LIBRARY i Efiflibfin State This is to certify that the thesis entitled Sand Dune Succession: A Comparison of Plant Life History Characteristics ga-« presented by Mary Lou Marino has been accepted towards fulfillment of the requirements for Ph.D; degree inflatanxJflant Pathology Peter G. Murphy Major professor Date JU]! 10, 1980 0-7639 w: 25‘ P" W per 1:» . “I W ,\_'.“~ 3;; a,” ' Place in book return to remove “W '4 charge from circulation records W “Hafiz-g, W 4311:3995 «3° "1805ng i I.“ . 1 _ _ M: _ i )‘i’mfimwt ,, «63. ~' jg =7; -. “”6‘8026920823. éfl-‘Nx s “00:17. i 9 0225 ' 1:1 , I “W 4 5 "we. r . L. m 2 3195-5 . @ Capyright by MARY LOU MARINO I980 SAND DUNE SUCCESSION: A COMPARISON OF PLANT LIFE HISTORY CHARACTERISTICS by Mary Lou Marino A DISSERTATION Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Botany and Plant Pathology 1980 ABSTRACT SAND DUNE SUCCESSION: A COMPARISON OF PLANT LIFE HISTORY CHARACTERISTICS BY Mary Lou Marino This study compares life history characteristics along a suc- cessional gradient on the sand dunes of Lake Michigan. Three life history features were examined: reproductive effort, seed character- istics (weight, number, caloric content), and mortality by age-stage class. Data on seed characteristics, reproductive phenology, and biomass allocation were collected for all reproducing herb and shrub species in the study areas (foredune, slack, dune forest). An energy alloca- tion study was conducted for six species (forest-Solidago £32513, Smilicina racemosa, Polyggnatum pubescens; slack-Monarda punctata; foredune-Solidago spathulata, Artemisia campestris). Demographic data were collected for three species (forestfig, gaggle; slackfifl, punctata; foredunefig, campestris). The reproductive effort studies showed a pattern of energy allocation like that predicted by 'r' and 'K' selection models. Results depended, however, on the measures used, if species of differing morphology were compared. Measures which incorporate more than the current year's allocation of energy may better represent reproductive effort in such comparisons. Mary Lou Marino Intermediate rather than early successional species had the smallest average seed size (.85 mg) and highest average seed number (70 seeds/g biomass). Forest and foredune means did not differ sig- nificantly. Consideration of seed caloric content did not change the observed patterns. The dune's xeric and unstable substrate gives large seeds an advantage on the foredune. The slack's topography, vegeta- tion, and greater soil maturity create more favorable microsites for establishment of smaller seeded species. Large seed numbers improve the chances that seeds will reach these microsites. The pattern of seed specific caloric content was opposite to that normally predicted for succession (5.9 kcal/g foredune, A.“ kcal/g forest). It is hypothesized that in early stages of dune succession, dispersal features and abiotic influences relative to imbibition rates must be balanced against energy requirements for seedling establishment. Demographic studies showed that forest Species had the lowest juvenile and adult mortality rates (37%, 19%). Reproduction followed establishment of a root-rhizome system. This vegetative structure likely allows the individual to compete effectively as well as survive herbivory. Both factors are likely major causes of death in the forest. The time of first reproduction can be delayed in both foredune and slack species depending on growing conditions. Such a delay may allow storage of required nutrients and keep individuals in less vulnerable age classes. The small seeded slack Species had the highest first year mortality (8#%). To my best and most loved field crew: Jack, Grace, and of course Stephen. ii ACKNOWLEDGMENTS My thanks go first to my committee: Dr. P. Werner, Dr. G. Schneider, Dr. J. Hanover, and especially its chairman, Dr. P. Murphy, for their assistance in the writing and editing of this thesis. I am particularly indebted to Dr. Werner who first introduced me to plant demography and whose ideas have greatly influenced my own thinking. I would also like to thank the Luther, Uhl, and Dennison families for the use of their property for this study. I am most indebted to Grace and Jack Marina and to Stephen Welch. Without their help, support, and encouragement this study would not have been possible. TABLE OF CONTENTS Page LIST OF TABLES ........................................ . ........ v LIST OF FIGURES ... .......................... . .................. vii INTRODUCTION ..................... . .............. . .............. 1 Chapter I. DESCRIPTION OF STUDY SITES ...... ........ .............. 6 Location and Floristic Features .. ................. 6 Environmental Comparison ..... ..................... ll Methods . .......... .. ..................... . II Results and Discussion ........................ IA II. REPRODUCTIVE EFFORT . ................................. . l9 Methods ........................................... l9 Results .. ......................................... 2h Discussion . ..................................... .. 28 III. SEED CHARACTERISTICS ........ ...... .................... 36 Methods ........................................... 36 Results .......................................... . 37 Discussion ...... .................. . ............... 57 IV. DEMOGRAPHIC STUDIES .............................. ..... 73 MethOdS o ccccc on. ooooooooo coo-000.00.00.00.oooooooo 73 Results ........................................... 75 Life Cycles and Survivorship ......... ......... 75 Overall Comparison of Life Stages and Life Expectancies ... ........ ...... ......... . 98 Survivorship Curves . ................ . ...... .. lOl Finite and Instantaneous Growth Rates ......... lOS Seed Characteristics ..... .................... . 109 Discussion .. ...................................... IIB V. SUMMARY ............................................... I23 LIST OF REFERENCES ............................................. l28 APPENDIX ....................................................... 135 iv IO. ll. l2. i3. LIST OF TABLES Summary of the environmental comparison among the three study .......................................... Soil nutrient and pH characteristics of the study areas .. ........................................ Reproductive effort estimates for six dune species using annual accumulation (AAR) percentages . ......... Reproductive effort estimates for six dune species using total accumulation (TAR) percentages ........... Mean total accumulation (TAR) percentages for the three study sites .. ................ . ................. Total accumulation (TAR) percentages for semelparous species .................................. Selective advantages and disadvantages of large and small seed sizes in the early stages of dune succession . ...................... . ............ .. Intersystem comparison of seed sizes .. ............... Survival probabilities by age class for species studied demographically ... ...... ... ................ .. Sources of seed set mortality ........................ Proportion of plants reproducing for the first time, by age class ......... ........ ...... ............ Comparison of overall survivorship and life length for the life stages of A, campestris, ‘M. punctata and S, caesia . ........ ........ ......... .. Finite and instantaneous rates of increase for A, campestrls, M, Eunctata, and S, caesia . ........... Page l5 I7 25 26 28 29 62 65 83 85 96 99 l06 Table 1h. Hypothetical rates of increase for species studied demographically ......... . .................... 15. Summary of seed characteristics and reproductive effort for A, campestris, M. punctata, and S, caesia . ..................... ........ .............. Appendix Table A1. Importance values for species from the three study sites ................ ...... . ...... . ............ A2. Degree-hour values for the three study areas ......... A3. Percent-hour data for the three study areas .......... Ah. Windspeed at plant height for the three study sites . .................... . .................... A5. Windspeed at seedling height for the three study areas, 1978 . ............ ..... ....... . .......... A6. Sand movement levels for the three study areas ....... A7. A list of the species studied ....... ........... . ..... A8. Untransformed means of reproductive effort and seed characteristics for foredune species .. ........ .. A9. Untransformed means of reproductive effort and seed characteristics for slack species ...... ....... .. AlO. Untransformed means of reproductive effort and seed characteristics for forest species . ...... . ...... All. Transformed (Ioglo) seed characteristic means of foredune and slack overlap and non-overlap species ..... .... ............. . ........... . ....... .... A12. T-test comparison of transformed (10910) seed characteristic means for foredune and slack overlap and non-overlap species .. .................... vi Page 108 110 139 Ihl 142 lh3 1AA th 1A6 lh8 IH9 150 151 152 10. 11. 12. 13. 1h. LIST OF FIGURES Aerial photograph of study area ..................... . Cross section through foredune site . ................. Schematic of methods used to determine reproductive effort . ...... ...... ..................... Seed weight (a) and energy per seed (b) frequency distributions for herbaceous species from each study site ......... . ............... Energy per gram of seed tissue frequency distributions for the herbaceous species from each study site .. ........ . ...... . ............... Averages of seed characteristics for foredune and slack overlap and non-overlap species ............ Adjusted seed number distributions for herbaceous species from each study site ...... ......... . ......... Seed weight (a) and energy (b) per seed frequency distributions for shrub species from each of the study sites . ........... . ............ Adjusted seed number frequency distributions for shrub species from each of the study areas ....... Artemisia campestris life history . ................... Monarda punctata life history .............. ......... . Solidago caesia life history ......................... Monarda punctata seed with fungal damage ............. Survivorship curves for A, campestris, .M. punctata, andlS. caesia .................. . ....... . ‘ vii Page 21 39 #2 A6 48 52 55 77 79 81 87 103 Appendix Figure A1. A2. Level I soil moisture data for the three study areas . .......... .... ........... . ......... Level II soil moisture data .................... viii INTRODUCTION The sand dunes along the eastern shore of Lake Michigan provide an ideal area to examine the life history characteristics of plants within a context of primary succession. The seral sequence for this area is well established (Cowles, 1899, Olson, 1958). The seral stages are in proximity to one another, and it is assumed, therefore, that the over- all differences observed in life history characteristics between plants from different seral stages should be primarily due to selective factors inherent in dune succession. Few comparative life history studies have been done for primary successional systems. Most research has dealt with secondary succes- sion, in particular old field, prairie, and grassland succession. Some examples are: Blake, 1935; Keever, 1950; Gadgil and Bossert, 1972; Abrahamson and Gadgil, 1973; Gaines et a1., 197A; Rabotnov, I975; Platt, 1975; Werner and Platt, i976; Newell and Tramer, 1978. Several studies have focused on individual dune species or groups of Species with some emphasis on life history factors (Hicks, 1938; Laing, 1958; Van Asdall and Olmstead, I963; Pemadosa and Lovell, I97ha and b, 1975; Hansen, 1976; McLeod and Murphy, 1977a and b; Watkinson, 1978; Watkinson and Harper, 1978; Van der Meijden and Van der Waals-Kooi, 1979). None, however, have made a comparison between seral stages. An additional reason for undertaking this study relates to some of the current life history models, the most frequently cited of which is 'r' and 'K' selection. In this model, the levels of density depend- ent and independent mortality are the prime selective factors in life histories (MacArthur, 1962; MacArthur and Wilson, 1967; Pianka, 1970, 1972; extended to plants - Gadgil and Bossert, 1970, reviewed by Stearns, 1976). This model predicts that in environments where mortal- ity is primarily density independent, species will tend to allocate more resources to reproduction, have larger clutch sizes and smaller offspring, and wait a shorter period before beginning reproduction. These are referred to as 'r' selected Species. Those species found in environments where mortality is commonly density dependent are referred to as 'K' selected and are predicted to have characteristics opposite to those of the 'r' selected species. These sets of traits are often placed within a successional context by associating 'r' selection with early seral stages and 'K' selection with the later stages (Odum, 1969; Harper, Lovell, and Moore, 1970; Gadgil and Bossert, 1972). Grime (197A, 1977) recently suggested that stress as well as levels of density independent and dependent mortality can act as a prime selec- tive factor. It is often inferred from the 'r' and 'K' model that, in environments where density independent mortality is high, resources are not limiting because competitive levels are low. However, this is not always true; some environments are naturally low in one or more important resources, making growth difficult. Thus, Grime feels there are three primary life history strategies which have evolved for plant species.I 'Note that this does not necessarily imply there are only three strategies. Many Species are exposed to differing levels of stress and/ or competition resulting in a gradation between the three strategies much like that suggested for 'r' and 'K' selection by Pianka (1970). His ruderal and competitive strategies have characteristics similar to the 'r' and 'K' selected species mentioned earlier. The third type, stress selected, evolves in situations where primary production is dif- ficult. The traits evolved will include an even lower level of repro- ductive effort than those of competitively selected species, a perennial habit, and a slow growth rate. He makes no comment on progeny size, and gives as examples species from arctic and alpine tundras, deserts, re- source depleted soils, and forest floors. Factors other than stress and levels of density independent and dependent mortality have also been suggested as important in the selec- tion of life history traits. Such factors have been used in the recent literature to explain discrepancies between observed life history char- acteristics and predictions of 'r' and 'K' selection. Examples of such factors are soil moisture levels relative to seed size (Baker, 1972; Werner and Platt, 1975), foraging strategies relative to clutch mass (Vitt and Congdon, 1978), and levels of physical stress relative to clutch Size (Menge, 197A). Wilbur, Tingle and Collins (1974) have sug- gested that close scrutiny of the selective factors (e.g. density of a population relative to its resources, the trophic and successional position of a population, and the predictability of mortality patterns) acting on organism is more important than an assumed model such as 'r' and 'K' selection which takes into account only one factor. Grubb (1977) and Werner (1979) have recently emphasized the importance of studying the entire life cycle of a plant to the under- standing of life history characteristics. Of particular significance are the seedling and juvenile stages. Pelton (1953, 1951), Curtis (1952), Stevens and Rock (1952), Glendining (1941), Shrive (1917), and Ganong (1907) have expressed similar ideas. However, Grubb and Werner approach the life cycle from a highly developed framework of niche and evolutionary theory. Particularly emphasized by Werner is the use of plant demography. Sand dune succession fits well within the framework of 'r' and 'K' selection since density independent mortality likely drops and density dependent mortality likely increases as the soil matures and the vegeta- tion becomes more dense, diverse, and structured. However, nutrients are very low in the initial stages suggesting the stress selection of Grime (1974, 1976). In addition, soil moisture can be limiting espe- cially in the seedling stage. Thus, an array of factors more inclusive than those of 'r' and 'K' selection may have to be considered to explain the life history patterns observed. Three aspects of life histories were examined in this study. The first was the relationship of reproductive effort to successional age. The ideas of Gadgil and Bossert (1970), Harper and Ogden (1970), Ogden (1968, 1972) and Gadgil and Solbrig (1972) were examined as to their applicability to sand dune succession. The second aspect investigated trends in seed characteristics (seed size, number, and caloric content) relative to successional age. The observed patterns were compared to the ideas of Salisbury (1942); Johnson and Cook (1968); Harper, Lovell, and Moore (1970); Levin (1974); MacNaughton (1979); and Werner and Platt (1976). To help explain the patterns observed in the first two studies, a demographic study was conducted for one representative species from each of the successional stages examined. It was hoped that the observed patterns of mortality by life stage would give clues to the type of selective factors acting on dune plants and, as suggested by Grubb (1977) and Werner (1979), help explain why Species with certain char- acteristics are found in particular successional stages. This dissertation is organized into five chapters. In the first, the sites chosen for study are described and their environments com- pared. In the next three chapters, each of the three aspects (repro- ductive effort, seed characteristics, and demography) of life histories considered are examined in detail. Each chapter contains methods, results, and discussion sections. The final chapter is an overall discussion of the findings. CHAPTER 1 DESCRIPTION OF STUDY SITES Location and Floristic Features The study was conducted on private property near Saugatuck, Michigan (Allegan County, Range 16 west, Townships 3 north and 4 north), immediately north of where the Kalamazoo River empties into Lake Michigan. The study area is the large tract of land, approximately 405 ha, outlined in black in Figure 1. Due to its relative inaccessi- bility to the public, this area is fairly undisturbed. Three study sites were selected to represent a sequence of early, middle, and late dune succession based on the botanical and physical descriptions of Cowles (1899) and Olson (1958). These sites were studied for three years (1976-1978). The foredune area (Site 1, Figure 1) represents the earliest seral stage and was located adjacent to Lake Michigan. Figure 2 is a cross section through the area, indicating east and west boundaries relative to vegetative and topographic features. The west border was placed near Ammophila breviligulata's lakeside edge of colonization. Beyond this edge, there were few plants in 1976. The eastern boundary was placed 10 meters beyond the crest of the first dune ridge, where lee- ward windSpeeds were similar to the windward side's. The study site had an area of 1.2 ha with a north to south distance of 234 m. Figure 1. Aerial photograph of study area. The private property the Study was conducted on is outlined in black. Location of the foredune (l), Slack (2), and dune forest (3) sites are shown. .mcouoe om >_oume_xoLaam m_ momcmocson “mo: ocm ammo ecu cooZuop oucmum_o one .ou_m ocsooLOm cmaoccu co_uoom mmOcu .N ocsmmm zones :30 8.6.. 9:6:u . 03E .. . . ct . Stooges . .mp2, foocson Lmom , a. 10 Floristically the foredune is dominated by A, breviligulata. Perennial herbs such as Artemisia campestris, Asclepias syriaca, Solidago spathulata, and Lithospermum carolinlense are less common, but become prevalent near the east boundary. Calamovilfa longifolia is found on the windward face and top of the dune ridge. The annuals Cakile edentula and Euphorbia polygonifolia are more abundant along the western boundary and in the several blow out areas found within the study site. Woody plants are found mostly along the dune ridge. £52333 jagglyg is the most prevalent followed by Sggggg_stolonifera, Populus tremuloides, and Ptelea trifoliata. Table A1 of the Appendix lists importance values for these Species. The Slack area (site 2, Figure 1) represents an intermediate stage of dune succession. However, in appearance it is more similar to the foredune than to the climax situation because it is mainly covered by an herbaceous layer and there has‘as yet been little soil development. The study area was 0.5 ha in size. The slack contains most of the Species found on the foredune with the exceptions of the two annuals Sgfijlg_edentula and Euphorbia poly- gonifolia. The perennial grass, A. breviligulata, is much less dominant, being replaced by Calamovilfa longifolia, Panicum virgatum, and Andropogon scroparius. Additional species not found in the foredune site are Monarda punctata, Arabis lyrata, Hypericum kalmianum, and Oenothera biennis. Both Salix glaucophylloides and Salix syrticola are found along the wet pannes which border the study site. Several Quercus velutina seedlings and saplings were found in the area 11 indicating the advance of subsequent seral stages. Table A1 of the Appendix lists the importance values for these Species. The dune forest site (site 3, Figure 1) represents the oldest successional stage studied. The 0.73 ha site was dominated by a mixture of Acer saccharum and Quercus rubra. Other tree species found in the area include 35293; serotina, Egggg_grandifolia, Tilia gmericang, Quercus ellipsoides, Ostrya virginiana, and Tsuga canadensis. A list of the forest floor species and their importance values is given in Table A1 of the Appendix. Among the most prevalent were Smjlg§_ rotundifolia, Polygonatum pubescens, Viola papilionacea, Mitchella repens, Hepatica americana, Viburnum acerifolium, Solidago caesia, Oryzopsis racemosa, and several Species of Lycopodium. Environmental Comparison Methods Environmental data were collected for each of the study areas so that the sites could be characterized and compared quantitatively. The differences between the slack and foredune areas are important since these areas appear physically similar, but have several floristic dif- ferences. Environmental factors of particular interest were air tem- perature, relative humidity, wind speed, sand movement, and soil moisture because these all have bearing on seedling survival. Monthly hygrothermograph (Belfort, Model 5-594) readings were taken in each area from July 1977 to July 1978, excluding December 12 through March. The instruments were placed on the groundz, near the center of the study site, and ventilated A-frame shelters were used to eliminate radiation error. The curves obtained from the hygrothermographs were integrated over time so that data for a given interval could be reported and compared as one number. The area under the curves was used rather than an aver- age because it was felt that this better represented the temperature and relative humidity experienced by a plant during the indicated interval. For the integration of temperature curves, 32°F was used as a lower boundary because most plant metabolic processes stop at or very near this temperature. Thus, temperature data (degree-hours) represent the amount of heat above the threshold temperature which occurred during the indicated interval. The choice of a threshold for relative humidity was more arbitrary; thirty percent was chosen because it was a multiple of ten and was slightly less than the lowest humidity reading recorded. Thus, humidity data (percent-hours) indicate the total amount of rela- tive humidity above 30 percent for the indicated time interval. Sand movement was measured by placing lO-ounce Styrofoam cups in the 5011 so that their rims were even with the soil surface to allow deposition of wind blown sand. The amount accumulated in the cup approximated the quantity of material deposited on a surface equal in size to the mouth of the cup for the period of time the cup was left out. A grid system was used to place the cups in each study site with 2These readings represent air temperature at O-lO cm from the ground surface. 13 nine cups in the woods, eleven in the slack, and twelve in the foredune. Collections were made once a month during the growing seasons of 1977 and 1978. A hot-wire anemometer (Hastings airmeter, Model RBI) was used to measure wind speed. Readings were taken during the growing season of 1977 and 1978 in each area at marked stations set up along a transect. In the foredune site, the transect ran perpendicular to the beach, with four stations: one on the dune plain (see Figure 2), one on the side of the ridge, one at the ridge crest, and one near the east boundary. In the slack and forest sites, the transect ran through the center from the east to the west boundary with three stations placed equidistant from one another. Twenty readings were taken at each station at five second intervals to allow for the variability of wind speed. Readings were taken at 1 cm and 30 cm heights to approximate seedling3 and adult heights, respec- tively. Because there is a diurnal variation in wind Speed, the read- ings for each study area were taken as close together in time as possible within the period of 0900 to 1700. An average wind Speed was obtained for each area by combining all data from all the Stations in each area. Soil moisture information was collected in 1978 at two week inter- vals, using Simple gravimetric measurements. Sample cores (4.5 cm deep by 6 cm diameter) were taken at the stations used for wind Speed. At each station three samples (termed Level 1 samples) were collected from 3Data on seedlings were taken only in 1978. 14 the tap 4.5 cm of soil using six ounce soil cans. Three more samples (termed Level 11 samples) were taken from the 4.5 cm of soil directly beneath the first three. Samples were collected only to this depth because information on moisture at the seedling level was desired. Due to the drastic difference in soil texture between the forest and the other sites, moisture content was eXpressed as grams of water per volume of soil as suggested by Salisbury (1952) and L. R. Stone (pers. comm.). Averages for each study Site at each soil depth were calculated. Data on soil nutrients (nitrate, total potassium, and total phos- phorus) and pH were obtained from 20 random Six inch soil cores taken at each of the three study sites. The analysis was performed by the Soil Testing Laboratory of Michigan State University. Results and Discussion In the analysis, the data for all sampling dates for a given environmental factor were taken as a group for each study area and a comparison was made. Where strong periodicity led to overlapping values for different sites, a pairwise comparison (Fisher Distribution Free Sign Test described on page 39 of Hollander and Wolfe, 1973) which took date into account was used. Otherwise, the Wilcoxon Rank Sum test sufficed. Table 1 summarizes the results of these comparisons. The l‘The sampling design was set up through the help of L. R. Stone of the Evapotranspiration Laboratory of the Department of Agronomy, Kansas State University. 15 ._o>o_ oocmo_m_cm_m oumo_oc_ meonESz .m~m_ .mnm_ .oe_oz ucm totem—.0: mo mm mama eocu .om_oz tam Leone—.0: mo mm ommo EOLu m N .oococoem_o acoo_m_cm_m o: no: meosu memos cm_m _m:uo c< .mc_omo; ocu co ou_m cozuo may mo omozu coco Ammo—v Louoocm >_ucou_m_cm_m ego: anoLm o no coxou con; £0.53 mucoEocammoE on; mc_omo; can—co ozu c. awe—m ooum__ omen ozu umzu mo__as_ nvv.A we >Laco c< _ 53m xcmm coxoo__3 mo.o.v. mo.o.v. n _. _o>04 Eam xcmm coxoo__3 mo.o.v mo.o.v mo.o.v _ _o>04 AmEU\mv ocaum_oe __0m masm xcmm coxoo__3 mo.o.A mo.o.A mo.o.A A>mo\MEo\mv ucoeo>oE ocmm .0902 is. 8:: 36K .o.o.A 35K .5 _ moumo__aoc ooe_mm _o.o.A _o.o.A n Eu cm Auom\EV woman oc_3 moumo~_aoc ooc_o¢ .o.o.v _o.o.v mo.o.v. AmL2051ucouLoov >u_v_E:; o>~um_o¢ Nmouoo._aoc ooc_mm .o.o.A _o.o.A mo.o.A Amcao;noocmoov ocaumcoaeop amok umoLOuuxum_m umoLOmnocsooLou xom_mnoc:ooE0m o_ummcouooco;o _oucoEc0n_>cu mmc_oooc ass—00 .oocmaeoo mm: mou_m mo L_ma ;o_;3 oumu_uc_ .mou_m >osum oocsu one moose cam_umoaou _mucoscoL_>co ocu mo >LoEESm ._ m4mcou:_ oucoo_mcou acouLoa mm oghw .oo. x.NWA>mcoco boom—asauum .ou0uv\flco_uo:o0caoc cu >mcoco m.coo> acoccaywa 8.. - mé 2.3 .. mé n.o :.m: m... mcoomonam153umcom>_0a unocou a; t 98 as. .. .3: m._ m.mm m... amoeoumc ocuu___sm umoLOu :3 - .1: ad .. mé o.~ m._m 3.: m_momo ammo—.0m umoLOu a; .. ad :6. - ma: o.m m.n_ .m.m_ oumuocsa mucosa: xom_m ad - ..3 GA: . mg: m.m ..n. m.:_ mum—scumow Omoo__0m ocaoocou «8.3 - 58 «3.3 ._ ma: w... m... «.mu m.cumomaoo o_m_eouc< ocaooLOm cmo> ucoccao ¢<< scam acoocoo c. c. oooom >mcoco co_uu:ooc _ m_uo:o0caom .: u4mAge of Site where Species Found—9 Figure 7. 47 Adjusted seed number distributions for herbaceous species from each study site. A 10910 transformation was used to nor- malize the distributions. Thus, the parameters given with the distributions have been transformed. The mean for each distribution is Shown and the bars to either side indicate the standard error. Frequency ( percent) 48 Foredgne 3? =1.49 15 .. SE. = .097 Mean N g 79 t—r-I 10 ‘- Fl 5 - 'l - r11 1- “1 H- , II- I '5 .1. £19.93. x = 1.85 SE.= 12 101- T N3 115 5 w h . . Ha mfi II” lHIrH—i I'll IT‘I '5“ :rest x = 1.35 SE. = .058 '0 __ _ I N s 145 1 H .. 5 .1. .rLHTIT Henri O .4 .e 1.2 1.6 2.0 2.4 2.8 32 Log.o (adjusted seed number) 49 distributions. Again, the data were separated into the overlap and nonoverlap groupings described earlier. The slack nonoverlap Species' average seed number is significantly larger (P < O.OI)ZI than the foredune nonoverlap group's average seed number as well as that of both overlap groupings (bottom, Figure 6). (Table All shows actual numbers and statistical comparisons.). Though the foredune nonoverlap group has the smallest mean, this mean is not significantly different from the averages of the two overlap groupings. Combining the information on seed weight (or energy) with that of seed numbers, it would appear from Figure 6 that dune Species (foredune nonoverlap) frequently colonizing the most environmentally stressful and least Stable portions of the dune system tend to have a smaller number of large seeds whereas those species (slack nonoverlap) found most frequently in the more stable and less stressful slack area tend to have a large number of small seeds. Those species found in both types of areas appear to exhibit seed characteristics intermediate between these extremes. Up to this point we have examined only the foredune and slack distributions of seed characteristics in detail. Examination of the forest's distributions is also enlightening. Returning to Figure 4, it can be seen that though the Slack and foredune distribution of seed weight and energy per seed are skewed, the forest's distributions encompass the entire range of values exhibited by both of these earlier 2lThe Behrans-Fisher Test was used. 50 seral stages. Use of the Kolmogorov-Smirnov Test shows a significant difference (P < 0.05) between the forest's distribution and those of the slack and foredune. For seed numbers (Figure 7), the forest's distribution primarily encompasses a smaller range of values than do the slack and foredune distributions. Thus, it would appear that though forest Species exhibit seed weights and energy per seed values which, on the average, are larger than those of the slack and equal to those of the foredune, the individual species exhibit a wider range for these seed characteristics. This is less true for seed number. Distributions of (loglo transformed) seed weight and caloric content per seed for Shrubs Species are shown in Figure 8. As was the case for herbs, the slack averages are smallest, and the foredune and forest means are Similar in size both for seed weight and for energy per seed. However, it could not be determined whether these differ- ences were statistically significant due to the trimodal distribution exhibited by slack species for these characteristics. However, com- parison of the distributions themselves is enlightening. All are Significantly different from one another (P < .001).22 The foredune distributions are not as variable as the slack's, but are more so than the forest. The largest seeds of the foredune belong to Prunus pumila, which is the most abundant shrub of the foredune area. It is found on the foredune crest as well as the dune plain, and can establish on blowing sand (Cowles, 1899). Of middle seed size is the far less 22The Kolmogorov-Smirnov Test was used. Figure 8. 51 Seed weight (a) and energy (b) per seed frequency distributions for shrub species from each of the study sites. All data were transformed using a 109.0 transforma- tion. The transformed means are indicated on the distributions as are the standard errors. Frequency (percent) 52 a. b. 30 _m 30 . Foredune Y '1.°74 M ‘2' '1.815 5.5 - .0431 5.: - .0517 n-eo . zo~- 20. u so M sue m x- - 1.175 30" Ynez 30.. senor S.E.- .109 "-100 u -100 20 " 2° . lo 5 4% '0 - l—H7 I I I l ,,4 . r-s . . . H A '1 . n A I A . nI-l‘ 40 Forest Forest Y-use Y-Ieeo 3° S.E.I.0136 3° 55-0142 "“0 11-40 20 . 20‘ 1° ‘ '0 ., -1fo . 4: ‘-.2 0 .2 e 10 14 Is -.4 0 a .e 1.2 1.6 20 2.4 Lot:l0 (seed weight(mg)) Log.0 (kilocalories per seed) 53 abundant Egglgg trifoliata which is able to colonize bare sand, but at a high rate of mortality (McLeod and Murphy, 1977b). It establishes best in thickets of more stabilized areas. The smallest seed belongs to Spgpg§_stolonifera. The two thickets present in the study are both in protected areas just beyond the crest of the first dune ridge. Though a good dune former, its seedlings most frequently establish in moist protected areas (Cowles, 1899). The slack area, with its trimodal distribution, has species which have three different Site requirements for germination. The largest seeds (ErunuS pumila) can colonize open sandy areas. The smallest (Hypericum kalmianum and Salix glaucophylloides) are dune swamp species and need moist protected areas for their seedlings to survive (Cowles, 1899). Species of middle seed size (Arctostphylous uvi-ursi and 11515 riparia) invade only areas which have been stabilized by other species (Cowles, 1899; Olson, 1958). The forest shows the least variability of the three areas in seed size since only two species are represented (Viburnum acerifolium and Euonomus obovatus). Both must develop woody tissue in shaded closed communities, thus, selection on seed size may leave little room for variability. Figure 9 contains the distributions of adjusted seed numbers for Shrub Species. The forest mean (0.599) is significantly smaller 23 The latter (P < 0.05) than both the foredune and slack area means. two means cannot be distinguished from one another (1.828 and 1.820, respectively). 23The Kolmogorov-Smirnov Test was used. 54 Figure 9. Adjusted seed number freguency distributions for shrub Species from each of the stugy areas. Data were transformed using a loglo transformation. Transformed means and standard errors are indicated on the dis- tributions. Frequency (percent) 55 30 1' M 7:18:18 s.1-:.=.1874 20 .. Mea_n_‘ _ N= 16 10 ~- 11 11 51 r 30 " Slack 3?: 1.320 3.15.: .1818 20 “t N = 34 __ 1—1—4 10 «~ T h-] l T 1 n J 41 L l 30 -- Ems: :1 3?: .599 '1 S.E.= .0786 2° " N = 17 10 -- 2 —.6 10* ‘14 18 l 22 21.6 3.0 L 3:4 Log|o (adjusted seed number) 56 The distributions for the Slack and foredune areas are not Sig- nificant by the Kolmogorov-Smirnov Test. Both are spread over a range from 0.6 to 2.8 with the smallest seeded plants having the largest seed numbers. However, for the slack, the largest seeded species have intermediate numbers. The forest distribution for adjusted seed numbers is more variable than it was for seed size. However, it is much less Spread than the other areas. It appears that the forest Species have large seeds, but few of them. Medianzu kilocalories per gram of seed tissue for shrubs are: 5.734 kcal/g for the foredune, 5.189 kcal/g for the slack, and 5.713 kcal/g for the forest. None are significantly different from one another since the ranges are so large (4.476-5.937 kcal/g for the fore- dune, 4.862-5.94l kcal/g for the Slack, and 5.171-6.320 kcal/g for the forest). Shrub and herbaceous medians are similar for both the slack and foredune areas (foredune--5.734 kcal/g for Shrubs versus 5.68 kcal/g for herbs; slack--5.184 kcal/g for shrubs versus 5.281 kcal/g for herbs) whereas those of the forest are significantly different (5.713 kcal/g for Shrubs versus 4.411 kcal/g for herbs, P1< 0.05 by Wilcoxon Rank Sum). 2liMedians are shown because the Wilcoxon Rank Sum Test was used, and this test compares medians rather than means. 57 Discussion Unlike other studies which have Shown that seed size increases with successional age (Salisbury, 1942; Harper, Lovell, and Moore, 1970; MacNaughton, 1975; Werner and Platt, 1976; Newell and Tramer, 1978), this study found that in dune systems seed size is Similar for the earliest and latest seral stages. It is in intermediate seral stage that the smallest average seed size is observed. The pattern for seed numbers was exactly the reverse of this with the largest numbers occur- ring in the intermediate stage and the smallest in the earliest and latest seral stages. Again, this contradicts the findings of other studies (Salisbury, 1942; Cody, 1966; Johnson and Cook, 1968; MacNaughton, 1975; Werner and Platt, 1976; Newell and Tramer, 1978). These apparent contradictions can be eXplained, however, if the selec- tive pressures acting in the different seral stages are examined. The early stages of sand dune succession are edaphic deserts (Salisbury, 1952). The low moisture holding capacity of the sand, drying effects of the wind, and high soil temperatures (sometimes > 50°C) result in moisture levels in the first few centimeters of sand being less than 0.05 percent of field capacity (Salisbury, 1952; McLeod and Murphy, 1977b). As indicated by the environmental data, the soil moisture in the first 4.5 cm of the foredune sand can fall below 0.002 g of water per cm3. This, combined with the high air tempera- tures, wind speed, and soil temperatures, result in a poor environment for seedling establishment. However, soil moisture increases with depth and at approximately 25 cm is permanently wet (Salisbury, 1952; McLeod 58 and Murphy, 1977b). The closer a seedling roots are to this area of permanent wetness, the greater the seedling's chances of survival (McLeod and Murphy, 1977b). Soil nutrients are also low, especially on the foredune, making root growth difficult. However, a large seed would allow the development of an extensive root system in the face of these low nutrients much as it would in the competitive situation. Also, due to the instability of the dune substrate, the chances are high that a seed will become buried. A large seed would provide reserves for stem growth to the sand surface, and a deep root system would provide anchor- age if sand is blown from around the seedling. Evidence for the impor- tance of large seed Size in xeric systems also comes from Baker's (1972) study of the California flora and Went's (1955) study of desert peren- nials. The environmental section demonstrated that there were differences in windspeed, air temperature, humidity, soil moisture, and sand move- ment between the foredune and slack areas. With the combination of these differences, the foredune becomes the least hospitable for seed- ling establishment. However, xeric and low nutrient conditions exist in the Slack area as well. The question arises as to why there is such a difference in average seed Size and number between these two areas. The answer lies in the comparison of overlap and nonoverlap species. As was discussed in the Results section, those species found pri- marily on the foredune, on the average, have the largest seeds and the lowest seed numbers, whereas those found only in the slack are just the reverse. Those which overlapped were intermediate in size and numbers. 59 It is possible that the combined effects of reductions in the Stress factors mentioned earlier allow enough favorable microsites so that smaller seeded Species can establish. Of particular importance would be the higher soil moisture and nutrient levels as well as protection provided by topographic relief and vegetative cover. The importance of the microsite in the establishment of seedlings has been demon- strated by Harper et a1. (1965), Sheldon (1974), and Domes and Elberse (1976). Even on the foredune, some of these smaller seeded species can infrequently be found; however, these have usually established in pro- tected areas behind the crest or in Shaded and more vegetated areas. Since most foredune or slack species are wind dispersed, the higher seed number of small seeds found in slack nonoverlap Species would allow greater dispersal distance and would increase the probability that enough suitable microsites will be found for replacement and/0r population in- crease. The large numbers of seeds would also be important, because the mortality of seedlings established from these small seeds is likely high, due to xeric conditions of the slack area and to the random location of favorable microsites. The larger seeds of the overlap group allow establishment in either area. Of interest in the overlap group are the two Species which have the smallest seeds (Agtemisia campestris and Solidago spathulata). In both Size and numbers, they are Similar to the slack nonoverlap group yet they have been able to establish sizable papulations on the fore- dune, apparently contradicting what was said earlier. However, both have adaptions to compensate for their small seed Size. The fruit of 60 .A- campestris is an achene which when wet provides a mucilagenous seed cover. Mucilage allows a much faster uptake of water and can make germination levels less sensitive to water supply (Harper and Benton, 1966; Oomes and Elberse, 1976). In addition, the dispersal unit is 25 far from spherical , increasing the surface area and imbibition rate (Harper and Benton, 1966). The specific energy content of the seed is among the highest of any species found in the three study areas. The high imbibition rate and lower moisture requirement for germination could allow for fast germination and root system establishment. The high caloric content allows for an adequate food reserve. The seeds of the other small seeded overlap species, Solidago Spathulata, appear to germinate on the foredune only infrequently, and if they do, the seedling may survive only a short time. .S. Spathulata seedlings were looked for when A, campestris and A. punctata seedlings were surveyed at two week intervals. Though several were found in the slack area, none were found in the foredune. This Species can, however, reproduce vegetatively by budding from its roots. Once the ramet is large enough and its own root system is well established, the connection with the parent plant degenerates. The distribution of this Species is highly clumped on the foredune and in many cases radiates from behind the dune crest and/or from more highly vegetated areas. Thus, it would appear that seedlings can infrequently establish in protected areas, and further colonization Slowly radiates from these 25The length is more than two times the width and the width is 1.75 times the depth. 61 points through vegetative propagation and very rare seedling establish- ment . Though a large seed can be advantageous in a xeric environment, there are concomitant disadvantages as well. These are listed in Table 7 along with the advantages discussed to this point. (The same is done for small seeds.) These disadvantages may place constraints on how large a seed may become. Though average foredune seed Size was larger than the forest's, there were several forest species of much larger seed size than any foredune Species. A possible constraint is imbibition rate. Imbibition takes longer in larger seeds because of their low surface to volume ratio (Harper and Benton, 1966; Harper, 1977). In the foredune and slack areas, periods favorable for germina- tion are Short. If a seed germinates too Slowly, conditions may become too harsh to allow establishment of a root system. Thus, the advantages of a large seed Size would be lost. Imbibition rates are even slower if the seed is on the soil surface because water is more easily lost from the seed to the surrounding atmosphere than if the seed were surrounded by soil (Harper and Benton, 1966). The probability is higher that a seed would be buried on the foredune than on the slack, due to the greater instability of the foredune's substrate. It is possible that seeds of equal size may more frequently have imbibition difficulties in the slack than they would on the foredune. Thus, the slack areas may have even greater constraints on large seed size than the foredune does. This may help explain why larger seeded species are teem >n o0_c:n m. poem w_ oummcam sumac team an oopsm c0» > . a oo_c:n m. boom m_ u_:0_mm_o 0L0: N ouomcam cu oumcuocoa o. >__mmo aces coo goesm .~ 6 econ mm_ co_umoo__o 0>_u0:o co_umoo__m tsp—n c_ o m . ragga; co>~m m com o>_uuaooLa0c c0>_m econ icosucm mu_E_A m oooaooca mooom ocoz .m o cow count mc_3o_n c_ own race mooom Luzon .m recsocm mo>oLaE_ .m mco_u_ocoo unmaoLo coon: co_umc_Ecom t.amc o.;mco>_>cam egos cu mcmooo_ cowumc_Ecom unmaoLo . . moumc coma >m_0o >mE moon; gm30cgu n.5m maum0__oeoo < t_n_ne_ L0;m_z .N comu_n_ne_ L030; .N 1co>_>c:m mouoeoLa .< ucoeao_0>oo “00L o_amc mc0_u_v Low a_nm__m>m .mmcoam_o _omeoa taco u:o_cu:c 30_ c. moccaomoc Luzon ._ oc_3 e0>0caE_ ._ tm_o oc_3 c0L00a ._ ucoeao_o>oo poo; o_am¢ .— mommuco>vmm_o mommuco>o< mommuco>omm_o . momoucm>v< mooom ..mEm mooom omcoml. H in E .c0_mmoouzm ocao mo mommum >_Lmo one cm m0~.m boom ..oEm ocm omen. e0 momouco>omm_v can mommucm>om o>_uuo_om .h m4m9 .23 (77) -12 -12 .26 (78) .845 (78) Juvenfle .84 (77) .86 (78) I1 I11 ($7.! ..1.’ I.»- h.“ I) 1 l 82 Table 9 contains the probability of survivorship for each of the stages found in Figures l0-l2. A separate table was needed to present these because the numbers in Figures lO-lZ combine survivorship and the proportion of a given life stage which took a given path to the next life stage. Comparing Figures lO-l2, all three species release seeds in the fall:.fl. punctata in mid-September, A, campestris in mid to late October, and _S_. _c_a_e_§_i3 in late October to early November. In l977, .fi. punctata had the highest seed viability at dispersal (75 percent, Table 9). A, campestris was next at 6] percent, and §, gaggig_was lowest at 40 percent. Viability does not appear to be consistent from year to year since I978 figures for 5. punctata (5h percent) and A, campestris (65 percent) are lower than l977 and s, 533313 is almost two times as high. Comparing unchilled and prechilled treatments, all three species do not need a cold treatment to germinate since percent- ages are similar for both treatments. Some of the causes of inviability are shown in the second grouping of numbers in Table 10. All of the categories are self explanatory except 'empty fruit'; this represents seeds which developed, but the endosperm had been eaten without any visible outside damage. This was less frequently a cause of inviability in A, campestris and g, ggggig (3.79 and l0.h0 percent, respectively), but was significant in‘fl. punctata (l6.9 percent). A fungus, Coelomycetes (H. lmshaug, pers. comm.). was always present inside these seeds (see Figure 13). However, it could not be determined if the fungus actually caused the death of the seed or if it invaded after a seed became inviable. 83 _m. mm. m:. mm. m0. m0. __mco>o oz oz on. oz 0.0 oz Lmu> vacuum oz oz mm. 00.. do. #0. Lmo> among nu_:u< mm.1mm. oz ma. _m. mm. _m. __mco>o oz oz oz oz mm. oz Loo> ucooom marmm. oz om. 3. 3. om. .89» ow... _ h. m uo__co>:w mu. MN. Nm_. m_. _N. mu. co_umc_ELom gossamimc_oam <2 <2 Ro. oz 8. oz :2 “2.25% :8 <2 <2 mm. oz mm. oz Aco_umc_ELom __mmv Louc_3co>o N N m “Lmo> umc_m :uo. v.0. muoo. m1o_Xm mwoo. duo—xm gossamumc_cam <2 <2 mn_oo. No.o m:_o. ~o.o __mu J "co_umc_Ecoo mom. R9 mom. Rm. mom. Rm. €833 m_momu omov__0m mumuocza monocoz m_cumooemu m_m_EouL< _.>__mu_;amcmosov po_v:um mo_ooom Lem mam—o 0mm zo mo_u___nmnoLa _o>_>L:m .m woo vacuum c_o:u cz >__o:uum mom mucm_a mommum o__co>:m ogu mo Lmo> «mg.» osu m_ m_:h m m .o_om—_m>m want 02 .mo_uoam m_;u op >_aam uoc moOo+~ .cOmmom mc_zocm mum. c. vooo_o>ov mc__voom pan .mum. mo _—mm c. poumc_ELom 200mm .cOmoom mc_zoLm mum. c_ voao_o>ou mc__voom was .mnm. *0 __mm c_ paumc_ELom voomu .uxou c. _m ouOcHOOm vcm w_1o_ moo:m_u oom co_umELOmc_ Loougam ooh .umsm:< concocu __ca< mo vo_coo ozu Low mm; co_umc_Ecom Love: >Loo0umo .LoEE:m1mc_com_ och .Loneo>oz new concuoo mo mcucoe ocu Lam mm: cooumc_ELom cove: >Lomoumo ...mm. ooh .uooavoLa momma mo Logan: ozu cu o>_um_og voumc_ELom mpoom mo co_uL0aoLo ecu mm ovum—:o_oo moo: mo_u___nmooLa co_umc_Eooo ._~Lo<1p_E o» Loo0u001v_s mo vo_ooo guc081w ozu Low m_ m:_couc_3Lo>o Lmo> awn—u .Lmoz oco ocm oc_couc_zgo>o Lmo> umc_$ vcm co_umc_Ecom com macsu uaouxo mo_comoumo __m com uo_coa oE_u ooh— .m. _m. <2 <2 o.o oz mcmuszLooccoz u:32 mom. -m_ mom. Nomo momo msmw zoomoumo m_momu ommv__0m mumHUcam mpomcoz m_cummemu m_m_Eouc< .Au.ucouv m uoo<» 85 .mo>—omEocu muoom co acou_>o 0:02— m.mm m.mm _.m 0.5m E:__0chuou >o n.0m ~.o~ :._: o.: ommEmv _mmcam Lo uuomc_ >o o.om o.on m.om o._: "o.om_>c_ _. ~.~N ~.m~ o.o. >Lo>ouoL uoxuma >o paumc_ELom m. w._ m.ou o.o: E:__oumL90u >o oo.o_ o.:~ o.o: o.mm A "2%.; v " mvoOm mum. .E:__oumcuou Louu__ mmo_ ucmm mumuu m:_couc_3Lo>o mo mu_:mo¢ Ao.mmV Ao.mhv Ao._mv AE:__0NMLuou >3 o.om_>v ~.m_ m.m «.mm E:__0chuou >o o_nm_>c. o.mm _o.o _o.o ommEmv gamma. :.o_ o.o. w.m poacm >uaEm "Amvoom mum. .E:__oncu0uv voumou mvoom mo ucoogoa m mm _mmcoom_v um mooL30m >u__mucoz o.oo o.oo o.om e=__oumouop "mum— o.~: o.mn o.om co_uoc_ELom no___;oco o._: o.mn m.~m co_uocmELom vo___;oocm o.mm o.os o._m E:__o~mouoh "num— o “_mmgoammp um >u___nm_> acoogoo g 8 g mumuocam m_Lumomamu m_mooo omop__0m mvcmco: m_m_EouL< >Lom0umo .>u__mucos uom coon mo mooL30m .o. w4o 0.05. Only numbers with standard errors were compared statistically; standard errors are parenthesized. 2The comparison between foredune and Slack is not significant. 3 The comparison between foredune and forest is not Significant. 111 species are significantly larger. This is partly because the Solidago has the smallest (non-parasitic) seed of the forest species examined. If another species had been chosen from the forest, this species could have had the largest seed; however, the Slack Species would still have had the lowest weight. The forest Species has a significantly greater Specific caloric content (P<0.05) than that of the two earlier successional species. The slack and foredune species were not significantly different. This does not follow the overall pattern observed earlier for the three study sites. This is explained by the fact that only for forest species is there a negative correlation between seed size and specific caloric content. Because S, ggggjg has the smallest seeds of the forest, it has one of the highest Specific caloric contents and thus deviates greatly from the average for the area as a whole. If a large seeded Species had been studied instead of S, ggggjg, the seed size pattern would not have been observed but the pattern for specific caloric content would have. The relationship among the species for energy per seed is similar to that observed for seed size. However, because of the larger Specific caloric content for S, £22313, there is no significant difference be- tween A, campestris and S, caesia in seed energy content. In 1977. A, campestris had the highest seed number followed by ‘A. punctata and then by S, caesia. This pattern still applies when l‘oThe Student t-test was used. 112 numbers are adjusted by plant Size (category 5). This does not reflect the overall pattern for the areas, but does reflect A, campestris's semelparity. For 1978, A, campestris's seed numbers were greatly decreased, and, thus, dropped below those-of.A. punctata. The seed number for both S, ggggjg and'A. punctata increased slightly in 1978. The expected life time seed production for an adult plant was calculated by using the following summation: I'l x..N. :EE: 'J J i=1 i' probability a plant lives to reproduce in year i J for species j. number of seeds produced per year for species j. Z I i - the number of adult years, 1 to n. species. 1.... I It was assumed that the plant had reached the first year of adulthood. Looked at this way, the slack adult has produced more than two times more seeds in its lifetime than an A, campestris adult (1977 numbers). For 1978 the difference is even higher. However, A, campestris com- pensates for this by having higher seedling survivorship than A. punctata. A, campestris can more than replace itself in a good year if seed production is at 1977 levels, but it will likely never reach the density on the foredunes which A, punctata achieves in the slack area. 113 The lifetime seed production of the S, ggggig adult is much lower than the other two Species. Though the adult plant has a long life, reproduction iS restricted because in some years the plant may not reproduce. If at some time plants can return to flowering, the number of seeds produced will increase, but for the calculation it was assumed that this did not occur. Even if an adult is unable to resume flower- ing, this population, as its rates of increase indicate, is more than replacing itself due to the higher survivorship of all life stages. The last category of Table 15 shows reproductive effort percent- ages. A, campestris is highest for both the AAR and TAR percentages. It is followed by.A. punctata and then S, ggggjg, This pattern was examined more closely for these species in the chapter on reproductive effort. Discussion The results from the demographic studies parallel the findings of the seed characteristics study. The species exhibiting the most 'r' selected traits (shortest generation time, highest rates of increase, smallest seeds, and largest life time seed production) is the Slack species, A. punctata, rather than A. campestris which appears earlier in the successional sere. As expected, S, ggggjg exhibits more 'K' selected traits than the others (largest generation time, lowest life- time seed production, very high adult survivorship). But even this species exhibits unexpected characteristics such as a seed size smaller and a rate of increase higher than A, campestris's. However, as was the 119 case with overall seed characteristics, these seeming contraditions may be explained by examining the selective factors likely acting on each of the species. 0f the three species, S, ggggig is likely the least affected by changes in environmental factors. This was well demonstrated by the dry conditions of the 1977 growing season. The germination levels of S, caesia dropped only 30 percent whereas those of A, campestris and .A. punctata dropped 77 percent and 90 percent, respectively. S, 222513 adults showed less signs of water stress than those of the other species, as well. Biotic causes of death appear to be the most important for S, £32313. Seedlings often showed signs of predation before they died. Frequently, the seedling would completely disappear, whereas A, campestris and A. punctata seedlings would show signs of desiccation before dying or disappearing. Most of the S, ggggig adults which died were those which were highly insect damaged in the previous growing season. Competition, though not experimentally demonstrated, likely plays a role as well. Roots, in the dune forest, are densely matted in the thin upper soil layer. Though the forest soil is the most mature of the successional sere, it iS Still very low in nutrients. Thus, resources are likely limited for the species found there. S, £22313 had the highest germination levels of the three species Studied. This is understandable since S, £22313 seeds have a more continuously moist environment in which to germinate than do.A. punctata and A, campestris seeds. In some years, germination conditions are so 115 frequently unfavorable in the Slack and foredune areas that almost no germination occurs there. Seedling survivorship levels are likely higher in S, Egggj§_than in A, punctata because the slack Species has a small seed and germinates in a xeric environment. The small food reserves make it difficult for a seedling to develop the root system necessary to allow survivorship through low moisture periods. Thus, mortality is high for this stage. However, a small seed size likely contributes to mortality in S, Egggjg seedlings as well. S, gggSjg has one of the smallest (nonparasitic) seeds found in the forest herb flora. Thus, competitive factors may contribute to mortality among these seedlings, and may exclude S, 523512 from certain areas of the forest floor. The most frequent germination Sites and areas of highest seedling survivorship for this species were places of Sparse vegetation and low litter cover. Seed predation likely contri- butes to this pattern as well because the study of overwintering via- bility demonstrated that seed predation by insects was higher in leaf litter than in sand. A small seed size may also place S, £32312 seedlings at a dis- advantage when fed upon by predators. Without the resources to quickly establish a root rhizome system, any loss of leaf material would be critical in the low light levels of the forest. This may explain why seedlings with predator damage frequently died. With these disadvantages of a small seed, the question arises as to why S. caesia's seed is not larger. The answer may well lie in this 116 Species' dispersal mechanism. S, 532312 is wind dispersed, as are all other members of the genus Solidago. Wind levels are low in the forest. Thus, seed size must be small to assure that seeds are carried over a large enough area to secure the microsites necessary for the species to reproduce itself. There is evidence, however, that there may have been selection for a larger seed in this species Since S, caesia's seed is among the largest in the genus Solidago (Werner and Platt, I976). The seed numbers for S, £22513 are among the highest for forest herbs. This probably helps to balance seedling mortality. It also increases the probability that enough seeds will reach favorable micro- sites to allow survivorship of the Species in the forest. The root rhizome system is likely very important to the survivor- ship of the S, £33513 adult. An S, £32312 plant does not reproduce until it is at least in its fourth year. During this period the root rhizome system develops as observations of first year and juvenile plants Show. The root rhizome development is probably slow due to the light constraints of the forest floor. As the system develops, the likelihood of plant survivorship increases, and once adulthood is reached, survivorship is 80 percent. The root rhizome system helps the plant establish its place in the forest floor by aiding in competition for limited resources. In addition, if the plant is preyed upon, the root rhizome system is a good reserve source which can make up for the photosynthetic loss. Several of the marked S, £23313 plants had most of their above-ground biomass eaten, but were able to survive the following growing season. 117 The reproductive effort of this Species is also low relative to the other species studied. Most of the resources available go to vegetative functions. This supports the prediction of Gadgil and Schaffer (1975) that in environments where mortality of adults tends to be low, energy expenditure for survival of the adult should be high in comparison with energy expenditure for reproduction. There is some indication, as well, that reproduction may slightly reduce the sur- vivorship of adults since mortality was higher for reproductive clones than it was for nonreproductive ones. Rates of increase indicate that as long as germination levels and adult survivorship remain like those of 1978, S, £22313 can maintain its p0pulation within the dune forest. This is mostly due to the high levels of adult survivorship, since reproductive allocation is low and first reproduction does not begin before the fourth year. This species is among the most prevalent herbs of the forest floor as the importance values of Table A1 indicate. Life history characteristics of A, campestris reflect the strong abiotic element of the dune system. Of the three Species studied demographically, its seeds were the largest. This combined with muci- lagenous fruit probably accounts for a higher seedling survivorship than A. punctata's and levels equal to S, gaggjgfs. The A, campestris juvenile has the highest chances of survivorship of any life stage of this species. In this stage, the plant has a well developed root system and can withstand the xeric dune conditions. In addition, the resource drain of reproduction is not present. 118 It takes at least three years before an A, campestris individual will reproduce. For many plants this period is even longer. This is contrary to the characteristics of early successional species pre- dicted by the 'r' and 'K' model. The delay in reproduction likely , allows storage of nutrients needed for the seemingly suicidal repro- duction of this semelparous Species. When reproducing, 30 percent of the A, campestris plant's current year's net energy budget is devoted to reproductive structures. Additionally, tall (up to 100 cm) bolts serving both photosynthetic and reproductive functions also appear to be necessary. Thus, before the plant produces seeds, many of the vegetative Structures necessary to assure survival of the adult through reproduction must be already present. This is supported by the fact that most adults died in the winter after reproduction occurred rather than before it. Accumulation of the resources needed is likely slow for this species due to the low nutrient soil and xeric summer con- ditions of the foredune. The storage of a requisite amount of re- sources has been observed for other semelparous species: Dipsacus fullonem (Werner, 1975) for Species of Aggxg (Schaffer and Gadgil, 1975) and Scenecio jacobaea in dune systems (Van der Meijen and Van der Vaals-Kooi, 1979). It has also been discussed by Harper (1977). The examination of alternate life histories indicated that A, campestris's level of increase could be much higher if it were an obligate biennial. However, Hertz (1971a, b) and Van der Meijden and Van der Vaals-Kooi (1979) have suggested that prolonged age at first reproduction is advantageous in declining populations since extinction 119 can be avoided or postponed. It increases the probability that good conditions will be encountered again and that renewed growth can occur. This is particularly true if the environment is variable and the sur- vivorship of older nonreproductive age classes is high. In A, campestris rates of increase appear to fluctuate widely from one year to the next, and in some instances conditions can become so stressful that greatly lowered germination levels are so low that rates of in- crease are negative. Such years can be followed by favorable years as was the case in 1977 and 1978. The detrimental effects of an unfavorable year can be minimized if many individuals are in less vulnerable life stages (e.g. juvenile). The very low levels of seed production in 1978 kept the foredune Artemisia population from replacing itself as the examination of alter- native life histories indicated. The reason for the sharp decline in seed numbers between 1977 and 1978 is unknown; however, pollination difficulties due to low density may be a possibility since the foredune population was very small in 1978. The slack A, campestris papulation maintained a much higher density level and its seed production declined only Slightly in 1978. The life history of'A. punctata also reflects the effects of abiotic factors. Harsh conditions can delay time of first reproduction as the comparison of the transition probabilities for juveniles in 1977 and 1978 indicated. However, if conditions are favorable, an individual can reproduce in a period shorter than A, campestris. Less stressful abiotic factors in the Slack may contribute to A. punctata's ability 120 to flower sooner than A, campestris. However, A, campestris is present in the Slack and appears to delay reproduction there as well (personal observation).hl The morphology and vegetative phenology of A, punctata are perhaps the keys. Bolts are present in A, punctata both in repro- ductive and nonreproductive plants. Those of the reproductive plants are taller and more robust than those of nonreproducing individuals; however, they reach neither the size (31 cm versus 59 cm) nor the woodiness characteristic of A, campestris. There is also little dif- ference in mortality between juveniles and adults in A, punctata. Thus, ‘A. campestris may have a resource drain much like that hypothesized for Aggxg Species (Schaffer and Gadgil, 1975). In A, punctata, seed number is high and reaches levels similar to the foredune population of A, campestris. This, as well as higher adult survivorship, account for Monarda's higher reproductive rates during good years. However, such rates may be needed to balance unfavorable years, Since seedlings are highly vulnerable and seeds do not have mucilage which allows germination under low moisture levels as sug- gested by the drastic drop in germination levels as well as in rates of increase for 1977. Thus, this Species not only had the highest rate of increase for the three species in the favorable year, 1978, but the lowest rate of increase in the very dry year, 1977, as well. ‘A. punctata's lower allocation to reproduction likely contributes to higher adult survivorship than found in A, campestris. However, the hlSeveral A, campestris seedlings and juveniles from the slack area were marked and followed. 121 more highly stressful abiotic conditions likely keep this survivorship for Monarda well below that of S, ggggjg. ‘A. punctata allocates its vegetative resources very efficiently. When fruit is developing, much of the energy stored in leaves is trans- ferred and most leaves are lost. (Energy content drops from 9.55 k.cal./ g for green leaves to 3.93 k.cal./g in dead ones). This occurs during periods of greatest moisture stress (late July and August). By ridding itself of leaves, A, punctata individuals can reduce transporetion. levels. Some productivity is lost by this process, but this portion of the growing season is probably of lesser importance for growth Since processes which reduce water loss also inhibit photosynthesis. Stomata close and, thus, stop the flow of carbon dioxide as well as water (Kramer, 1969). 'A. punctata forms a basal rosette in the fall, when moisture conditions are again favorable for growth. Overwintering in this form, photosynthesis can begin immediately in the spring when conditions be- come favorable again. Thus, this species grows during the most favor- able periods of the growing season, and then transfers some of the accumulated energy to fruit development during the worst periods. A, campestris does not Show such an efficient vegetation phenology, and this may further contribute to its semelparity. This species loses some leaves during July and August, but most loss occurs during the fall when fruit is developing. Thus, the adult cannot take full advan- tage of the good growing conditions of this period. In addition, many leaves are present during conditions of greatest moisture stress, and tall flowering bolts place leaves into areas of higher wind speeds 122 because wind speeds increase with distance from the ground (Ranwell, 1972; Murphy, P. G., unpublished data). Thus, some adults die before fruit is fully developed. This is not as true for A. punctata (0.19 for A, campestris versus 0.08 for A, punctata). A, punctata, with its high rates of increase during favorable periods and its efficient vegetative phenology, would appear better adapted to the dune environment than A, campestris. However, as the demographic study has shown, the seed and seedling stages of this Slack Species are the most vulnerable portions of its life cycle. The harsher conditions found on the foredune combined with the smaller seed Size of this species likely explain why A. punctata is rarely found close to the lake. CHAPTER V SUMMARY 1. Findings from the reproductive effort studies support models based on 'r' and 'K' selection since reproductive effort declined with suc- cessional age. High allocation to reproduction probably helps balance the mortality levels for all age classes found in the early successional Species studied demographically. This would help to offset the periodi- cally very dry years like that of 1976-77 when the finite rates of in- crease for A, campestris and A. punctata can be less than one. Repro- ductive effort was also high for monocarpic Species. This was expected since these species have only one chance to reproduce and must devote proportionately greater resources to accomplish what the perennial has several years to do (Williams, 1966; Gadgil and Bossert, 1970; Hart, 1977) - 2. Species from the earliest successional stage tend to have a small number of large seeds. Xeric and low nutrient factors have probably played a role in selecting these characteristics. Seedlings must estab- lish in the face of scarce resources just like in competitive situations. In low nutrient sand, a large seed allows rapid development of a deep root system which confers resistance to the low water levels often occurring in the first few centimeters of sand. A large seed also permits the seedling to grow up through several centimeters of sand if the seed becomes buried. 123 129 3. AS the dune community matures, vegetative and topographic features ameliorate abiotic factors. Species of small seed Size and high seed numbers, such as A, punctata, A, 115232, and A. articulata, begin to appear due to the greater numbers of sites favorable to the survival of their seedlings. It is in the Slack area that Species come closest to having characteristics often associated with ”colonizing” species (Baker and Stebbins, 1965; Harper, Lovell, and Moore, 1970). However, xeric factors are still important since, in the comparison of seed size for Species common to both old fields and sand dune slack areas, the slack population had the highest average seed Size. Apparently, dryness plays the same role in selecting for large seed Size that competition does in more mature systems. 9. As more dense and diverse vegetation enters the dune community, competitive factors become more important. Thus, dune forest Species have a small number of large seeds. However, due to the diversity of vegetative structure and corresponding reduction in the constraints imposed by abiotic factors, a greater variety of seed characteristics is observed than is seen in the earlier stages. 5. Though selection appears strong for large seeds in the early stages of dune succession, there appear to be factors present which may con- strain seed size. This is evidenced by the pattern of Specific caloric content which is opposite to that normally predicted for succession. It is hypothesized that abiotic influences relative to imbibition rates and dispersal factors are balanced against energy requirements for seedling establishment. Large, slowly imbibing seeds may not successfully 125 germinate in the short favorable periods available or may not be able to reach suitable germination sites. Additionally, Since increases in seed size are often accompanied by decreased seed number (Harper, Lovell and Moore, 1970; Stebbins, 1971; Werner and Platt, I976; Werner, 1980), there may be a point where further reductions in numbers (and thus in potential offspring) may negate survivorship gains due to larger seeds (Werner, 1980). This is particularly true given the low survivorship for all life stages observed in the demographic studies of A, campestris and A. punctata. 6. Demographic studies further illustrate the influence of stress on the early stages of dune succession. Both A, campestris and A. punctata can exhibit delayed first reproduction if conditions are too stressful. In addition, A, campestris is a long lived monocarp, waiting at least until the third year before reproducing. Such delays may be necessary for this Species because of low soil nutrients and the large amount of resources this species devotes to structures associated with reproduc- tion. S, £35312 waits four or more years before reproducing. However, this is probably to establish vegetative structures necessary for the plant to compete effectively as well as to survive herbivory. 7. Of the three Species Studied demographically, S, 233312 is the least affected by changes in environmental factors. During the dry, l977 growing season, germination levels for this species dropped by only 30 percent and adults showed little Signs of moisture stress. Germination levels for A, punctata and A, campestris dropped by 90 and 70 percent, 126 respectively. As a result both Species, for 1977, had finite rates of increases which were less than one. 8. Survivorship levels for the three Species studied demographically reinforce findings of the reproductive effort and seed characteristics Studies. Seedling survivorship was lowest for the slack species 'A. punctata which has the smallest average seed size. Once plants of this slack Species develop beyond this vulnerable stage, survivorship levels are higher than those of A, campestris. A, punctata's lower allocation to reproduction likely contributes to higher adult survivor- ship than found in A, campestris. However, the more highly stressful abiotic conditions probably keep the survivorship well below that of S, ggggjg, This latter Species had the highest juvenile and adult survivorship; once the root-rhizome system is established, mortality declines suggesting this structure probably helps the plant survive competition and herbivory. Low reproductive effort in the adult may help keep survivorship high for this Stage since mortality of reproduc- tive adults was only slightly lower than those which did not reproduce. 9. Throughout this summary the influence of stress and abiotic factors has been emphasized; thus, it would seem this study advocates the adop- tion of Grime's model. However, it is felt that this model is too simplistic. While it does point out the importance of Stress factors, it leaves out such important influences as age Specific mortality, the availability of favorable microhabitats and their fluctuation through time, trophic positions, and environmental predictability (Charnov and Schaffer, 1973; Schaffer, 1979; Wilbur, Tinkle, and Collins, 1979; Platt, 127 1975; Grubb, 1977; Hart, 1977; Werner, l977; Vitt and Congdon, 1978; Michod, 1979; Werner, 1979; Whittaker and Goodman, 1979; Gross, 1980). It would seem a far better procedure to first isolate the factors of selection likely to be acting on a particular Species within a given system or stage of succession. Then, the observed set of life history characteristics should be explained in the context of these factors rather than factors deemed important in other systems (or situations). The process should not Stop here, however, experiments should be set up which test these explanations (or hypotheses) and provide more than correlative evidence that certain selective factors are important. Viewed within this context, this study of dune succession becomes only the first or inductive step toward an understanding of dune succession. LIST OF REFERENCES LIST OF REFERENCES Abrahamson, W. G. and Madgave Gadgil. 1973. Growth form and repro- ductive strategies in goldenrods (Solidago, Compositae). Amer. Nat. 107:651-661. Association of Official Seed Analysts. 1970. Rules for Testing Seeds. Vol. 60, no. 2. Association of Official Seed Analysts. 1970. Tetrazolium Testing Handbook. Contribution 29 to seed testing. Baker, H. 1972. Seed weight in relation to environmental conditions in California. Ecol. 53:997-1010. Baker, H. G. and G. L. Stebbins. 1965. The Genetics of Colonizing Species. The International Union of Biological Sciences Symposia on General Biology. New York: Academic Press. Bell, 0. J. 1979. Studies on the ecology of a streamside forest: Composition and distribution of vegetation beneath the tree canopy. Bull. Torrey Bot. Club 101:19-20. Box, M. J. 1965. A new method of constrained optimization and a comparison with other methods. Computer Journal 8:92-52. Bratton, S. P. 1979. Resource division in an understory herb community: Response to temporal and microtOpographic gradients. Amer. Nat. 110:679-693. Charnov, E. L. and W. M. Schaffer. 1973. Life history consequences of natural selection: Cole's result revisited. Amer. Nat. 107: 791-793. Cowles, H. C. 1899. The ecological relations of the vegetation on the sand dunes of Lake Michigan. Bot. Gaz. 27:95-117, 167-202, 361-391. Curtis, J. T. 1952. Outline for ecological life history studies of vascular plants. Ecology 33:550-558. Fretwell, S. 1969. The adjustment of birth rate to mortality in birds. Ibis 111:629-627. Gadgil, M. and W. H. Bossert. 1970. Life historical consequences of natural selection. Amer. Nat. 109:1-29. 128 129 Gaines et a1. 1979. Reproductive strategies and growth patterns in sunflowers (Helianthus). Amer. Nat. 108:889-899. Ganong, W. F. 1907. The organization of the ecological investigation of the physiological life histories of plants. Bot. Gaz. 93: 391-399. Glendening, G. E. 1991. Development of seedlings of Heteropogon contortus as related to soil moisture and competition. Bot. Gaz. 102:689-698. Good, N. E. 1978. Professor, Department of Botany and Plant Pathology. Michigan State University. Personal communication. Grime, J. P. 1977. Evidence for the existence of three primary Strategies in plants and its relevance to ecological and evolutionary theory. Amer. Nat. 111:1169-1199. Grime, J. P. 1979. Vegetation classification by reference to strategies. Nature 250:26-31. Gross, K. L. 1980. Colonization and local extinction of Verbascum thagsus (Mullein) in early successional fields. J. of Ecol. In press. Gross, K. L. 1978. Ph.D. candidate. Kellogg Biological Station. Michigan State University. Hickory Corners, Michigan. Unpublished data. Grubb, P. J. 1977. The maintenance of species-richness in plant communities: the importance of the regeneration niche. Bio. Rev. 52:107-115. Hansen, E. N. 1976. Primary production and metabolism of an Ammophila breviligulata (American Beachgrass) community. Ph.D. dissertation. Department of Botany and Plant Pathology, Michigan State University. Harper, J. L. 1977. Population Biology of Plants. New York: Academic Press. Harper, J. L. and R. A. Benton. 1966. The behavior of seed in soil, part 2. The germination of seeds on the surface of a water supplying substrate. J. Ecol. 59:151-166. Harper, J. L., P. H. Lovell, and K. G. Moore. 1970. The Shape and sizes of seeds. Annual Review of Ecology and Systematics 1:327-356. Harper, J. L. and J. Ogden. 1970. The reproductive Strategy of higher plants I: The concept of strategy with special reference to Senecio vulgaris L. J. of Ecol. 58:681-698. 130 Harper, J. L., J. T. Williams, and G. R. Sagar. 1965. The behavior of seeds in 5011. Part I: The heterogeneity of soil surfaces and its role in determining the establishment of plants from seeds. J. of Ecol. 53:273-286. Hart, R. 1977. Why are biennials so few. Amer. Nat. 111:792-799. Hickman, J. C. and L. F. Pitelka. 1975. Dry weight indicates energy allocation in ecological Strategy analysis of plants. Decologia 21:117-121. Hicks, M. L. 1938. Growth of Artemisia caudata in relation to environmental conditions. University of Chicago. Department of Botany. M. Sc. dissertation. Hollander, M. and D. A. Wolfe. 1973. Nonparametric Statistical Methods. New York: John Wiley 5 Sons. Hubbell, S. P. and P. A. Werner. 1979. On measuring the intrinsic rate of natural increase of population with heterogeneous life histories. Amer. Nat. ll3(2):285-292. lmshaug, H. 1978. Professor. Department of Botany and Plant Pathology. Michigan State University. Personal communication. Johnson, M. P. and S. A. Cook. 1968. Clutch size in buttercups. Amer. Nat. 102:905-911. Keever, C. 1950. Causes of succession on old fields of the Piedmont, North Carolina. Ecol. Mono. 20:229-250. Kramer, P. J. 1969. Plant and Soil Water Relationships: A Modern Synthesis. New York: McGraw-Hill Book Company. Laing, C. 1959. The ecological life history of AmOphila breviligulata on Lake Michigan sand dunes. Ph.D. dissertation, Department of Botany, University of Chicago. Laing, C. C. 1958. Studies in the ecology of Ammophila breviligblata l. Seedling survival and its relation to population increase and dispersal. Bot. Gaz. 119:208-216. Levin, D. A. 1979. The oil content of seeds: An ecological perspective. Amer. Nat. 108:193-206. 11 Linkola, K. 1935. Uber die Dauer ard Jahres klassenverhalthasse des Jungenstadium bei einigen Wiesenstauden. Acta Forst. Feen. 92:1-56. Ku, Kwen Pao. l977. Specialist. Department of Animal Husbandry. Michigan State University. 131 Leith, H. 1968. Measurement of calorific values of biological materials and the determination of ecological efficiency. In Functioning of Terrestrial Ecosystem at the Primary Production Level. Ed. by F. E. Ecardt. UNESCO Publication, Belgium. MacArthur, R. 1962. Some generalized theorems of natural selection. Proc. Nat. Aca. Sci. 98:1893-1897. MacArthur, R. and E. 0. Wilson. 1967. Theory of Island Biogeography. Princeton: Princeton University Press. McLeod, K. W. and P. G. Murphy. l977a. Establishment of Ptelea trifoliata on Lake Michigan Sand Dunes. Am. Midl. Nat. 97:350-362. McLeod, K. W. and P. G. Murphy. 1977b. Germination ecology of Ptelea trifoliata. Am. Midl. Nat. 97:363-372. McNaughton, S. J. 1975. 'r' and 'K' selection in Typh . Amer. Nat. 109:251-261. Menge, B. A. 1979. Effect of wave action and competition on broodling and reproductive effort in the seastar Lgptasterias hexactis. Ecol. 55:39-93. Mertz, D. B. 19713. Life history phenomenon in increasing and decreasing populations. In Statistical Ecology. Ed. by G. P. Patil, E. C. Pielou, and W. E. Watars. 2:361-399. Mertz, D. B. 1971b. The mathematical demography of the California condor population. Amer. Nat. 105:937-959. Michod, R. E. 1979. Evolution of life histories in response to age specific mortality rates. Amer. Nat. 113:531-550. Moss, E. H. 1938. Longevity of seeds and the establishment of seedlings in species of Populus. Bot. Gaz. 99:529-592. Murphy, F. G. 1978. Associate Professor. Department of Botany and Plant Pathology. Michigan State University. Personal communication. Newell, S. J. and E. J. Tramer. 1978. Reproductive strategies in herbaceous plant communities during succession. Ecol. 59:228-239. Odum, E. F. 1969. The strategy of ecosystem development. Science 169:262-270. Ogden, J. 1968. Studies on reproductive strategies with particular reference to selected composites. Ph.D. dissertation. University of Wales. 132 Ogden, J. 1979. The reproductive strategy of higher plants 11. The reproductive strategy of Tussilago farfara L. J. Ecol. 62:291-329. Olson, J. S. 1958. The rates of succession and soil changes on southern Lake Michigan sand dunes. Bot. Gaz. 119:125-170. Oomes, M. J. and W. T. Elberse. 1976. Germination of Six grassland herbs in microsites with different water contents. J. Ecol. 69:795-755. Pelton, J. F. 1953. Ecological life cycle of seed plants. Ecology 33:619-628. Pelton, J. F. 1951. Outline for ecological life history studies in trees, shrubs, and stem succulents. Ecology 32:339-393. Pemadasa, M. A. and P. H. Lovell. 1976. Effects of timing of life cycle on the vegetative growth of some dune annuals. J. Ecol. 69:213-222. Pemadasa, M. A. and P. H. Lovell. 1975. Factors controlling germination of some dune annuals. J. Ecol. 63:91-59. Pemadasa, M. A. and P. H. Lovell. 1979a. Factors controlling the flowering time of some dune annuals. J. Ecol. 62:697-657. Pemadasa, M. A. and P. H. Lovell. 1979b. Interference in populations of some dune annuals. J. Ecol. 62:855-68. . Pianka, E. R. 1972. 'r' and 'k' selection or b and d selection. Amer. Nat. 106:581-588. Pianka, E. R. 1970. On 'r' and 'k' selection. Amer. Nat. 100: 592-597. Pielou, E. C. 1977. Mathematical Ecology. New York: Wiley Interscience. Platt, W. J. 1975. The colonization and formation of equilibrium plant Species association on badger distributions in a tall grass prairie. Ecol. Monogr. 95:285-305. Rabotnov, T. A. 1975. The study of coenotic populations for the purpose of elucidating life history strategies of plant Species. Bull. Moscow Soc. Naturalists, Biol. Dept. 30:5-17. Ranwell, D. S. 1972. Ecology of Salt Marshes and Sand Dunes. London: Chapman and Hall. Salisbury, E. 1975. The survival value of modes of diSpersal. Proc. Royal Soc. of London, Sec. B 188:183-188. 133 Salisbury, E. 1952. Downs and Dunes: Their Plant Life and Its Environment. London: Bell 5 Sons, Ltd. Salisbury, E. 1992. The Reproductive Capacity of Plants. London: G. Bell 5 Sons, Ltd. Schaffer, W. M. 1979. Optimal reproductive effort in fluctuating environments. Amer. Nat. 108:783-790. Schaffer, W. M. and M. D. Gadgil. 1975. Selection for optimal life histories in plants. In Evolution of Communities. Ed. by M. L. Cody and J. M. Diamond. Cambridge: Bell Knap Press of Harvard University Press. Sheldon, J. C. 1979. The behavior of seeds in soil III. The influence of seed morphology and the behavior of seedlings in the establishment of plants from surface lying seeds. J. Ecol. 62:97-66. Shrive, F. 1917. The establishment of desert perennials. Jour. Ecol. 5:210-216. Stearns, S. C. 1976. Life history tactics. Quar. Rev. Bio. 51:3-97. Stebbins, G. L. 1971. Adaptive radiation of reproductive character- istics in angiOSpermS II: Seeds and seedlings. Ann. Rev. Ecol. Syst. 2:237-260. Stephenson, S. N. 1978. Associate Professor. Department of Botany and Plant Pathology. Michigan State University. Personal communication. Stevens, 0. A. 1932. The number and weight of seeds produced by weeds. Amer. J. Bot. 19:789-799. Stevens, 0. A. and L. Rock. 1952. Outline for ecological life history studies of herbaceous plants. Ecology 33:915-922. Struick, G. J. and J. T. Curtis. 1962. Herb distribution in an Acer sacchraum forest. Amer. Midl. Nat. 68:285-296. Tayla, O. R. 1979. Professor. Systematics and Ecology. University of Kansas. Lawrence, Kansas. Thompson, L. M. 1977. Seed Analyst. Seed Laboratory. Michigan Department of Agriculture. Personal Communication. 139 Van Asdall, W. and C. E. Olmstead. 1963. Corispermum hySSOpifolium on the Lake Michigan Dunes: Its community and physiological ecology. Bot. Gaz. 129:155-171. Van der Meijden, E. and R. E. van der Waals-Kooi. 1979. The p0pulation ecology of Senecio iacobaca in a sand dune system 1. Reproductive strategy and biennial habit. J. Ecol. 67:131-159. Vitt, L. J. and J. D. Congdon. 1978. Body shape, reproductive effort, and relative clutch mass in lizards: Resolution of a paradox. Amer. Nat. 112:595-608. Watkinson, A. R. 1978. The demography of a sand dune annual: Vulpia fasciculata II. The dynamics of seed populations. J. Ecol. 66:35-99. Watkinson, A. R. and J. L. Harper. 1978. The demography of a sand dune annual Vulpia fasciculata I. The natural regulation of populations. J. Ecol. 66:1-19. Went, F. W. 1955. The ecology of desert plants. Scien. Amer. 192:68-75. Werner, P. A. 1979. Competition and Coexistence of Similar Species. In Topics in Population Biology, edited by 0. T. Solbrig, S. Jain, G. B. Johnson and P. H. Raven. Columbia University Press, N. Y. Werner, P. A. 1977. Ecology of plant papulations in successional environments. Syst. Bot. 1:296-265. Werner, P. A. 1975. Predictions of fate from rosette Size in Teasel (Dipsacus fullonum L.). Oecologia 20:197-201. Werner, P. A. and W. J. Platt. 1976. Ecological relationships of co-occurring goldenrods (Solidago: Compositae). Amer. Nat. 110:959-971. Whittaker, R. H. and D. Goodman. 1979. Classifying species according to their demographic strategy 1. Population fluctuation and environmental heterogeneity. Amer. Nat. 113:185-200. Wilbur, H. M., D. W. Tinkle, and J. P. Collins. 1979. Environmental certainty, trOphic level, and resource availability in life history evolution. Amer. Nat. 108:805-817. Willson, M. F. 1971. Life history consequences of death rates. The Biol. 53:99-56. Yoda, K. T. Kira, H. Ogawa, and K. Hozumi. 1963. Self thinning in overcrowded pine stands under cultivation and natural conditions. J. Biol. 19:107-29. APPENDIX 135 Figure Al. Level 1 soil moisture data for the three study areas. Brackets enclose two standard error lengths. 9 water/ cc soil .008 - .007 - .006 - .005 - .004 - l .05 l .03 l .02 .003 - .002 - .001 Forest Foredune l i l 19 28 J 10 I36 J J J l 24 e 29 3 Date of Sample May May May June June July July Aug Aug Aug Sept Sept Sept 6 19 J 3 I P 15 23 137 Figure A2. Level II soil moisture data. Brackets enclose two standard error lengths. 9 water/ cc soil .008 - .006 - .004 - .003 - .002 - .1 - .08 '- .06 - .05 - .04 '- .03 - .02 - .01 - Forest Foredune 138 .00 l I I I I 1 I 1 1 r 1 1 1 1 May May May June June July July Aug Aug Aug Sept Sept Sept 6 I9 28 10248293111931523 Date of Sample 139 TABLE A1. Importance values for species from the three study sites. Chimaphila maculata Direa palustris Hepatica americana Mitchella repens 0ry20psis asperfolia Oryzopsis racemosa Osmorhiza claytoni Polygonatum pubescens Smilicina racemosa Solidago caesia Viola pgpilionacea Violai(Unknown) m =3 Importance Area Species value Foredune Herbaceous Ammophila breviligulata 289.5 Artemisia campestris 1.1 Calamovilfa longifolia 7.2 Solidago spathulata 7.1 Shrubs Prunus gumila 200.0 Open area 2 100.0 Ptelea trifoliata 0.0 Cornus stolonifera 0.0 Trees Populus tremuloideg: 3 0.0 Slack Herbaceous Ammophila breviligulata 22.9 Andropogon scoparius 3.7 Artemisia campeSIris 1.2 Asclepias syriaca 2.6 Hudsonia tomentosa 98.5 Ajthospermum carolinlense 3.8 Monarda punctata 12.3 Egpicum vergatum 29.2 Solidago Spathulata 11.6 Taxicadendron radicans 5.2 Cryptogams 75.7 Shrubs Hypericum kalmlgnum 229.6 Prunus pumila 70.9 Forest Herbaceous Aralia nudicaulis WO‘N d 0‘..- owooxwuoomwnuow mooo—m—oowwmwtom 190 TABLE A1 (cont'd). Fagus grandifolia Ostrya virginiana Prunus serotina Quercus borealis Quercus palustris Tilia americana Tsuga canadensis ImportanceI Area Species value Forest (cont'd) Herbaceous (cont'd) Cryptogams 98.1 Ferns 0.6 Shrubs Amelanchier . Sgonymus abovatus 1 . Hamamelis virginiana Rubus Sp. Smilax rotundifolig 3 . Taxicadendron radicans . Viburnum acerifolium 5 . Trees Acer saccharum ;r_.th300h3hJJr U1h3h3C>C>O\-‘ mmmumrwm mmurwmn NNU‘l-P—‘N IData were taken from projects performed for Botany 950 of Michigan State University. The Transect Method was used unless otherwise noted. 2This Species was present in the Study area, but was not sampled in the transects due to its low density and clumped distribution. 3Also present as a Shrub. “Data were collected using the Point-quarter Method. 191 TABLE A2. Degree-hour values for the three study areas. The period of measurement is 1030-1830 unless otherwise Stated. Degree (C°)-hours Date Foredune Slack Forest lQZZ July 25 67.3 58.1 50.0 August 23 72.2 69.5 55.1 September 28l 50.1 37.8 28.8 October 29 25.3 28.6 23.9 lgzg May 18 59.3 98.9 50.1 June 17 58.0 55.7 99.9 July 20 80.2 63.7 62.5 IMeasurement period was 1030-1630. 192 TABLE A3. Percent-hour data for the three study areas. The period of measurement was 1030-1830 unless otherwise noted. Percent-hours Date Foredune Slack Forest 1.22.7. I 2 June 27 65.6 ND 205.2 July 253 136.5 209.7 503.9 August 23 219.5 318.2 501.9 September 284 192.8 205.9 365.7 October 29 903.5 379.0 979.9 ‘SSZS May 18 279.1 288.9 291.0 June 17 518.1 551.9 557.9 July 20 237.5 967.1 5311.2 1Time period was 1130-1600. 2No data due to equipment failure. 3The period was 1030-1600. “The period was 1030-1630. 143 2_o.ov o.o Amo.ov o._ zoo.ov o.o zoo.ov o.o Amo.ov o.o : eooee>oz A_o.ov o.o Amo.ov o.o Amo.ov o._ Amo.oo o.o Amo.ov o.o mo eoosoeoom A_o.ov o.o Amo.ov o.o Amo.ov o.o “mo.ov o.o Amo.ov o.o no eooeoeoem 2_o.ov o.o Aoo.ov o.o Amo.ov ~.o zoo.ov e.— Amo.oo ~._ m eooeeeoom zoo.ov o.o Amo.ov o.o Amo.ov ... Amo.oo o.o Amo.ov o.o _m emooe< Amo.ov o.o Amo.ov o.o Amo.ov o.o 2_:.ov o.o zoo.oo e._ m. emoo=< zoo.ov o.o Aoo.ov o.o Amo.oo _.o zoo.ov o.o Aoo.ov o.o __ emeoeo mNmfl zoo.ov o.o Amo.ov m._ Aoo.ov o.o Am~.ov ~._ 2_o.ov m._ no eoooeoo zoo.ov o.o zoo.oo 9.. zoo.ov o.o zoo.oo o._ zoo.ov m._ ow eooeoeoom Amo.oo o.o Aoo.ov 4.. Amo.ov N.o zoo.oo ..N zoo.ov o._ m. emeoe< Amo.ov o.o Amo.ov o.o Amo.ov o.o Amo.oo o.o Amo.ov o.o mo >_=o NNmfl omoLOu xon_m xumo once; __meo>o ouoo ocaooLOm .ooN_mo;ucoLoo 02m mcoLco oeoocmum oou oco vacuum Loo mcouae oeo mu_c: .ocsoem as» 802% so on coxou ago: mucoeoLzmmoE _—< .mou_m >oaum ooezu 0:» com ucm_o: ace—m um ooaamoc_3 .:< w4ooz mu eooeouoom ma eooeueoom m coneouoom _m umooo< :— umama< omega; xom_m xoom 0:02; __meo>o oesoOLOm ouoo lull .ooN_m0;ucoema ocm meageo oemocoum ecu ocm ocooom Loo meoooe mom mu_co .oeooLm 0:» 502* So _ coxmu 0203 mucoEoeommoE __< .mhmz.4mmoLo >ooum omega any 20» ozone; mc__ooom um vooqmoc_3 .m< moooz - om .eeem zeoo.oo _o.o A~.oo ... Ao._v o.o 2_._V o._ oo .eoem 1 _o .oe< 2_oo.ov moo.o 2_.oo o.o 2:.ov o.o 2:.oo o.o _m .o=< 1 m .o=< 2_o.ov «o.o Auoo.ov o.o Ao.~v o.o 2o._o o._ m .o=< 1 o >_=o 38.8 mooo 3.8 o.o 8.8 N; 3.8 _._ o :2. .. R .6: A_o.ov «o.o Am.ov o.o 2m.~v o.o AA.~V o.o AN zez - mo __eo< MNNH. zeooo.oo _o.o Amo.oo ..o 2o.~o o.o 2_.oo o.o : .>oz - m. .euo “_o.oo _o.o Ao._v o.o 2~.~V o.o 2m.~v «.6 m. .eoo - __ .eeem zoo.oo mo.o 2m.ov o.o A~.ov A.o 2~.ov mo.o __ .eeoo - o. .o=< zoo.oo «o.o 2~.~V o.o Am.~V o.o 2o._o o.o oz .o=< 1 o. >_=o NNwH. umoLOu xom_m ucoLm __oeo>o oo_coq,oszk Loo team *0 mEmem ago mu_co .oo~_mozoc02mo ago mcoLeo oemocMHm ago ocm >mo can ocom mo eoooe_ucuo .mmoem >ozum 002;» any com m_o>o_ ucoEo>oe ocmm .m< m4om mm_oo_0m< ~:.~o~ -.o~ no.6 o.o.u No... _u.:m e__oe_eoo>_eo o.oeoeeom oo.om~ :~.o_~ oo:.o mom.oo o-.o_ «o.om e_eeeoee e__xeo o.o._ _o.o_ mom.: :Am.~_ omo.~ m~.o_ eee_=o___>eeo e__eeoee< o~.¢m o_.o_k o:~.: .~_.o mo_.~ _o.o e._oe_meo_ ee__>eee_eo Lasso: Conan: Eaem Loo ooom Loo Amev o;m_03 whammo oommwmo< ooom mo_eo_ou m0_eo_ou ooom _o>_uooo maoooom -o__z -o__z -oeooz .m0_ooom oesoaLOm Lam mu_um_20uomemnu ooom ocm “Lawwo o>_uosooLo02 mo memos ouELOmmcmeuc: .m< m4oo mao.>eeeum000e< .m.mmm m¢.o~oo :.~.m «mm.o mono.o am.m seem.e.mx 530.202»: :35 mod. .8... oo....~. SJN 3.2 61.21634 6:; mm~.o ~m.~o. Amm.m mo.mn. ow.m~ «o.o mc.uoeom mseoeo mm.mom m...m- mm..m eoo.m moo.o m~.om m.ee0.o meoeeoeoo ~:.m- om.:o.m oe~.m moo.m mm:.o mo.m mem.ozemew come..om :o.mm. mo.mmmm meo.m .mm._ mmm.. m:.m~ m.eumoo560 m.m_eoue< m~.mo. mn.m..~ omm.m NNo.. mam..o «o.o. momoucoo muemeoz no... mode. :26 mood. io.~ No... 85.... .28 3% «o.oou mm.wmm omm.m o.o.om m..m o..m. mue.e>m mm.qo.0m< om.mm. o~.ooo mom.: om:.~ m:m.o om.: mmoeeoeoe e.eomeoz n~.om~. mn.omm. _m4.m m~n.o on...o ~o.o~ mane». o.ome< :~.m~:. mm.mom o.~.: oo:.~ mom.o ~:.m~ 606.30.026 m..oeom>.oe oz no.~o. mmo.: mom.o. mom.~ oz ..mx m.0w.mm 3.3 2.:3 no; o.o.o .3; 82 3% o~..: em.mmm. muo.: o.o.u mum.. o~.m asymme.> 2:0.emo m_.m_ mm.:n: nmm.m _mm.m mom._ _o.m m__om~meo_ nm__>05m_ou Lasso: Lassa: Emem can. ooom Lon. mmev acmzoz ueommo ooom ovum mo_eo_mu m0_eo_ou ooom o>_ou:o mo_ooom eoemeoe< -o..z -o..z . -oeooo .mmmUGQm xUQ—m 1.0m. monummLMHUMLM£U 300m flew ”Thou—$0 0>~903UOLQOL “—0 WCMOE VELOWmCMLuC—J .m< w4m mm.w mm.m No:.m ~m_.nm :_.m_ oz msum>030 mae>c03u oz oz mm~.m :mo.mm., om.:~ oz me.uoeom moeoee :~.:_o~ oo.m~_~ ome.: mom.o m~_.o :~.om __ocOE:eo m.ome< wo.mm _m.mm_ mm.m :Nm._ om~.o #0.: o_mom0 omoo__0m no... :~.- mmo.¢ mom.N. «em.~ :m.m meoooe m..ozoo.z mm.:: om.mm m:o.: omo.o. o-.~ o~.o meaneeoe.o e:..nm oz .o.o:~: oz oz m.oo.o oz mum—some m._eome.eo mm.mn mm.:m mmo.m mm..: «No.0 mm.m omoumco..amm n.0.> m:.wm om.mm mmw.: Nom.m~ oo:.m mm.o~ _c0u2o_o m~_;LOEmo em.m. oo.m. omm.m o.o.m oo~.. mo.. m.me0umcmo m.m0..:u< om.:m om.w: amo.m mom.m omm.. mm.o. memo.eoEm no.0mo0: mm.m m:.- mmm.: mm..ww m~.o~ _w.__ mcoomonam.e:umc:o>_0o mm.oN mo.: omm.m em~.m: no.m. m~.m. omcoemcmo Eaeozuem.mz mm.o~ om.m~ wmh.m mom.o__ w~.mu om... amoeoooe me_o___Em mu.:m Nm.mm~ :d:.m _m:.~ «_n.o mn.~ mu:_o>coo xoemu m_.m :m.m. mum.m mom.o. .mm.m no.0 mmoeoom. m.moo~>eo _n.m no.mw _~_.: :No.w~ omn.m mo.m m__0meoomm m.mo0~>eo N:.~: _:._o_ _no.: mmo.: noo._ mm.m mmaono monumom Loose: Lasso: Emem Loo, beam can Away u;m_03 onmeo noon team mo_eo_mo mo_eo_oo ooom 0>_uo:o mo_uoom eeemeoee -o..z -o..z . -oeoee .mo_ooom amoeom cam mu_um_eouomeoco even one onmmo o>_ou:o02ooe mo memos ooELOmmcmeuc: .o_< moo