EFFECT OF fMMIGRATl‘ON THE GK THE ESTABLISHMENT OF WILD CARROT POPULARONS IN EARLY 0LD~F1ELD 'SUCCESSMN Thesis for the Degree of Ph. D. MICHIGAN STATE UNIVERSITY. . ' BUFORD 'R. HOLT A ‘ 1969. ‘ 4-_-,_ mm I 12 WWW" 756 5438 1 5‘7: rue-u? LIB R A R Y Michigan ‘5: U :3 mt y _ (fun This is to certifg that the thesis entitled E7‘*‘Ci‘tm ’7? T ” "T-“. ”17/" '7 rn" - ~ .--7 mvr -,-r1'1‘ -~-'- T .vv ~41'm ("'7‘ ' ; L V1, .au ,7, .. .41 J. 4- 4'4 ‘V i .4 14 a. ....L _.4J.-.) - 4.4 . , d1. T "1"" ~-\' ’1 ' hrs rfi -* f7"- ' m‘fifi ~k1 I 1 T; nT 17 T " 771' -T " rs" firs ' .jx (1 I 1 . .. ~ I- .Lu slant... \Jfi ._ 4-_,.4_s..1..-\..u_) .i 3.44.2. --'._ \J__:V/ 4....-44 .4.) D .4.,'..44... v . has been accepted towards fulfillment of the requirements for T" fir?" ‘zl‘rj‘ ‘ ri'ir'f'v I“. (_ ’ 71" . H”! ---v\’ ..--.., 1.. _)— CF .__...,_D_.i_1‘)u _ ... degree In BO .AL . 5 Qmfi, Major professor 0-169 . " Rel ABSTRACT EFFECT OF IMMIGRATION TIME ON THE ESTABLISHMENT OF WILD CARROT POPULATIONS IN EARLY OLD-FIELD SUCCESSION BY Buford R. Holt The effect of a small difference in immigration time on the life table characteristics and consequent establish- ment success of Daucus carota L. (Umbelliferae) was examined in early old-field succession in southern Michigan. Seedling emergence, survival to flowering, and seed production were monitored and correlations between these phenomena and the percent cover of the more abundant associated species were examined. The study area was a 100 x 100 meter section of a former corn field which had been divided into 25 16 x 16 meter plots separated by four meter buffer strips. Seeds were sown at three seed rates in December 1966 in replicated plots at the end of their first or third fallow year. Seeds were sown in 260 0.24m2 quadrats in each fallow. Of these 260 quadrats, :10 were sown with 188 seeds each (low rate), 10 with 814 seeds (high rate), and 240 with .412 seeds (intermediate rate). 1k) insure the presence of bare plots during the spring Buford R. Holt germination pulse, non—carrot vegetation on 50 of 240 quadrats sown at the intermediate rate was clipped prior to seeding and at intervals during the first growing season. The vege— tation composition was assessed in all quadrats by visual estimation of the percent cover of each species. -In December 1967 seed were sown in 30 quadrats in each of three sod types: Agropyron repens, Poa compregga, and bare ground. Individual emerging seedlings in these quadrats were marked and their survival monitored during their first growing season. Establishment of reproductive pOpulations was most af- fected by events preceding seedling emergence, and was nega- tively correlated with the abundance of other plant species. In unclipped quadrats sown with 412 seeds (1740 seed/m2) mean seedling densities in the first growing season repre— sented 58% of the seed sown in the younger fallow and 18% in the older fallow. Early post-emergence mortality was not determined in 1967, but in 1968 was equivalent to only 5 to 14% of the total seedling emergence in quadrats sown the preceding fall. In the quadrats sown in 1966, percent post— emergence mortality during the first two years of growth was relatively independent of both intraspecific density and the percent cover of other species. Juvenile mortality was highest during the winter and during summer droughts. No Daucus reproduction occurred in the older fallow in the first growing season, but in the younger fallow, reproductive Buford R. Holt plants represented 0.7% of the seed sown, and 5.5% of the live plants. Second year reproductives represented 0.2% and 5%tof the seed sown in the young and old fallows re- spectively, i.e., 7 and 57% of the individuals surviving to the second summer. The first year seed cr0p was largely destroyed by unknown herbivores, but second year seed pro- duction was equal to 89% of the seed sown in the older fallow, and 1500%'of the seed sown in the younger fallow. Both the number of flowers and the air dry weight of seed produced per quadrat were linearly related to the number of reproductive plants. The number of reproductives was pro— portional to carrot density irrespective of age of fallow. The probability of reproduction in the second growing season was negatively correlated with the more common grass species and figlilotus alba, the most abundant forb. Small delays in the arrival time of Daucus seed may significantly alter both its ecesis success and the rate of subsequent population buildup. The magnitude of the effect varies with the composition of the previously established vegetation. Since Daucus can persist for several decades in Michigan old-fields it may be representative of non- vegetatively propagated old-field species. If so, seedling establishment and age of initial reproduction are the life history phenomena most sensitive to the time of immigration and.mechanisms reSponsible for these effects warrant further study. EFFECT OF IMMIGRATION TIME ON THE ESTABLISHMENT OF WILD CARROT POPULATIONS IN EARLY OLD—FIELD SUCCESSION By \ o Buford R?VHolt A THESIS Submitted to Michigan State University in partial fulfillment of the requirements for the degree of DOCTOR OF PHILOSOPHY Department of Botany and Plant Pathology 1969 ééy//Ca,z._ )1— /2"70 ACKNOWLEDGEMENTS I would like to acknowledge the assistance of Dr. John E. Cantlon, my major professor, and Drs. William COOper, John Beaman, and Norman Good, my guidance committee. The staff at the Kellogg Biological Station facilitated this study. In particular I wish to thank Dr. George Lauff, Mr. Harold Webster, Mr. Paul Hartman, and Mr. Art Weist. This study was supported by National Science Foundation grants GB-1220 and GB-6941X. ii TABLE OF CONTENTS LIST OF TABLES . O O O O O O O O O O O O O 0 LIST OF FIGURES O O O O O O O O O O O O O 0 INTRODUCTION. . . . . . . . . . . . . . . . MATERIALS AND SITE. . . . . . . . . . . . . METHODS O O O O O O O O O O O O O O O O O 0 Effect of Immigration Time . . . . . . Partitioning of Seedling Establishment Statistics . . . . . . . . . . . . . . RESULTS Q C O O O C O O O C O O O O O O O 0 Effect of Immigration Time . . . . . . Failure of Seedling Establishment. . . DISCUSSION 0 C O C O O O O O O O O O O O O 0 CONCLUSIONS 0 O C O O C O O O O O O O O O 0 Conclusions Concerning Daucus. . . . . Conclusions General to Succession. . . LITERATURE CITED. . . . . . . . . . . . . . iii Failure Page iv 10 10 21 24 26 26 50 59 69 69 71 72 TABLE 1. 2. LIST OF TABLES Verification of the Mean Number of Seeds Weighed and Placed in Seed Packets. . . Number of Seeds Lodged in Seed Packets after SOWing O O O O O O O O I O O O O O O C 0 Correlation of Daucus Density with the Cover of the More Prominent Vascular Plant Species Occurring in the Quadrats . . . . . . . Correlations Between the Percent Cover of the Prominent Vascular Plants and the Probability of Flowering in the 2nd Growing Season. iv Page 11 11 38 45 LIST OF FIGURES FIGURE 1. 11. 12. 15. Approximate Distribution of Herbicide Treat— ments Applied in the Fall-1962 and Spring- 1963 as Part of Management Prior to Initia- tion of Old-Field Studies. . . . . . . . . . . Distribution of Blocks, Strips of Daucus Quadrats and the Small Indigenous Daucus BOpu— lation O O O O O O O O O O O O O O O O O O O 0 Number of Flowers/Inflorescence as a Function of Inflorescence Size. . . . . . . . . . . . . Regression of Loglo Air Dry Seed Yield per Plant on Loglo Stem Base Diameter. . . . . . . Time Courses of Net Seedling Emergence in Old and Young Fallows. . . . . . . . . . . . . . . Mean Number of Rosettes Resulting from Emi- grant seeds. 0 D 0 o o o o o o o o o o o o o 0 Time Courses of Rosette Survival in Old and Young Fallows. . . . . . . . . . . . . . . . . Relation Between Total Grass Cover and Rosette Survival . . . . . . . . . . . . . . . . . . . Effect of Sowing Rate on Rosette Survival. . . Effect of Clipping the Associated Vegetation on Daucus Survival . . . . . . . . . . . . . . Relation Between Total Grass Cover and Repro— duction Probability. . . . . . . . . . . . . . Regression of Estimated Flower Production on the Number of Reproductives/0.24m2 . . . . . . Regression of Seed Production in Grams Air Dry Weight on the Number of Reproductives. . . . . Page 17 19 28 50 55 35 57 4O 45 47 49 LIST OF FIGURES--continued FIGURE 14. 15. 16. 17. Frequency Distribution of Seed Production/ Biennial Plant. . . . . . . . . . . . . . . Histograms of Mean Seedling Emergence and Concurrent Mortality. . . . . . . . . . . . .Mean Number of Live Daucus Plants Resulting from Seed Sown in December 1967 in Four Sod Types . . . . . . . . . . . . . . . . . . . Summary of Percent Seedling Emergence, Sur— vival and Reproduction in Quadrats Sown at the Rate of 1740 Seeds/m2 in Two Ages of Fallow. . . . . . . . . . . . . . . . . . . vi Page 52 54 56 65 INTRODUCTION Plant species characteristic of the initial stages of old-field succession undergo rapid and quasi—predictable changes in population size. These changes, which are fre— quently associated with the appearance or disappearance of other species, have been extensively described in cross- sectional studies (e.g., Oosting, 1942; Bard, 1952; Beckwith, 1954). but experimental examination is essentially lacking. The effects of fertilization (Davidson, 1962, 1967), size of area of denudation (Davis, 1968), and faunal impoverishment (Malone, 1969) on vegetation composition have been investi— gated in the field, and possible causes of dominance shifts have been examined in the laboratory (Keever, 1950; Rice, 1964, 1967). The role of changing life table characteristics of constituent species is, however, unknown. The objective of this study was to evaluate, in an old- field succession, the effect of a small difference in immi— gration time on the life table characteristics and consequent establishment success of a plant population. Specifically, the objectives were to determine the effect of arrival time on seedling emergence, survival to flowering, and seed pro- duction; and to determine whether these effects are correlated with the amounts of the more abundant species of the previously established vegetation. MATERIALS AND SITE Daucus carota L. (Umbelliferae), the wild carrot or Anne's Lace, was selected for its ease of manipulation, and virtual absence from the study area, a site on which it would normally occur. This near absence apparently was due to the limited immediate seed source. Individual plants are easily recognized, short—lived, and easily censused. Flowers and seeds are large enough to be counted easily. In southern Michigan, naturally dispersed Daucus seed germi— nate primarily in the early spring, although very slight additional germination occurs throughout the remaining grow- ing season. .Surviving seedlings form rosettes which either bolt and flower during the plant's first growing season, or overwinter and flower in subsequent years. The plants die as the seeds mature and the seeds are shed in the fall and winter. The study area was a 100 x 100 meter section of a former corn field set aside for studies on early old-field succession. The field is located on Gull Lake Biological Station property at the intersection of Gull Lake Drive and B Avenue, Ross Township (T. 1 S, R. 9 W.), Kalamazoo County, Michigan. The soil is well drained Fox Sandy Loam (Typic Hapludalf) on flat to gently rolling glacial drift of Cary Age. The site had been farmed for approximately a century, and in the recent past has had a varied crop history. It was planted with hybrid walnut trees about 1958 (Harold Webster, personal communication; USDA photo BDB-5-50), but returned to general farming between 1950 and 1955 (USDA photo BDB-1G-95 and BDW-1P-47). Between 1960 and 1964 the site was planted with wheat, alfalfa, and corn (§§§_m§x§). Fertilizer (250 pounds/acre 6-24-24; 100 pounds actual nitro- gen) was last applied in 1964 with the final corn crop. Herbicides were last applied in November, 1962 and May, 1965 for a demonstration of quackgrass.(Agropyron repens) control (Figure 1). Cantlon et al. (unpublished) initiated basic studies of old-field succession in the study area in the fall of 1964. They divided the area into 25 16 x 16 meter plots separated by four meter buffer strips, and grouped the plots in five blocks. Blocks III, IV, and V were established primarily with respect to the uneven topography, but blocks I and II, both on level ground, were partitioned to minimize the anticipated number of missing plots per block. Although these plots (I-68, II-68) were subsequently retained, blocks I and II were irregularly partitioned. Each year since 1964 one plot from each block has been selected at random, plowed and left fallow (Figure 2). Davis (1968) and Cantlon et al. (unpublished data) have documented FIGURE 1 Approximate Distribution of Herbicide Treatments Applied in the Fall-1962 and Spring-1965 as Part of Management Prior to Initiation of Old-Field Studies. The borders of the 25 20 x 20 meter plots of Cantlon et al. (unpublished) are represented by the background grid. The remaining walnut trees are indicated by circles. The location of strips of contiguous carrot quadrats are indicated by checkered bars for plots plowed in 1966 and by solid bars for plots plowed in 1964. The widths of the checkered and solid bars are not to scale. Br Burned during the last half of May 1965. L = Lorox (Linuron) [5-(5,4-dichlorophenyl)- 1—methoxy-1-methylurea] was applied at 1.1/2 pounds/acre. Atrazine (2-chloro-4-ethylamino-6- isoPropylamino-1,5,5-triazine) was applied at the rate of 4 pounds/acre; 2 pounds/acre on May 6 and 2 pounds/acre on May 20, 1965. AT]. AM = Amitrol-T (5-amino-1,2,4—triazole plus ammonium thiocynate) was applied at 2 pounds/acre on May 6, 1965. AT2 = Atrazine was applied at 2 pounds/acre in the spring of 1965. DAL 2,4-D = Dalapon (2,2-dicholopropionic acid) was applied at 12 pounds/acre on November 7, 1962. 2,4—D (2,4—dichlorophenoxyacetic acid) was applied at 1 1/2 pounds/acre on May 20, 1965. AT2 2,4—D = Atrazine was applied at 2 pounds/acre in the spring of 1965. 2,4-D was ap- plied at 1 1/2 pounds/acre on May 20, 1965. AT3 2,4-D = Atrazine was applied at 4 pounds/acre on November 7, 1962. 2,4-D was applied at 1 1/2 pounds/acre on May 20, 1965. HERBICIDE TREATMENTS 'DAL'fiAT3 24D 2 4D l I L AT] AM AT2 / l.l.l/ / IIII/ 12w. /%%%VA2 BR' FIGURE 2 Distribution of Blocks, Strips of Daucus Quadrats, and the Small Indigenous Daucus Population. The background grid is the same as in Figure 1. The numbered squares repre- sent the portions of the plots committed to the previously initiated study by Cantlon et al. Widths of the strips representing quadrats are exaggerated (X 2). The Roman numerals indicate the block number, and arabic numerals indicate the year of fallowing. The topography character- istic of each block is: Block I Flat upland. Block II Flat to very gently sloping upland. ,Block III Flat to gently sloping. Block IV Gently sloping upland. Block V Flat to gently SIOping lowland. -oZ SCALE HO M ;\\\\\' INDIGENOUS DAUCU§ aoooo‘ BLOCK BORDERS COLONY Q TREES — W QUADRATS FIGURE 2 vegetational change since 1964 in the 100 x 100 meter study area and adjacent parts of the former corn field. Daucus was absent from the field except for a small colony at the edge of the study area (Figure 2). Vascular plant nonmenclature follows Gleason and Cronquist (1965). Voucher specimens, collected in coopera- tion with Mr. Roger Davis (1968), have been deposited in the Beal-Darlington Herbarium, Michigan State University. METHODS Effect of Immigration Time Seeds were collected October 8, 1966 from a naturally occurring population 5.5 km from the study area. Seeds were taken from first and second order umbels (nomenclature of Borthwick, 1951), and stored in ventilated containers at room temperature (25-2800). Immigration time was controlled by sowing seeds in fields of different ages. Seeds were sown at three rates in a replicated randomized block design on December 16 and 26, 1966 in plots last plowed or cultivated June 27, 1966 and early summer 1964. The 260 0.24 square meter (47.5 x 50 cm) quadrats used in each age of fallow were located in strips of 26, placed along the two more homogeneous borders of each plot, i.e., parallel to the plow furrows (Figure 2). Within. each strip, 24 quadrats were sown with approximately 412 seeds, one with approximately 816 seeds, and one with approximately 188 seeds, i.e., the rates were approximately 1740, 5440, and 790 seeds per square meter (Tables 1 and 2). Seeds were broadcast by hand within each quadrat to simulate natural dispersal. Seed numbers were estimated by weight. 10 11 TABLE 1 Verification of the Mean Number of Seeds Weighed and Placed in Seed Packets Planting Rate Number of Seed Packets Counted Mean Number of Seeds per Packet (:1: 3.15:.) High 6 818.: 6.0 Intermediate 5 414 i.7.8 Low 7 191 i.5.5 TABLE 2 Number of Seeds Lodged in Seed Packets After Sowing Planting Rate Number of Seed Packets Counted Mean Number of Seeds per Packet (45.3.) High Intermediate Low 19 118 55 1.8.i 0.5 2.0.: 0.5 2.5 i.0.5 12 Foliage and litter were removed from three quadrats per strip prior to sowing, to insure the presence of cover— less quadrats during the initial germination pulse. Continued removal of all cover was not feasible, but the non-carrot vegetation in these quadrats was partially re- moved by clipping on May 2, June 15-17, and August 10-11, 1967. Vegetation composition was assessed in each quadrat by visual estimation of the percent cover of each species. Initial cover readings and carrot censuses were taken con- currently between April 11 and May 18, 1967. Subsequent cover readings were taken July 10 to August 10, 1967, and July 6 to July 21, 1968. Minimal estimates of seed emigration, after sowing, were based on rosette counts outside the sown quadrats in the second growing season (July 5-6, 1968). Counts were made in 0.24 m2 quadrats, laid end to end to form transects. These transects were perpendicular to, and extended one meter upslope and two meters downsloPe from, three randomly selected quadrats in each strip. Transects were approxi— mately parallel to the direction of the gentle slope of the field. Conversion of rosette counts to seed emigration estimates assumed: 1) 100% germination; 2) survival prob- ability equal to that observed for the same date in the permanent quadrats (P=0.56 in both fallows); 5) independence of germination percentages and seed densities; 4) absence 15 of further seed immigration into the field. Data of Cantlon et al. (unpublished) for changes in Daucus cover values in 150 meter square quadrats per age of fallow justify the latter assumption. .Estimates of potential field germination in the first and second growing seasons were based on laboratory germi- nation tests. In December 1966 fifteen lots of 100 seeds each were placed on moist blotters in petri dishes, then placed in -50C (six dishes) or in 40C (nine dishes) for 50-58 days. They were then removed to room temperature (25-280C) and germination was monitored for 20 days. In March 1968 two ZOO-seed lots of these same seed were placed on moist blotters at room temperature for 28 days. Cold treatment was not applied since previous germination tests had shown that seeds were unaffected by cold treatment fol- lowing prolonged storage. Field estimates of seedling emergence and rosette sur- vival were based on inventories during the first two grow- ing seasons. Carrots were censused in all quadrats April 11 to May 18, 1967; September 5 to 11, 1967; March 27 to April 15, 1968; June 26 to July 5, 1968; and September 25 to 28, 1968. Multiple counts were made during March and April 1968 since new leaf production during the first two weeks of April caused apparent increases in the number of overwintering rosettes. Counts were continued until a constant estimate was obtained. A randomly selected 14 subsample of 41 quadrats in each age of fallow was also inventoried May 25 to 27, 1967; July 25, 1967; October 24, 1967; April 10, 1968; August 21 and 22, 1968; and April 26, 1969. On April 28 and May 25, 1968 new seedlings in the subsample quadrats were marked with plastic toothpicks, permitting estimation of minimal second year germination percentages. Estimates of potential seed production in 1967 were based on empirically derived relationships between the number of flowers per inflorescence and the diameters of the discoid inflorescences. Estimates for inflorescences greater than 5 cm in diameter were based on a regression of flower numbers on 10910 inflorescence diameter (N=19, r= 0.95). Estimates for smaller inflorescences were based on the assumption that the flowers and inflorescences are circular and that the number of flowers/umbel can be approximated by the number of circles of arbitrary size which can be packed in a larger circle of given diameter. The diameter of the small circles (5.5 cm) in this model was chosen to fit both the observed values for small in- florescences and the regression value for inflorescences 5 cm in diameter (Figure 5). Due to the frequent occur- rence of ellipsoidal umbels, two measurements were made per inflorescence (major and minor axes), and their mean was used as the estimate of umbel diameter. Each umbel was measured when the central flowers were in anthesis. 15 Diameters of severely damaged (chewed) umbels were esti- mated by the distance between tips of umbellet peduncles, or if totally destroyed, were assumed to equal the mean size of measurable umbels in the same quadrat. Each umbel measured was labeled with a paper tag. Conversion of flower numbers to potential seed numbers assumed that all flowers possessed two viable ovules. Since most flowers in the first growing season were borne in primary umbels, this assumption probably caused minor overestimation of potential seed pro- duction (cf. Quagliotti, 1967). All umbels were measured in 1967; none in 1968. Actual seed production per quadrat was estimated by air dry weight of cleaned seed. The seed were hand threshed, sieved, and cleaned in a South Dakota model A seed blower. In 1968, data for seed yield/quadrat were inadvertantly lost for 45 quadrats. .These losses were replaced by estimates constructed from the sums of yield estimates for individual plants. These estimates of seed yield per individual were based on a regression (r=0.89, F1,57 = 5.16) of 10910 seed weight on 10910 stem base diameter (Figure 4). Since the slope of this log-log regression line was greater than one, overestimates of seed yield tended to greatly exceed the underestimates. To correct this bias, estimates of seed yield per quadrat were adjusted by use of the regression loglo corrected seed weight = -0.164 + 0.801 loglo estimated seed weight (F1 60 = 105.6). The correction was quite large 16 FIGURE 5 Number of Flowers/Inflorescence as a Function of Inflor- escence Size. The solid line is based on a regression of flower numbers on 10910 inflorescence diameter (y = -2700 + 4740K). The broken line is based on a model of tightly packed circles (see text). 17 3000 P “D O O O l NO. FLOWERS / INFLORESOENCE 6 O O I L" /1 1 1 1 1 J 1 1 L 1 2 4 6 8 IO 20 INFLORESCENCE DIAMETER — CM. FIGURE 5 18 FIGURE 4 Regression of Loglo Air Dry Seed Yield per Plant on Loglo Stem Base Diameter. The mean air dry seed weight was approximately 0.79 i.0.09 mg/seed. 5000 I000 I00 SEED PRODUCTION / PLANT - no. I TITIIII 19 woo]O Y=0.7I +3.33 Loomx IO 1 1 1 1 1 1 l1 1 4 2 4 6 IO STEM BASE DIAMETER -m1. FIGURE 4 20 (i.e., mean actual seed weight equaled 22 and 75%;of the mean estimated weight in the old and young fallows re— spectively) but it was needed in less than 10% of the quad— rats. The term stem base, as used in the estimation of individual seed yield, refers to the easily identified ring of leaf scars at the root-stem junction. In exceptional cases where this area was grossly enlarged relative to the adjacent portions of the stem-root axis, measurements were taken just below the leaf scar zone. All seed were harvested in 1967 and 1968. In 1967 seed bearing umbels were harvested individually over a period of eight weeks, as their outermost seeds turned brown. In 1968, the umbels within each quadrat were harvested simultaneously. All quadrats were harvested between September 19 and 22, 1968. This harvest was timed to permit as many seed as possible to be harvested at the same maturation stage as seed harvested in 1967, without permitting appreciable seed loss from early maturing umbels. In both years seed pro- duced by plants outside the quadrats were harvested and discarded. Establishment success of the second generation was assessed by sowing part of the 1967 seed crop in quadrats near the parent plant, i.e., in areas where seed presumably would have fallen. .Eighteen 0.06 m2 (25 x 25 cm) quadrats were placed in vegetation subjectively assessed as resembling that in which the parents had grown. To determine whether 21 marked differences existed in the emergence of seedlings from the seeds of annuals and biennials, six of these quadrats were sown with seeds from putative biennials. Seeds were harvested October 12, 1967 from known annuals in the study area, and from putative biennials in the population used as the 1966 seed source. All seeds were sown December 15, 1967 at the rate of 50 seeds per quadrat (800 seeds/m2). .Emerging seedlings were marked.with plastic toothpicks until May 16, 1968, the end of the initial germination pulse. Subsequent findings, however, suggest that very small numbers of new seedlings continued to appear throughout the summer. Partitioning of Seedling,Establishmentfifailure The role of post-emergence mortality as a source of dif- ferential seedling establishment was assessed in the field by labeling individual seedlings and monitoring their survival. Seeds collected October 12, 1967, from putative biennials, were sown in 90 0.06 square meter quadrats, 30 of which were placed in each of three vegetation types: bare ground, Agropyron repens sod, and Poa compressa sod. These quadrats were placed in areas of homogeneous sod near the quadrat strips used in the main experiment. Emerging seedlings were marked with plastic toothpicks and censused at least weekly during the spring germination pulse. Censuses were infre- quent during the summer. 22 Estimates of potential germination were based on labora— tory tests conducted immediately after seed harvest (October 12, 1967). Ten lots of 25 seeds each were placed on moist blotters in the cold (-1 to 1.5°c) for 16 days. Observation stopped two weeks after seeds were returned to room tempera- ture. In the greenhouse, the assessment of post-emergence mortality was combined with an attempt to separate the ef- fects of live grass and litter on seedling establishment. Seven sod and litter combinations were used: 1) bare soil; 2) Agrogyron repens sod plus its litter; 3) A, repens sod alone; 4) A, repens litter alone; 5) Poa compre§§§_sod plus its litter; 6) g, compressa sod alone; 7) g, compressa litter alone. Sod, soil, and fresh litter (i.e., standing dead) were collected October 51, 1967 within the 4 hectare (10 acre) field containing the study area. Litter was air dried and.stored at room temperature. Sod and soil were stored le an unheated greenhouse prior to potting on November 26 anud 27, 1967. Twenty-four five-inch (15 cm) porous clay EKTts were prepared of each of the seven sod-litter combina- tixans. Old litter was removed from all sod pots and was replaced where apprOpriate with sufficient fresh litter to ccD‘Ver the soil surface (1 to 4.8 gm)° The grass was clipped 'UD a.height of 5 cm, then half of the pots were placed in a cold frame on December 6 to simulate field conditions, While the other half were retained at 170C. Pots were 25 randomized in the cold frame and their bases were packed in peat moss. Seeds collected October 16, 1967 were sown at the rate of 20 seeds per pot. Pots were then covered with 1/4 inch mesh hardware cloth. ~Emergence, age specific mortality, and early develop- ment of Daucus seedlings were monitored in the pots retained in the greenhouse. To insure high potential germination in the absence of Cold treatment, these pots were sown with thoroughly after-ripened seed gathered in 1966. Twenty seeds were sown per pot (approximately 1500 seeds/m2) on December 8, 1967, and pot positions were randomized. Emerging seedlings were labeled with colored toothpicks, a different color being used at each census to permit evalua— tion of age specific survival and development rates. Esti- mates of potential germination were based on two lots of 25 seeds in petri dishes placed in the greenhouse December 19. Estimates of development rates were based on the rate of appearance of new leaves during the first 18 days follow- ing initiation of seedling emergence. At the end of the experiment, the sod, litter, and carrots in each pot were harvested. Sod and litter were air dried, then weighed on an O'Haus triple beam balance to 1.0.1 gm. Carrots were oven dried at 80°C for a minimum of one day and then weighed to.i 1 mg. Seedling emergence and post-emergence survival were monitored in the cold-treated pots after they were returned 24 to the greenhouse. Pots were moved into the 17°C green- house February 18, 1968. The bench was partitioned into two blocks, a central and a peripheral block, and pots were randomized within the blocks. A buffer strip of bare, soil—filled pots was placed around the peripheral block. Pots were watered at least weekly with 200-500 ml of dis- tilled water during the period of maximum emergence and with tap water thereafter. Emerging seedlings were marked with toothpicks. Temperature control was faulty throughout the period of study, but the variations were not serious until the latter part of the study (late March) when the temperature ranged from 11 to 50°C. Statistics Except for cover data, all percentages were trans— formed to angles (Snedecor, 1956). Cover data were not transformed since cover values, given an adequate sample, tend to approach a normal distribution (Holscher et al., 1958; Ursic and McClurkin, 1958). To meet the assumption of linearity, correlations of carrot performance with the cover of other species were based only on quadrats in which the sum of cover values of dominant species, other than the one being examined, was less than 20%. This value was arbitrarily chosen as a compromise between avoidance of small sample sizes and 25 the presence of significant quantities of species other than the two being correlated. Graphical analysis of selected sets of data indicate that this procedure restored linearity. Only quadrats sown with 412 seeds/Quadrat were included. Correlations of cover values and seedling emergence in 1967 were based on quadrats sampled between April 28 and May 14, a period in which net changes in Daucus abundance were absent. Correlations of litter, bare ground, and total cover were based on all of the 177 quad- rats sampled during this period. Tests of significance were based on standard proce- dures (t-test, ANOVA, Duncan's Multiple Range) given by Snedecor (1956) and Kirk (1968). Means cited in the text are accompanied by their standard errors. RESULTS Effect of Immigration Time In the first growing season seedling densities at the end of the spring germination pulse (Figure 5) were greater in the younger than in the older fallow (t148 = 9.46, P < .001), despite greater seed emigration in the younger fallow (Figure 6). Potential (laboratory) germination was 65%, but seedling densities in the two ages of fallow, after correction for emigration, represented only 42 and 18% of the seed sown in the young and old fallows respec— tively. Seedling densities were positively correlated with percent bare ground (F1,175 = 128.4, P < .001), and nega— tively correlated with each of the dominant perennial herbs (Table 5). Correlations with ghus typhina and Melilotus alb§_were not calculated since their April-May cover values were negligible. Estimated seed emigration (Figure 6) was significantly greater in the younger than in the older fallow (t47 = 6.28, P < .01). Loglo estimated seed loss was negatively correlated with litter (r = 0.81, F1,50 = 98.15, P < .01), and live cover (r = 0.49, F1,50 = 15.94, P < .01), implying that seed losses declined exponentially with increased cover . 26 27 FIGURE 5 Time Courses of Net Seedling Emergence in Old and Young Fallows. Points represent mean (1.2 S.E.) carrot densities in fields fallowed in 1964 (O) and 1966 (0). Seeds were sown in December 1966 at the rate of 412 seeds/0.24m2. PLANTS / 0.24112 l00 5 28 . YOUNG FALLOW o OLD FALLOW 01' l I I l J IO 20 30 40 50 60 DAYS AFTER I APRIL |967 FIGURE 5 29 FIGURE.6 Mean Number of Rosettes Resulting from Emigrant Seeds. Data are based on rosettes occurring outside sown plots during the second growing season. Data for the older fallow (O) are means of 22 quadrats and data for the younger fallow (0) are means of 27 quadrats ( ) or 25 quadrats (----). Two quadrats in the vicinity of the highly localized indigenous carrot population are excluded from the means connected by the broken line (---—). Only unclipped quadrats sown at the rate of 412 seeds/0.24m2 are included. NO. OF CARROTS / 0.24 M2 .0 10 'o 50 UPSLOPE DOWNSLOPE 4—— a l I l l l I I l J I l I 0.5 0 0.5 I 1.5 DISTANCE FROM SEED SOURCE-M FIGURE 6 51 Seedling emergence in the second growing season was negligible in both ages of fallow. -Estimated percent emerg- ence accounted for less than 1%Lof the seed sown in 1966,~ even though seed not accounted for by prior germination or emigration represented 42 and 80%»of the original seed in- put in the young and old fallows respectively, if one makes the unwarranted assumption that no mortality occurred prior to seedling emergence. Potential (laboratory) second season germination of the same seed was 81%. Seedlings were observed but not counted in the third growing season. The seedling densities appeared to be low, although considerably higher than in the previous year. (Mortality rates during the first two growing seasons were quite variable, but percent mortality was relatively independent of the age of fallow (Figure 7), the cover of the dominant herbaceous species (Figure 8), and intra— specific density (Figure 9). The correlations between_ Daucus density and the cover of particular dominant species tended to increase during the first two years (Table 5). Partial removal of the other Species, by clipping, had little effect on the survival of Daucus rosettes, but did enhance seedling establishment in the older fallow (Figure 10). Intraspecific density accounted for very little vari- ance although graphs of percent survival as a function of initial density suggested that the maximum survival prob- abilities were pr0portional to the reciprocal of intra- specific density. 52 FIGURE 7 Time Courses of Rosette Survival in Old and Young Fallows. Points represent means of 41 quadrats (:l: 2 S.E.) in each age of fallow (O = Old, 0 = Young). The same 41 quadrats per age of fallow were used for each inventory. |00 PLANTS / 0.24 M2 6 55 O—OYOUNG FALLOW O—OOLD FALLOW l l l 1 l l 200 400 600 800 DAYS AFTER I APRIL l967 14,1,1 1 1 1 1 1 1 11 l 1 1 11 1,1 1 1 1 1 1 11 1 A J O J A J O J A 1967 I968 1969 FIGURE 7 54 FIGURE 8 Relation Between Total Grass Cover and Rosette Survival. Data for the two ages of fallow are pooled. Quadrats are the same as in Figure 7. 55 fl 9., cuss covn b—i o-2o,u=27 o—-—o > 20-40, 11:23 ~—~ >40-60.N=18 o—a use-so. N = a >80-IOO, N I 4 100 T T rTlTIT I I 9 1 I IIIIFFT§ / / / TO ZERO PLANTS / 0.24 M2 I IUIIIII I l l l 400 600 800 DAYS AFTER 1 APRIL l967 200 L1 1 1 1 14 4 L1 1 4 1 14 11 1 1 14 l 1 1 1 14 A J o 'J A J o 'J A 1967 I968 1969 FIGURE 8 56 FIGURE 9 Effect of Sowing Rate on Rosette Survival. Seed rates were 188 (low), 412 (intermediate), and 816 (high) seeds/0.24m2 quadrats. Data for the intermediate seed rates are from Figure 7. N=10 for high and low seed rates in the young fallow and N=9* for high and low seed rates in the old fallow. Solid symbols (0. ‘1 I) represent values for the young fallow and hollow symbols (0, All) represent values for the older fallow. * The discrepancy in sample sizes in the two ages of fallow is a result of sowing two of the clipped quadrats at the wrong rates. I00 PLANTS / 0.24 M2 5 57 ' \ —-3IucH — g INTERMEDIATE 3 LOW l 1 1 I I 400 600 I 200 DAYS AFTER I APRIL I967 FIGURE 9 58 AH m.comnmmmv ucmwoflmmmoo sowumHmHHou Roe. v muAHAnmnoum ao. v muAHAnmnoAm mumuflmnq mo Hmnfisz *** ** oAumA A II II HEII II Z ***ao.mmm mam.OI oma *Imo.e Rom.OI mma IIImm.mH 6mm.OI am mammmu consaouwa IIIaa.ma .mma.OI as mm.m mem.OI mm «0.4 omm.OI mm mmmmuasoo «mm. IIImm.om 44A.0I Hm **Imm.am mmm.OI as *IIom.mm HRP.OI mm mamcmumua mom IIIHN.om mma.oI mm *Iam.ma amm.0I on I I I mnHm.mmmmAMAmm mm.m ¢NA.OI ma m>.o amH.OI an I I I mcflaazw warm m u z m u z m H z mmma qqam Pmma gnaw Rama oszmm mB¢MQ BZ¢ZHZOMQ Mao: W38 mBH3_MBHmZmQ mDUDfiQ m0 mZOHfififlmmMOU m mflmfifi 59 FIGURE 10 Effect of Clipping the Associated Vegetation on Daucus Survival. Unclipped data (----) are from Figure 7. Clipped data ( ' ) represent means of 28* and 50 quad- rats in the old (0) and young (0) fallows, respectively. Seeds were sown at the rate of 412 seeds/0.24m . * Discrepancy in sample size is a result of a sowing error. PLANTS / 0.24 M2 40 I00 ES UUIIII' l l 200 400 600 DAYS AFTER I APRIL I967 1111114111IIIIIIIIIJ A J O J A J 0 I967 I968 FIGURE 10 41 Highest mortality rates occurred during the winter and during summer droughts. Percent winter-kill was remarkably constant when corrected for normal post-fruiting mortality in the preceding fall. Drought was concurrent with exceedingly heavy grazing pressure in the summer of 1967 which resulted in complete defoliation of most of the rosettes. Grazers were not identified, but the defoliation coincided with the immigration of large numbers of mature and late instar Acridid grasshoppers following the harvest- ing of an adjacent hay field. Reproduction occurred in both the first and second growing seasons in the younger fallow, but was delayed in the older fallow until the second growing season. One indi- vidual in a clipped quadrat in the older fallow flowered in the first year, but failed to set seed. In the younger fallow, 5.5% of the plants alive at the end of the first growing season were reproductive. In the second year, 57% of the individuals in the younger fallow, and 7% in the older fallow were reproductive. Potential seed production based on estimates of numbers of flowers for the first growing season was 4900.1 5000 seeds per quadrat (0.24 m2) in the younger, and essentially none in the older fallow. Most of this potential seed cr0p was destroyed by herbivores prior to seed maturation. Actual seed production based on air-dry weight of seeds produced in the second growing season was 5500.: 1100 seeds per quadrat in the younger 42 fallow and 570 i.190 seeds per quadrat in the older fallow. Reproduction probability of Daucus in the second grow— ing season was negatively correlated with the cover of the more abundant species (Table 4). Reproduction probability declined non-linearly with increases in aggregate grass cover, i.e., the cover of the dominant herbaceous life form (Figure 11). The number of reproductives per quadrat was proportional to population density (r = 0.89, F1,518 = 1667). Since reproductive plants in the first year occurred in relatively few quadrats, correlations were not run between first year carrot reproduction and the percent cover of other species. Clipping the grasses and non-carrot vegetation during the first growing season markedly enhanced second year carrot reproduction. A mean of 28.0.i 5.1 individuals per clipped quadrat flowered in the younger, and 5.1 1.1.2 individuals flowered in the older fallow. These values represent two- and seven-fold increases relative to the unclipped quadrats, whereas densities in the clipped quad— rats at flowering were only 1.4 and 1.5 times the densi- ties in the unclipped quadrats. Flower and seed production were proportional to the number of reproductive individuals (Figures 12 and 15), implying that mean flower and seed yields per reproductive individual were nearly constant over the observed density range. The partitioning of seed production among the 45 TABLE 4 CORRELATION BETWEEN THE PERCENT COVER OF THE PROMINENT VASCULAR PLANTS AND THE PROBABILITY 0F FLOWERING IN THE 2nd GROWING SEASON SPECIES N r F P Eggs typhina 48 -0.492 14.68 < .001 Melilotus alga 69 -0.401 12.80 < .001 Egg pratensis 48 -0.841 110.85 < .001 Egg compressa 44 -0.588 7.44 < .01 Agropyron'repens 115 -0.726 127.97 < .001 Arcsin transformation (Snedecor, 1956) was used for repro- duction probabilities. 44 FIGURE 11 Relation between Total Grass Cover and Reproduction Prob- ability. Points represent means (1.2 S.E.) of untrans- formed data. Seed were sown December 1966 in plots fallowed since either summer 1964 or 1966 at the rate of 412 seeds/0.24m2. Data for the two ages of fallow were pooled. .° .5 9 *9}— PROBABILITY OF REPRODUCTION 9 IV .0 45 I \J 1968 I967 2b 40 éo 80 IOO 7. GRASS COVER FIGURE 11 46 FIGURE 12 Regression of Estimated Flower Production on the Number of Reproductives/0.24m2. Flower number estimates are based on two diameter measurements of each inflorescence. Data for the two fallows are pooled. The regression (r = 0.84, N = 94) is accompanied by confidence intervals. THOUSANDS OF FLOWERS/0.24 N' 30 25' N O G 5 47 Y=0.39+0.70 X FLOWERS - I967 20 4o ‘ so REPRODUCTIVES/O.24 M2 FIGURE 12 48 FIGURE 15 Regression of Seed Production in Grams Air Dry Weight on the Number of Reproductives. The regression (r = 0.86, N = 296) is accompanied by 95% confidence intervals. Data for the two fallows are pooled and quadrats without reproductives are excluded. GRAMS SEED / 0.24 M2 20 I6 i3 U 49 SEED - I968 L l l 20 4O 60 REPRODUCTIVE PLANTS /O.24 M2 FIGURE 15 50 reproductive individuals in a random sample of the pOpula- tion is shown in Figure 14. Approximately 50% of the seed sown in the vicinity of the parent plants in 1967 resulted in established seedlings in 1968, but annual reproduction by these individuals was absent. Very few new individuals (easily identified by their conspicuous cotyledons) were observed coming from this sowing in the second growing season (1969). The number of emerged seedlings resulting from seed of annuals and putative biennials was the same. These data were conse- quently lumped as "sparse Agrgpyron sod" (Figures 15 and 16) for comparison with concurrent germination in the second field eXperiment. Failure offSeedling Establishment Post-emergence seedling mortality had little effect on seedling establishment in 1968. ‘Mortality during the spring germination pulse was equivalent to only 5 to 14% of the total emergence (Figure 15). The highest percent mortality as well as the greatest emergence occurred in the presence of AgrOpyron sod. .The lowest seedling densi- ties were associated with gg§_compressa sod. Peak emergence rates occurred in late April but emerg- ence continued at low levels throughout the summer, despite the frequent occurrence of heavy rain that might be expected to give subsequent pulses of germination (Dale 51 .ucmHm me comm map Imam mo mam mm ommmmumxm ma pamfiw comm .5 mnsmwm mo mHmEmm umncmsq mm map Eoum mum mama .ucmam Hmwccmwm\sofluuspoum comm mo sowusnfluumfla mocmsvmnm dd flmDOHm 52 «a mmDOHm ._.z:UDQO~Em~_ ._<_ZZm_m 55 FIGURE 15 Histograms of Mean Seedling Emergence and Concurrent Mortality. Means (i 2 S.E.) are given for total emerg- ence (clear) and total mortality (cross-hatched) during the Spring of 1968 for seedlings resulting from seeds sown in December 1967. Seeds were sown on bare ground (B), in dense A ro ron repens sod (DA), in sparse Agropvron sod (SA), and in ggg_compressa sod (P). N = 50 for each treatment except Sparse Agropyron (N = 18). The B quadrats were in plots fallowed in 1967, the SA quadrats were in plots fallowed in 1966, and the P and DA quadrats were in plots fallowed in 1964. 54 r .fl CW 1. LT-.-.-L _ A _ _ 6 M. 2 O 8 6 4 2 I I I «2 3.0 x 323.. Lo $5.32 2.32 B SADA P 15 F I GURE 55 FIGURE 16 Mean Number of Live Daucus Plants Resulting from Seed Sown in December 1967 in Four Sod Types. The 1968 counts (clear) are all seedlings and the 1969 counts (checkered) are largely one-year-old rosettes. Counts were made May 8, 1968 and April 26, 1969. Rosettes resulting from emigrant seeds are included. Sample sizes and confidence intervals are as given in Figure 15. 56 «2 00.0 \ 2.2.415 LO mmmEDZ ZHuommmmH may no ucHGCAme may um OUOE mucsoo ucmmmumwu manEn: wuumwom .BOHHmm mo mmmfi 038 NEKmUmmm OSFH mo mumm on» um s3om mumupmso ca coHuosp Ioummm was Hm>a>nsm .mocmmumfim mcaapmmm usmoumm mo mumEEsm NH MMDOHm 65 n.“ HMDOHm Ow>O~=mmo mO¢U Owum v Z>>Om Omwm wt... ".0 thU¢mm m< 30:5. 30:5 02:0» 90 ammmmmaxm 20:039.. 8mm oz< 330 u n x 3330: "3.55 >9. 3 30:5 Lo m9: $5.950 z. a a _ Iv m I 0 AI & A1 a I 3 AI m2 AI 8 _ 1. P. 9 Doom. Sam «.0 m a no no -v m Al 0 Al 3. Al a A1 3 AI m2 I we _ A . A x o com. o No 0 an 2 2 p wwwwm O merDV>O E“: S: on :08 32v - 39 02.330: 532400-202 A3500 2 so 30.34“. “—0 mm0< O>>h Z. 9256.5 «2 \ mommm ownp .._O m._.<~_ 2.: ._.< ZZ/Om omh_>~5w .mUZmOmmim 0220mmw thU¢mm “.O >m<<fi23m AVSoc. I 230m 3mm 000— woo— koo— coo— 66 and 0.5 gm of seed, 40% produced between 0.5 and 1.0 gm, and 10% produced more than 1 gm. Koyama and Kira (1956) have shown that size distributions in plant pOpulations which are relatively free of interference are approximately normal, and tend to become log-normal with increasing inter— ference. This tendency is possibly accentuated in the seed yield distribution as a result of the allometric relation- ship between seed yield and plant size (Figure 4). .Further, since the reproduction probability of Daucus is correlated with plant size (cf. Hawthorn, 1951), the smallest plants do not appear in the reproductive class, resulting in a distribution approximating a log-normal truncated on the small size end. The seed yield per unit area, in contrast to the usually observed asymptotic or “parabolic" yield-density relationships (cf..DeWit, 1960; Bleasdale, 1967; Palmblad, 1968) observed in obligate annuals was directly proportional to density. Density, at the time of flowering, was deter- mined primarily by the initial density and the age at repro- duction, both of which were negatively associated with sod development (Table 5, Figure 11). The frequent correlation, in biennials (cf. Dickson, 1958), of the onset of reproduc- tion and plant size suggests that the reproductive delays in the field reflect sod-induced suppression of growth rates, such as were observed in the greenhouse. Similar reproduc- tive delays have been observed in grassland perennials in 67 the field (Linkola, 1955) and are implicit in the green- house data of Palmblad (1968). This suggests that reductions in reproductive rates associated with interfer- ence may, for perennials and facultative annuals, be ex- pressed as delayed reproduction rather than the reduced fecundity exhibited by obligate annuals. Delayed reproduction for a species like Daucus carota which lacks efficient mechanisms for both mass dispersal and vegetative multiplication, could cause a several fold difference in the size of the colonized area and possibly alter the persistence of the population. Short generation times would have little effect on population persistence if the site were to rapidly become seed saturated, i.e., if the effect of augmentation of the seed supply were negligible. If, however, seed saturation were absent or delayed, the shorter generation times observed in the early arriving populations might lead to rapid increases in density (cf. Cole, 1954; Lewontin, 1965). Since subsequent mortality appears to be independent of population density, higher initial densities would be expected to prolong the persist- ence of the population if conditions subsequently became unfavorable for further seedling establishment. If continued establishment remains possible, but re* stricted, the probability that a seed will land on a site suitable for germination would be approximated by the product of the number of "safe areas" (minimal areas of 68 substrate suitable for germination) and the number of poten- tially effective seeds (viable, nondormant seeds). If the total safe area is negatively correlated.with sod develop- ment, as the seedling emergence data suggest (Table 5), the number of safe areas would be eXpected to decline rapidly in young fallows. In older fellows, however, the smaller number of safe areas would probably decrease more slowly to a minimal level. As both the number of safe areas (safe sites of Harper et al., 1965) and the number of potentially effective seeds decline, their product, the probability that an effective seed will encounter a safe site, would be ex- pected to decrease very rapidly. Thus, an earlier arriving ,Daucus population which succeeds in establishing a high initial population density with consequent early and large seed production, would tend to have a higher probability of persistence than a later arriving one. Generalizing the findings of the present study sug- gests that the large density variations observed in natural- ly occurring populations in southern lower Michigan largely reflect various combinations of immigration time and con- tinuing diSturbance. In particular, the sharp disparity between the population densities in roadsides and adjacent, seemingly similar, old-field vegetations probably reflects enhanced seedling establishment in the periodically mowed roadside vegetation, coupled with occasional successful reproduction. CONCLUS IONS Conclusions Concerning Daucus Arrival in the third rather than the first year of a succession results in significant reduction in the density and potential growth rate of wild carrot popu- lations. For the years studied, post-emergence seedling mortality was less important than emergence success in determining the size of a Daucus pOpulation. Seedling establishment is negatively correlated with both total cover and the major component species of the established plant cover. The number of emerged seedlings is affected by the age of fallow, but the relative rate of emergence (numbers emerging per day as percent of total emergence) appar- ently is not. Intra-seasonal emergence rates decline rapidly follow- ing the spring emergence pulse. 69 10. 11. 12. 15. 7O Germination of seed surviving to the second growing season is modest and probably has little effect on population size of Daucus. Mortality rates of wild carrot populations are rela— tively independent of intraspecific density and the intensity of sod development. Mortality rates are highest during the winter and during summer drought. Reproduction is delayed in the presence of well- developed sods. Reproduction probability in the second growing season is negatively correlated with the percent cover of the more abundant species, and is markedly enhanced by partial removal of the shoots of associated plant species. Flower and seed yields per unit area of wild carrot are directly proportional to intraspecific density in the ranges encountered in this study. The partitioning of seed production among the members of the population is not normal, but rather appears to approximate a truncated log normal. The projected population duration, ignoring second and later generations, would approximate six to ten years. 1. 71 Conclusions General to Succession Small delays in arrival time may significantly alter both ecesis success and the rate of subsequent pOpula- tion buildup. The magnitude of the effect varies with the composition as well as the age of the previously established vege- tation. If Daucus is representative, the most sensitive life history determinants of population growth rate are seedling establishment and the age of reproduction. LI TERATURE CI TED LITERATURE CITED Bard, G. E. 1952. Secondary succession on the piedmont of New Jersey. Ecol. Monogr. 22:195-215. Beckwith, S. L. .1954. Ecological succession on abandoned farm lands and its relation to wildlife management. Ecol. Monogr. 24:549-576. Bleasdale, J. K. A. 1967. The relationship between the weight of a plant part and total weight as affected by plant density. J. Hort. Sci. 42:51-58. Borthwick, H. A. 1951. Carrot seed germination. Proc. Amer. Soc. Hort. Sci. 28:510-514. Cavers, P. B. and J. L. Harper, 1967. Studies in the dynamics of plant populations. I. The fate of seed and transplants introduced into various habitats. J. Ecol. 55:59-71. Clements, F. E. 1916. Plant succession: an analysis of the development of vegetation. .Carnegie Inst. Wash. Publ. No. 242, 512 pp. Cole, L. C. 1954. The population consequences of life history phenomena. Quart. Rev. Biol. 29:105-157. Dale, H. M. and P. J. Harrison. 1966. Wild carrot seeds: Germination and Dormancy. Weeds. 14:201-204. Davidson, R. L. 1962. The influence of edaphic factors on the species composition of early stages of the subsere. .J. Ecol. 50:401-410. Davidson, R. L. 1965. An eXperimental study of succession in the Transvaal HighveldI pp. 115-125. In: D. H. S. "Davis, D. De Meillon, J. S. Harrington, and M. Kalk (ed.), Ecological studies in southern Africa. W. Junk, The Hague. Davis, R. M. 1968. Effect of size area open to coloniza- tion on species composition in early old-field succession. 7M..S. Thesis. .Michigan State Univ. .104 pp. 72 75 DeWit, C. T. 1960. On competition. Versl. Landbouwk. Onderzock. No. 66(8):1-80. Dickson, H. M. 1958. The physiology and inheritance of flowering in carrots. Ph. D. Thesis. Michigan State Univ. 69 pp. Doneen, L. D. and J. H. MacGillivray. 1945. Germination (emergence) of vegetable seed as affected by different soil moisture conditions. Pl. Physiol. .18:524-529. Gleason, H. A. and A. Cronquist. 1965. Manual of vascular plants of northeastern United States and adjacent Canada. D. Van Nostrand Co., Inc., Princeton, N. J. 810 pp. Harper, J. L. 1955. The influence of the environment on seed and seedling mortality. VI. The effects of the interaction of soil moisture control and temperature on the mortality of maize grains. Ann. Appl. Biol. 45:696-708. Harper, J. L. 1965. Establishment, aggression, and co- habitation in Weedy Species, pp. 245-265.£ Ig; H. G. Baker and G. L. Stebbins. The genetics of colonizing species. Academic Press. New York. Harper, J. L. 1967. A darwinian approach to plant ecology. J. Ecol. 55:247-270. Harper, J. L., J. T. Williams, and G. R. Sagar. 1965. The behavior of seeds in soil. I. The heterogenity of soil surfaces and its role in determining the establishment of plants from seed. J. Ecol. 55:275- 286. Harrison, P. J. and H. M. Dale. 1966. The effect of graz- ing and clipping on the control of wild carrot. Weeds. 14:285-288. Hawthorn, L. R. 1951. Cultural studies of carrot steck- lings in relation to seed production. USDA Circ. 877:1-21. Heydecker, W. 1962. From seed to seedling: factors affect- ing the establishment of vegetable crops. Ann. Appl. Biol. 50:622-627. Heydecker, W. 1965. Laboratory studies in seedbed condi- tions. Hort. Rev. 2:97-105. 74 Holscher, C. E. (Chairman). 1959. Report of the committee on plant cover and competition, pp. 96-99. lg: Techniques and.methods of measuring Understory vegeta- tion, USDA, Southern Forest Exp. Sta. and Southeastern Forest Exp. Sta., Proc. Symp. Tifton, Georgia. Keever, C. 1950. Causes of succession on old-fields of the Piedmont, North Carolina. Ecol. Monogr. 20:229- 250. Kirk, R. E. 1968. »Experimental design: procedures for the behavioral sciences. ‘Wadsworth Publishing Co. Belmont, California. 577 pp. Kotowski, F. 1926. Temperature relations to germination of vegetable seed. Proc. Amer.-Soc. Hort. Sci. 25: 176-184. Koyama, H. and T. Kira. 1956. Intraspecific competition among higher plants. VIII. Frequency distribution of individual plant weight as affected by the inter- action between plants. J. Inst. Polytech. Osaka City Univ. Ser. D, 7:75-94. Lewontin, R. C. 1965. Selection for colonizing ability, pp. 77-94. IQ: H. G. Baker and G. L. Stebbins, The genetics of colonizing species. Academic Press. New York. Linkola, K. 1955. Uber die Dauer und Jahrenklassenver- haltnisse des Jugendstadiums bei einigen Wiesenstauden. Acta Forest. Fenn. 42(2):1-5. Malone, C. R. 1969. Effects of diazinon contamination on an old-field ecosystem. Amer. Midl. Nat. 82:1-27. Oosting, H. J. 1942. An ecological analysis of the plant communities of the Piedmont of North Carolina. Amer. Midl. Nat. 28:1-126. Palmblad, J. G. 1968. .Competition in experimental popu- lations of weeds with emphasis on the regulation of population size. Ecology. 49:26-54. Quagliotii, L. 1967. Effects of different temperatures on stalk development, flowering habit, and sex expres- sion in the carrot Daucus carota L. Euphytica. 16:85—105. Rice, E. L. 1964. Inhibition of nitrogen-fixing and nitri— fying bacteria by seed plants. Ecology. 45:824-857. 75 E. L. .1967. Chemical warfare between plants. Bios. 58:67-74. Snedecor, G. W. 1956. Statistical methods applied to ex- periments in agriculture and biology. Iowa State Univ. Press. Ames, Iowa. 554 pp. Ursie, S. J. and D. C. Clurkin. ,1959. Small plots for measuring vegetation composition and cover, pp. 70-78. In; Techniques and methods of measuring understory vegetation, USDA Southern For. Exp. Sta. and South- eastern For. Exp. Sta., Proc. Symp. Tifton, Georgia. Weigert, R. G. and F. C. Evans. .1964. Primary production and the disappearance of dead vegetation on an old- field in southeastern Michigan. Ecology. 45:49-65. MICHIGAN STATE UNIV. LIBRARIES III III II I 6 312931075 5438