Reproductive Potential and Minimum Reproductive Size of Ferocactus wislizeni (Cactaceae)

Item Type Article

Authors Bowers, Janice E.

Publisher University of Arizona (Tucson, AZ)

Journal Desert Plants

Rights Copyright © Arizona Board of Regents. The University of Arizona.

Download date 04/10/2021 01:51:49

Link to Item http://hdl.handle.net/10150/554307 Ferocactus Bowers 3

most cacti are hermaphroditic, some are monoecious, dio- Reproductive Potential and ecious, or even trioecious (Gibson and Nobel 1986, Fleming Minimum Reproductive Size et al. 1994), and although fruit set is generally low in her- maphroditic, outcrossing species in other plant families of Ferocactus wislizeni (Sutherland and Delph 1984, Sutherland 1986), high fruit set is apparently not unusual in the Cactaceae (McGregor (Cactaceae) and Alcorn 1959, McGregor et al. 1962, Johnson 1992, Suzán et al. 1994, Roller 1996, Bowers 1996); these and Janice E. Bowers other traits suggest that cacti could be instructively em- ployed in a variety of comparative studies.

U. S. Geological Survey This paper reports a four -year investigation of flower, fruit, 1675 W. Anklam Rd. and seed production of Ferocactus wislizeni ( Engelm.) Britt. Tucson, Arizona 85745 & Rose, a short -columnar, unbranched cactus that is widely distributed in the northern Sonoran Desert. Ultimately, these Abstract data will be used in studies of the reproductive biology and Flower, fruit, and seed production of the short -columnardemography of this species. Ferocactus wislizeni can live cactus Ferocactus wislizeni were studied in the northernas long as 45 years (Goldberg and Turner 1986). The plants Sonoran Desert from 1993 -1996. Most plants of Ferocactusbloom once a year from July- September. Medium -sized wislizeni start to reproduce when they attain a diameter of solitary bees are the main pollinators (McGregor and Alcorn about 19 cm (7.5 in). Number of flowers initiated and fruits 1959, Grant and Grant 1979); the flowers are self- incom- set increased linearly with plant diameter. Plants produced patible or largely so (McGregor and Alcorn 1959). Flow- 25 flowers /year on average, a very low value for a woody,ers are borne on subapical tubercules, nipplelike structures relatively long -lived plant. On average, 93% of initiatedthat contain vegetative and floral meristems (Gibson and flowers set fruit, and fruits contained about 600 -1000 seeds.Nobel 1986). As the plants mature, adjacent tubercules fuse Fruit set seemed unaffected by drought, but seed set wasinto vertical ribs. Annual height growth is largely a func- apparently diminished by unusually dry conditions. Al- tion of tubercule formation (Gibson and Nobel 1986, Nobel though the small number of flowers produced by Ferocactus1986), as is annual diametric growth, at least for part of wislizeni could be a substantial competitive disadvantage, the lifespan (Benson 1982). Mature plants have a large fruit set and fecundity are evidently high enough for popu-volume:surface area ratio. Specific questions addressed lations to maintain themselves. during this study were: 1) what is the minimum size for reproduction; 2) what is the relation between plant size Investigations into reproductive, floral, and pollination bi-and annual flower and fruit production; 3) what propor- ology often require quantitative measures of flower, fruit,tion of flowers set fruit; 4) how many seeds does a fruit and seed production. Such investigations might include al-contain and what proportion of fresh fruit weight do they location of plant resources to reproductive effort (Harperrepresent; and 5) does reproductive effort vary annually? 1977, Silvertown 1982), pollinator behavior on flowers and inflorescences (Gentry 1974a, 1974b; Thomson 1983), the Study Site and Methods trade -off between flower size and flower number (PrimackThe Tumamoc Hill study site (32°13'N, 111°05'W) is at 1985), the connection between breeding system and the 818 m (2684 ft) above sea level on a gently sloping bench relative sizes of ovary and mature fruit (Primack 1987),with a northeast aspect. Tumamoc Hill, an outlier of the the adaptive significance of excess flower production Tucson Mountains, Pima County, Arizona, has a maximum (Stephenson 1981), and the effect of drought on fruit pro-elevation of 948 m (3110 ft) and a basal elevation of 703 m duction and therefore fitness (Parker 1987). Precise repro-(2306 ft). The rocky, basaltic -andesitic slopes are domi- ductive data also contribute to demographic studies, as whennated by Cercidium microphyllum (Torr.) Rose & Johnst., age -specific seed production is used to calculate lifetimeCarnegiea gigantea ( Engelm.) Britt. & Rose, Larrea fecundity or to determine the maximum and actual rates oftridentata (DC.) Cov., Fouquieria splendens Engelm., increase of a population (Harper 1977, Silvertown 1982). Ambrosia deltoidea (A. Gray) Payne, Opuntia engelmannii Salm- Dyck., Ferocactus wislizeni and other woody plants Cacti are admirably suited to an examination of these and characteristic of the Arizona Upland division of the Sonoran other problems. Plant dimensions (stem height, stem di-Desert. Annual rainfall averages 250 mm (10 in), almost ameter, stem or cladode number) can be readily determined, half of which falls during July, August, and September. making it possible to correlate reproductive effort and plantMaximum temperatures in summer frequently exceed 40° size with unusual precision (e. g., Schmidt and BuchmannC (104° F). Minimum temperatures rarely drop below -6° 1986, Bowers 1996). Flower and fruit production of many C (21° F) in winter. Although freezing nights can be com- cacti are not so large as to hinder the monitoring of totalmon in winter, daytime temperatures always rise above 0° plant reproduction on a weekly or seasonal basis. Although C (32° F). 4 Desert Plants 1998

Flower and fruit production. In September 1993, a total ofthat Ferocactus plants invariably flower by the time they 24 Ferocactus wislizeni plants growing on a gentle north -exceed 30 cm (11.7 in) in height; the sample was therefore facing slope were randomly selected within four qualita-limited to plants less than or equal to 30 cm. All plants of tive size classes (cylinders, parabaloids, large spheres, small this size on the survey site were examined for flower buds spheres). The selected plants were tagged, and their heightsand flowers, and the height and diameter of each plant were and diameters were determined. For every plant, the num- measured. ber of ribs, flower buds, flowers, fruits, fruit scars, and aborted buds was recorded. Flower buds and flowers wereFruit and seed characteristics. Mean fruit weight, length, considered to be aborted when they shriveled before anthe- and width were determined from 17 ripe fruits collected in sis or when they abscised without having been fertilized. October 1995. The seeds from each fruit were removed, air The reproductive census was repeated for the same plantsdried, and weighed as a group to the nearest 0.1 gram, then in August 1994, September 1995, and September 1996.mean seed length and width were determined using five Sampling in the latter half of the flowering season made itrandomly selected seeds from each fruit. The average weight possible to count the entire reproductive effort of each plantof an individual seed was calculated from the collective at a single visit except in 1994, when the census was tooweight of 500 randomly selected seeds, then used to esti- early for an accurate assessment of fruit production andmate the average number of seeds per fruit. Total seed flower abortion. weight per fruit, average weight of a single seed, and esti- mated number of seeds per fruit were determined in the Numbers of flowers initiated, fruits set, and flowers aborted same way for 17 fruits collected in November 1995. A were plotted as a function of plant height or diameter, and Pearson correlation matrix was used to analyze the strength curves were fitted to the data points. Percent fruit set, cal-of the associations among the various fruit and seed traits, culated as the ratio of fruits set to flowers initiated, was and independent t -tests were used to compare between months plotted in the same way. Mean values for the four yearsthe weight of a single seed and total seed weight per fruit. were used in all regressions except for those involving fruit data, which were based on three -year means. One -wayResults ANOVA was used to compare flower and fruit productionFlower and fruit production. Of the original 24 plants, 19 among years. All sample plants, whether reproductivelywere reproductively mature in 1993 and 1994. In 1995, mature or not, were included in these analyses. A regres-one mature plant died and a young plant bloomed for the sion equation was used to examine the effect of drought on first time; in 1996, two more plants reached reproductive fruit production. Number of fruits expected on the basis ofmaturity. Over three years, 93% of initiated flowers set fruit plant diameter was determined from the equation, thenon average (range = 91% to 96 %). The sample plants compared in a paired samples t -test with the actual numbershowed no significant difference among years in number observed in 1996, when precipitation on Tumamoc Hill from of flowers initiated (F = 0.09, P > 0.80) or fruits set (F = January to July was 43% of the long -term average. 0.11, P > 0.80). Plant heights and diameters were highly correlated (R2 =0.78, P < 0.001). In linear regressions, plant Minimum Size at Reproductive Maturity. In July 1996, aheight explained a high proportion of the variance in num- survey was conducted to determine plant size at first flow-ber of flowers and fruits initiated (73% and 72 %, respec- ering. The 0.4 ha survey area, located on the north slope oftively, P < 0.001). Plant diameter accounted for an even Tumamoc Hill, included the plot used to monitor flowergreater proportion: 80% for flowers, 79% for fruits (P < and fruit production. Preliminary measurements suggested0.001) (Figure 1). Number of flowers aborted was only

--- Flowers Initiated - - Fruits Set 70 I 60-y = 1.36x-22.04 (flowers) Figure 1. Number of flowers initiated 50 R2 = 0.80, P < 0.001 and fruits set as a function of plant 40 y = 1.32x-21.45 (fruits) -- diameter. R2 = 0.79, P < 0.001 e 30 20 10

0 -10 10 15 20 25 30 35 40 45 50 Diameter (cm) Ferocactus Bowers 5

weakly correlated with plant height (R2 = 0.21, p < 0.03) Discussion and not at all with plant diameter (R2 = 0.16, P > 0.50).Collecting basic reproductive data such as annual flower, Neither height nor diameter explained any variance in per- fruit, and seed production makes it possible to compare cent fruit set. Fruit set in 1996, the drought year, was not reproductive strategies among and within plant families. significantly different from the predicted values (T = 0.248,For example, for Mammillaria heyderi Muehl. (Boke 1953) P > 0.80). and perhaps other small cacti, good reproduction one year results in poor reproduction the next. This is apparently Minimum Size at Reproductive Maturity. Because plantnot true of Ferocactus wislizeni, flower production of which diameter explained a higher proportion of the variance indid not vary significantly among years in this study. That flower and fruit production than plant height, diameter was 80% of the variance in mean flower number can be ex- used to determine minimum reproductive size. In the 104 -plained by plant diameter alone suggests that Ferocactus plant sample, the smallest plant to produce flower buds or wislizeni flowers should be a dependable source of pollen flowers was 15.0 cm (5.9 in) in diameter. More than 60%and nectar from year to year for solitary bees, the major of plants 19 -21 cm (7.5 -8.3 in) in diameter and all plants >pollinators. 27 cm (10.6 in) in diameter were reproductively mature (Figure 2). Fruit yield of Ferocactus wislizeni is also quite predictable from plant size (Figure 1). Certain cactus species, among

120 I I

100

Figure 2. Percentage of each 80 diameter class that showed reproductive activity in July 60 1996 (n = 104 plants). 40

20

1 0 I I 1=11 I I I 1 3 6 9 121518 2124 27 30 Diameter (cm)

Fruit and seed characteristics. Mean length, width, andthem Stenocereus thurberi (Engelm.) Buxb. and Opuntia weight of fruits and seeds in October 1995 are presented in engelmannii (Parker 1987, Bowers 1997) mature fewer Table 1. Fruit weight was significantly correlated with fruitfruits in drought years, but fruit production of Ferocactus width and total seed weight/fruit (Table 2). Fruits collected wislizeni in this study did not differ significantly among one month later were heavier on average (22.7 g versusyears, nor was there a significant difference between pre- 16.1 g) and devoted a higher proportion of their weight to dicted and observed fruit yield in 1996, a drought year. seeds (12.8% versus 8.1 %); the differences were signifi-Seed set is evidently more sensitive to dry conditions as cant (fruits, T = 5.0, P < 0.001; seeds, T = 4.6, P < 0.001). shown by embryo failure in fruits developed from early flow- The mean weight of a single seed was 0.003 g in Novem- ers in 1995. ber 1995 but only .0.002 g in October 1995. An examina- tion of randomly selected seeds (n =30) revealed that 97%Seeds of Ferocactus wislizeni are relatively small (< 2 mm of the November 1995 seeds contained embryos whereaslong and wide) and numerous (600 to 1000 /fruit). 40% of the October 1995 seeds did not, which probably Stenocereus thurberi produces even larger numbers of small accounts for their lighter weight. Embryo failure might have seeds (about 1900 /fruit) (Parker 1987). In the northern part been related to unusually dry conditions at the beginning of its range, population age structures indicate that recruit- of the 1995 flowering season. Late summer rains perhaps ment is episodic and that populations are often unstable, enabled mid -season flowers to develop normally. Based on with young plants too few to replace old ones as they die the weight of 500 counted seeds, the estimated number of(Parker 1993). That Stenocereus thurberi persists none- seeds /fruit (± SD) was 671 ± 378 in October and 971 ± 382 theless is due in part to its longevity and in part to high in November. fecundity (Parker 1987, 1993). This is probably true of 6 Desert Plants 1998

Ferocactus wislizeni, as well. During periods that are fa- Bowers, J. E. 1997. The effect of drought on Engelmann prickly pear vorable for germination and establishment, an abundant (Cactaceae: Opuntia engelmannii Salm- Dyck.) fruit and seed produc- tion. Southwestern Naturalist 42:240 -242. seed supply should enable populations to increase rapidly; Fleming, T. H., S. Maurice, S. L. Buchmann, and M. D. Tuttle. 1994. Repro- in addition, high annual fecundity should allow at least ductive biology and relative male and female fitness in a trioecious cactus, some recuitment in all but the most unfavorable years. Pachycereus pringlei ( Cactaceae). American Journal of Botany 81: 858- 867. Gentry, A. H. 1974a. Coevolutionary patterns in Central American Most plants of Ferocactus wislizeni start to reproduce when Bignoniaceae. Annals of the Missouri Botanical Garden 61: 728 -759. they attain a diameter of about 19 cm (7.5 in) (Figure 2), Gentry, A. H. 1974b. Flowering phenology and diversity in tropical presumably the size at which they possess enough water Bignoniaceae. Biotropica 6: 64 -68. and photosynthates to initiate and maintain a few (4 to 6) Gibson, A. C. and P. S. Nobel. 1986. The Cactus Primer. Harvard University Press, Cambridge, MA. flowers. Over a lifetime, annual flower production increases Goldberg, D. E. and R. M. Turner. 1986. Vegetation change and plant demog- by an order of magnitude; only the largest, oldest plants raphy in permanent plots in the Sonoran Desert. Ecology 67: 695 -712. initiate >60 flowers annually. In this study, plants produced Grant, V. and K. A. Grant. 1979. Pollination of Echinocereus fasciculatus 25 flowers /year on average, a very low value for a woody, and Ferocactus wislizeni. Plant Systematics and Evolution 132: 85 -90. Harper, J. L. 1977. Population biology of plants. Academic Press, New York. relatively long -lived plant. This is apparently not unusual Johnson, R. A. 1992. Pollination and reproductive ecology of acuña cactus, in the Cactaceae: Carnegiea gigantea averages only 295 Echinomastus erectrocentrus var. acunensis (Cactaceae). International flowers annually (Schmidt and Buchmann 1986), Opuntia Journal of Plant Sciences 153: 400 -408. engelmannii about 55 (Bowers 1996). Associated trees and McGregor, S. E. and S. M. Alcorn. 1959. Partial self- sterility of the barrel cactus. Cactus and Succulent Journal (U.S.) 31: 88. shrubs with comparable lifespans are more prolific by sev- McGregor, S. E., S. M. Alcorn, and G. Olin. 1962. Pollination and pollinating eral orders of magnitude (e.g., Simpson et al. 1977); sig- agents of the saguaro. Ecology 43: 259 -267. nificantly, their flowers are many times smaller than most Nobel, P. S. 1986. Environmental biology of agaves and cacti. Cambridge cactus flowers. Cactaceae are evidently one of the more University Press, New York. Parker, K. C. 1987. Seedcrop characteristics and minimum reproductive size extreme examples of a trade -off between flower size and of organ pipe cactus ( Stenocereus thurberi) in southern Arizona. Madroño flower number. Apical dominance limits branching in many 34: 294 -303. cactus species (Gibson and Nobel 1986) and thereby re- Parker, K. C. 1993. Climatic effects on regeneration trends for two columnar stricts the number of flowers that can be initiated. This cacti in the northern Sonoran Desert. Annals of the Association of Ameri- can Geographers 83: 452 -474. phylogenetic constraint is especially evident in unbranched Primack, R. B. 1985. Longevity of individual flowers. Annual Review of Ecol- forms such as Ferocactus wislizeni. ogy and Systematics 16: 15 -37. Primack, R. B. 1987. Relationships among flowers, fruits, and seeds. Annual Because Ferocactus wislizeni is common throughout most Review of Ecology and Systematics 18: 409 -430. Roller, R S. 1996. Distribution, growth, and reproduction of Pima pineapple of its range (Turner et al. 1995), one can assume that the cactus (Coryphantha scheeri Kuntz var. robustispina Schott). M.S. the- species is not endangered by low flower production. Evi- sis, University of Arizona, Tucson. dently fruit set and fecundity are high enough to offset what Schmidt, J. O. and S. L. Buchmann. 1986. Floral biology of the saguaro might otherwise be a substantial competitive disadvantage. (Carnegiea gigantea) I. Pollen harvest by Apis mellifera. Oecologia 69: 491 -498. Many factors determine the optimum flower number for a Silverton, J. W. 1982. Introduction to plant population ecology. Longman, plant species, among them availability of plant resources, New York. attraction of pollinators, predictability of the environment, Simpson, B. B., J. L. Neff and A. R. Moldenke. 1977. Prosopis flowers as a and phylogenetic constraints (Primack 1985). To untangle resource. Pp.84 -107 in B. B. Simpson (ed.), Mesquite: its biology in two desert scrub ecosystems. Dowden, Hutchinson & Ross, Stroudsburg, PA. these various elements and assign each its proper weight Stephenson, A. G. 1981. Flower and fruit abortion: proximate causes and ul- would be to gain a substantial understanding of the repro- timate functions. Annual Review of Ecology and Systematics 12: 253- ductive biology of Ferocactus wislizeni in particular and of 279. the Cactaceae in general. Sutherland, S. 1986. Floral sex ratios, fruit -set, and resource allocation in plants. Ecology 67: 991 -1001. Sutherland, S. and L. F Delph. 1984. On the importance of male fitness in Acknowledgments plants: patterns of fruit -set. Ecology 65: 1093 -1104. I thank Steven P. McLaughlin, Annita Harlan, and two Suzan, H., G. P. Nabhan and D. T. Patten. 1994. Nurse plant and floral biol- anonymous reviewers for reading the manuscript and mak- ogy of a rare night -blooming cereus, Peniocereus striatus (Brandegee) E Buxbaum. Conservation Biology 8: 461 -470. ing many helpful comments. Thomson, J. D. 1983. Component analysis of community -level interactions in pollination systems. Pp. 451 -460 in C. E. Jones and R. J. Little (eds.), Literature Cited Handbook of experimental pollination biology. Van Nostrand Reinhold, Benson, L. 1982. Cacti of the United States and Canada. Stanford University New York. Press, Stanford, CA. Turner, R. M., J. E. Bowers and T. L. Burgess. 1995. Sonoran Desert plants: Boke, N. H. 1953. Tubercule development in Mammillaria heyderi. Ameri- an ecological atlas. University of Arizona Press, Tucson. can Journal of Botany 40: 239 -247. Bowers, J. E. 1996. More flowers or new cladodes? Environmental correlates and biological consequences of sexual reproduction in a Sonoran Desert prickly pear cactus, Opuntia engelmannii. Bulletin of the Torrey Botani- cal Club 123: 34 -40. Ferocactus Bowers 7

Table 1. Characteristics of fruits and individual seeds of Ferocactus wislizeni (± S.D. in October 1995 (n = 17).

Fruit Seed

Length (mm) 42.5 (3.00) 2.4 (0.15) Width (mm) 31.0 (3.1) 1.8 (0.15) Weight (g) 16.1 (3.2) 0.003 ( --)

Table 2. Pearson correlation coefficients between fruit and individual seed characteristics of Ferocactus wislizeni in 1995 (n = 17). * = significant at P < 0.001. Seed number, which was calculated on the basis of seed weight, was omitted from the analysis.

Fruit weightFruit length Fruit width Seed weightSeed length

Fruit length 0.218

Fruit width 0.953* 0.148

Seed weight 0.832* -0.013 0.780*

Seed length 0.091 0.481 0.000 0.187

Seed width -0.113 -0.137 -0.054 -0.018 0.387