mexico, from 1954 to 1965. associated with a highly vegetative rhese characteristics were: 1. fruit plant condition induced mainly by production and seed production and rabbit clipping and some livestock viability, 2. site and soil influences on trampling. The rabbit and livestock establishment of creosotebush, 3. In numbers responsible for damage to fluence of range condition on estab­ creosotebush have generally not been lishment, 4. natural invasion and es­ damaging to the more desirable for­ tablishment, 5. growth and matura­ age grasses. Fruit production per tion, and (i. damage and death. plant in the creosotebush type aver­ In normal years, fruit production aged about 1700 annually, and on of plants in the creosotebush vegeta­ scattered plants it averaged 665 an- tion type varied from 72 to 518 fruits nually. A trend in the direction of re­ per 100 grams of oven-dry branch. duced fruit production with increased I IK average was 263 fruits. In one January to September precipitation year, diaracterized by extremely low was indicated. temperature in January, fruit produc­ Fill of carpels, or set of seed, varied tion dropped to an average of 1 per from 12.50 to 61.75 percent of carpels, 100 grams of branch. In contrast to and averaged 35.26 percent, less than filled the five- be related to growing-season (July to dlate ill September) precipitation. Average years but was not associated with growing-season precipitation would rainfall of the January-September be expected to bring about emergence pei iod. of about five percent of viable seed Germination percentages on a filled planted. carpel basis averaged 73.9 percent and Natural establishment of creosote­ varied from 54.8 to 89.6 percent. Sig­ bush was found to be associated with nificant differences among years were proxir id .lie ol revealed, but these were not found to seed-bearing plants and degree of soil be assodated with January-September instability. precipitation. Germination at the end Range areas at distances from seed- of seven years in storage showed no bearing plants may remain uninvaded marked drop below values obtained lor extended periods of time. With in earlier tests. the establishment of a single plant, No differences in emergence of the immediate area may still remain planted creosotebush seed was re­ free from further invasion, pending vealed among a sandy loam ridgetop the attainment of seed-producing size site supporting a developed stand of of the pioneer plant. This may be de­ creosotebush, a sandy loam upland layed by rabbit clipping and tram­ carrying a black grama grass cover, pling by livestock. Following seed and a deep day loam swale carrying production by the pioneer plant, a tobosa grass cover. Further, no seed­ seedlings appear at irregular intervals, lings survived at any of these sites apparently dependent to a large de­ four years after the last planting. Pot gree on infrequent favorable rainfall tests revealed very high levels of years and in some measure on degree damping-off of creosotebush seedlings of soil instability. Indications are that in all soils tested, including those the isolated pioneer plants should be from creosotebush types. No signifi­ destroyed before they attain seed-pro­ cant differences in growth, expressed ducing size. in weight of seedlings, at 8 and 12 Along the leeward margins of estab­ months, was found on any of the soils. lished creosotebush types, particularly Emergence of creosotebush seed­ if soil is unstable, young plants may lings from planted seed was signifi- become established and develop at cantly higher under good condition siu 11 tapid rates as to bring about the than under very poor condition black conversion of brush free-range to the grama range, 11.43 and 7.64 percent creosotebush type in 12 to 15 years. of viable seed planted, respectively. Control along such expanding mar­ The greater emergence under good gins is strongly indicated where fea­ condition was presumably due to ben­ sible. eficial microclimate and greater soil Dead creosotebush plants may be stability of the good condition range. found in most older stands, but the Range condition had no differentiat­ rate of loss is apparently so low that ing influence on final survival. The no significant effect is produced in overall rate of establishment of young the stand; creosotebush remains as plants was extremely low—0.37 plant the dominant cover species. Cause of pet thousand viable seed planted— death may be a combination of borer and essentially the same for very poor attack of crowns, secondary damage and good condition range. Emergence, by wood-decaying fungi, and drought although generally low, was found to Literature Cited

I \ss.Hi.umn of Olli.i.il Seed \n.il\sis. I'M.v Rules l,,i Seed resting. Proceedings ol the \s sociation of Official Seed Analysts. 54(2) .

2. Buffington, Lee C. and Carlton I lei bel. 1905. Vegetation Changes on a Semi-desert Grassland Range I < ol. Monog. 35 (2): 139-164.

3. Campbell. R. S. 1929. Vegetative Succession in the Prosopis Sand Dunes of Southern New Mexico. Ecol. 10(h

4. Canfield, R. II. 1 !>.">•>. I lie Effects ol Intensity and Frequency of Clipping on Density and Yield of Black Grama and Tobosa Grass. U.S. Dept. of Ague. I ech. Bull. 681.

5. Chew, Robert M. and Mice Eastlake (hew. 1965. The Primary Productivity of a Desert Shrub 'Larrea tridentata) Community. Ecol. Monog. 35(4) :355-375.

6. Cooperrider, Charles K. and Bernard A. Hendricks. 1937. Soil Erosion and Stream Flow on Range and lores! Lands of the Upper Rio Grande Watershed in Relation to Land Resources and Human Welfare. IS. Dept. of Agric. Tech. Bull. 507.

7. Dalton, Patrick D., Jr. Ecology of the Creosotebush tl ana iridenlata (DC.) Cov.). Ph.D. Thesis, University of Arizona. 1961.

8. Dick-Peddie, William A. 1905. Changing Vegetation Patterns in Southern New Mexico. New Mexico Geological Society, Guidebook Sixteenth Field Conference. Southwestern New Mex-

9. Fosberg, F. R. 1937. The Lowei Sonoran in Utah. Sc. 87:39-40.

10. Gardner, J. L. 1951. Vegetation of the Creosotebush Area of the Rio Grande Valle) in New Mexico. Ecol. Monog. 21 (4) :379-403.

11. Gerard, Jesse Blake. 1965. Factors and Treatments Affecting Fruit Fill, Seed Germination and Seedling Emergence of Fourwing Saltbush (Atriplex canescens (Pursh.) Nutt.). Master's I hesis, New Mexico State University.

12. Jardine, James T. and L. C. Hunt. 1917. Increased Cattle Production on Southwestern Ranges. U.S. Dept. of Agric. Bull. 588.

13. Knipe, Duane and Carlelon H. Herbcl, I960. Germination and Growth of Some Semidesert Grassland Species Treated with Aqueous Extract from Creosotebush. Ecology 47 (5) :765-781

14. Leithead, Horace L. 1959. Runoff in Relation to Range Condition in the Big Bend—Davis Mountains Section of Texas. Jour. Range Mgt. 12 (2) :83-87.

15. Paulsen, H. A., p.. and 1 N. Ares. L962. Grazing Values and Management of Black Grama and Tobosa Grasslands and Associated Shrub Ranges of the Southwest. U.S. Dept. of Agric. Tech. Bull. 1270.

16. U.S. Department of Commerce, Weather Bureau. 1948. Climatological Data. New Mexico Annual Summary. Vol. LII No. 13.

17. U.S. Department of Commerce, Weather Bureau. 1951. Climatological Data New Mexico Annual Summary. Vol. LV No. 13. 18. U.S. Department of Commerce, Weather Bureau. 1962. Climatological Data. New Mexico Annual Summary. Vol. 66 No. 13.

19. U.S. Department of Commerce, Weather Bureau. 1966. Climatological Data New Mexico Annual Summary. Vol. 70 No. 13.

20. U.S. Department of Interior. Geological Survey—Department of the Army. Corps of En­ gineers. San Diego Mountain Quadrangle. New Mexico, Dona Ana County. Federal Center, Denver, Colo. 1947-48.

21. Yang, Tein Wei. 1961. The Recent Expansion of Creosotebush (Larrea divaricata) in the North American Desert. The Western Reserve Academy Natural History Museum. Hudson. Ohio.