HORTICULTURAL ENTOMOLOGY Flies (Diptera: Muscidae: Calliphoridae) Are Efficient Pollinators of Allium ampeloprasum L. (Alliaceae) in Field Cages

STEPHEN L. CLEMENT,1,2 BARBARA C. HELLIER,1 LESLIE R. ELBERSON,1 1 3 RUSSELL T. STASKA, AND MARC A. EVANS

J. Econ. Entomol. 100(1): 131Ð135 (2007) ABSTRACT In conjunction with efforts to identify efÞcient insect pollinators for seed multiplication of cross-pollinated plant species stored and maintained by USDAÐARS Western Regional Plant Introduction Station (WRPIS), experiments were conducted to assess and compare the efÞciency of the house ßy, Musca domestica L. (Diptera: Muscidae), and Calliphora vicina Robineau-Desvoidy (Diptera: Calliphoridae), and different densities of each ßy species, to pollinate leek, Allium ampe- loprasum L., plant inventory (PI) accessions in Þeld cages for seed yield maximization and high germination. Cages with ßowering plants were exposed to 0 ßies or stocked with 100, 250, and 500 M. domestica or C. vicina pupae per week for 6 (2002) and 7 (2004) wk. Seed yield (weight per cage) increased linearly as ßy densities (C. vicina or M. domestica) increased from 0 to 500 pupae per week, with 500 ßy cages averaging 340.7 g (C. vicina) and 70.5 g (M. domestica) of seed in 2002 (PI 368343) and 615.3 g (PI 168977) and 357.5 g (PI 368343) in 2004 when only the C. vicina was used. For 0, 100, and 250 ßy cages, seed yields averaged between 2.3 and 175.3 g in 2002 and 10.7 and 273.1 g in 2004. Mean 100-seed weights between treatments ranged narrowly between 0.4 and 0.5 g in 2002 and 0.3 and 0.4 g in 2004, and germination rates of seed lots from “ßy cages” were mostly Ն80% in both years. The C. vicina is an efÞcient and cost-effective pollinator ($388.97 for pupae and shipping, compared with $2,400 for honey bee, Apis mellifera L. [Hymenoptera: Apidae], nuclei) for caged leek accessions, with 250 and 500 C. vicina pupae per week required to produce sufÞcient seed (130 g) to Þll an accession storage bag in the WRPIS gene bank.

KEY WORDS leek, Calliphora vicina, Musca domestica, germplasm regeneration, insect pollination

Regeneration of seed-propagated plant species is one plant species, and a seed lot inventory of Ϸ72,500 of the most important tasks for gene bank curators and accessions) in the U.S. National Plant Germplasm Sys- managers. This activity, which normally involves Þeld tem. A sizeable percentage of WRPIS germplasm is nurseries with multiple plant accessions, is necessary cross-pollinated species, including species in the ge- to replenish seed stocks in cold storage that are low in nus Allium (Alliaceae). Among the 105 species and 978 viability and supply. However, there is a risk of unin- accessions of Allium in WRPIS collections are 204 leek tended hybridization among accessions of allogamous accessions, A. ampeloprasum L. [sometimes classiÞed (outcrossing) entomophilous plant species if acces- as A. porrum (L.)]. Leek is a vegetable belonging to sions are clustered in Þeld nurseries. To prevent in- the onion family and is used as an ingredient in soups sect-meditated cross-pollination between clustered and salads. There are three other horticultural groups accessions in open-Þeld nurseries, gene bank curators in A. ampeloprasum (Kurrat, Great-headed Garlic, and use spatial isolation of accessions or Þeld cages over Pearl Onion), which are mainly cultivated in Europe plants with suitable insect pollinators (Clark et al. and the Middle East, but on a smaller scale than leeks. 1997, Schittenhelm et al. 1997). Leeks are an important crop in Europe. Wild A. am- The USDAÐARS Western Regional Plant Introduc- peloprasum originates from Portugal in the west tion Station (WRPIS), Pullman, WA, is a major gene through the Mediterranean countries to Iran and else- bank for plant collections (358 plant genera, 2,194 where in the Middle East (van der Meer and Hanelt 1990, Brewster 1994). Leeks, which do not normally bulb like onions, are grown from seed. Mention of a trademark or a product does not constitute a guarantee A leek seed stalk can approach 1.2 m in height and or warranty by the USDA and does not imply its approval over other suitable products. is terminated by a single umbel containing hundreds 1 USDAÐARS, Plant Germplasm Introduction and Testing Research of bell-shaped ßowers with light purple petals. These Unit, 59 Johnson, Hall, Washington State University, Pullman, WA protandrous ßowers facilitate out-crossing. However, 99164-6402. protandry only offers a partial barrier to self-pollina- 2 Corresponding author, e-mail: [email protected]. 3 Department of Statistics, Washington State University, Pullman, tion because pollen from ßowers that typically open WA 99164-3144. over 2Ð4 wk can fertilize receptive stigma of more 132 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 100, no. 1 advanced ßowers in the same umbel (McGregor 1976, CA (M. domestica) and Forked Tree Ranch, Porthill, Brewster 1994). This potential for self-pollination ID (C. vicina). Once a week between 8 July and 12 complicates seed regeneration because leeks are sub- August 2002, these vendors express-mailed 5,000Ð15,000 ject to severe inbreeding depression, which can lead pupae of their respective ßy species to Pullman, WA. to severe loss of vigor and reduced survival in seedlings In 2004, only C. vicina pupae were required for this derived from self-pollination (Currah 1986, Brewster study, and 15,000 pupae were purchased from Forked 1994, Silvertand et al. 1995). Therefore, leek acces- Tree Ranch and shipped weekly, 6 July to 16 August. sions must be regenerated without risk of contamina- Caged Field Experiments. The experiments were tion by pollen from neighboring accessions, yet under conducted, 2001Ð2002 and 2003Ð2004, at a major conditions that encourage high rates of cross-pollina- WRPIS seed regeneration site (25 ha) bordering the tion. These requirements can be accomplished by cou- Snake River and near Central Ferry in southeastern pling a suitable insect pollinator with caged plants of Washington (46Њ 40Ј 13Љ N, 117Њ 45Ј 8Љ W). Plots were a leek accession. established using 4-wk-old seedlings grown from The honey bee, ßies in the genera Calliphora, Lu- seed in a glasshouse and transplanted to the Þeld on cilia (Diptera: Calliphoridae), and Eristalis (Diptera: 26 and 27 September 2001 and 15Ð17 October 2003. Syrphidae), and the solitary bee Osmia rufa L. (Hy- The seedlings were transplanted into black plastic menoptera: Megachilidae) have been used to polli- ground cover for weed suppression and watered nate caged onions, A. cepa L. (Moffett 1965; Bohart et with drip irrigation as needed. Nitrogen was applied al. 1970; Free 1970; Schittenhelm et al. 1997; Currah before transplanting at 45 kg ⅐ haϪ1. No pesticides and Ockendon 1983, 1984). Although some of these were applied and weed control was by cultivation investigators found that ßies and honey bees were before transplantation and hand hoeing thereafter. equivalent in their suitability as pollinators for caged A plot (4 by 7 m) consisted of four 6-m-long rows onions (Currah and Ockendon 1983, 1984), others on 0.3-m centers, with each row containing 18Ð20 reported that calliphorid and Eristalis ßies were more evenly spaced transplants for a total of 75Ð80 plants suitable than honey bees for cross-pollinating onions per plot. The spacing between plots was 1.5 m. Each in cages (Bohart et al. 1970, Free 1970). It is well plot was thinned to 53 and 49 bolting plants on 8 May known that onion ßowers are unattractive to the 2002 and 16 June 2004, respectively. In June 2002 and honey bee (Bohart et al. 1970, McGregor 1976, Silva et 2004, before plants ßowered, each plot was covered al. 2003), so it is not surprising that some ßy species with a screen Þeld cage (4 by 7 by 2 m) (Synthetic might be better pollinators for caged onions, as well as Industries, Gainesville, GA) and randomly assigned a other alliums. Schittenhelm et al. (1997) did not rec- ßy density treatment: 100, 250, and 500 M. domestica or ommend the solitary bee O. rufa as a pollinator for C. vicina pupae, and one 0 ßy cage (no pupae added) caged onions because it was repelled by the plants. in 2002; and 0, 100, 250, and 500 C. vicina pupae in 2004. Without published accounts on the efÞciency of These assignments resulted in seven treatment cages different insects for pollinating A. ampeloprasum in per block in 2002 (one leek accession [PI 368343] ϫ cages, but armed with the results of the aforemen- two ßy species ϫ 3 ßy rates ϩ one 0 ßy cage) and eight tioned reports on A. cepa pollination, we chose to treatment cages per block in 2004 (two leek accessions evaluate two ßy species for their efÞciency to pollinate [PI 368343, PI 168977] ϫ one ßy species ϫ 4 ßy rates). leek accessions in cages. In addition, ßy pupae were Each experiment was replicated four times in random- available from commercial suppliers, whereas honey ized complete blocks for a total of 28 and 32 cages in bees, Apis mellifera L. (Hymenoptera: Apidae), were 2002 and 2004, respectively. Cages were secured to the unavailable at the start of this study. ground by using metal stakes and bottom edges along The research objective of this multiyear study was each side were buried in soil. to assess and compare the efÞciency of the house ßy, Fly pupae were placed in cages at the onset of Musca domestica L. (Diptera: Muscidae), and Calli- ßowering and until ßowering was complete. There phora vicina Robineau-Desvoidy (Diptera: Callipho- were six weekly release dates (8 JulyÐ12 August) in ridae), and variable densities of each species, to pol- 2002 and seven (6 JulyÐ16 August) in 2004. Pupae were linate caged leek accessions in the Þeld for seed yield placed in a plastic pot (15 cm in diameter) positioned maximization and high germination. Using our results, on its side on the soil surface of each cage. The pot was we identify the best ßy pollinator for the WRPIS leek held in place by a few grams of top soil. This technique regeneration program. Last, we compare the econom- sheltered pupae from wind and direct sunlight and ics of ßies versus honey bees for pollination of caged allowed us to easily monitor adult emergence. Flies leek accessions. began emerging from pupae within one day and usu- ally continued for another 2Ð3 d. A plastic pan (29 by 35 cm) Þlled with 15 cm of water-saturated vermicu- Materials and Methods lite (water added as needed) was placed in a corner Sources of Plants and Flies. The plants selected for of each cage to provide a constant supply of moisture this study were grown from seed of two plant inven- for ßies during periods of ambient temperature ex- tory (PI 368343, from the former Yugoslavia; PI tremes (10.2Ð41.8ЊC, 2002; 12.5Ð39.7ЊC, 2004). 168977, from Turkey) accessions of A. ampeloprasum We thinned to 47 primary umbels per cage (one per obtained from the WRPIS gene bank. In 2002, ßy plant) on 22 August 2002 and 16 September 2004. The pupae were purchased from Rincon-Vitova, Ventura, mature umbels were hand-harvested during the last February 2007 CLEMENT ET AL.: FLY POLLINATION OF A. ampeloprasum 133

Table 1. Partial ANOVA results for the effects of fly pollinator species and pollinator density on seed yield of A. ampeloprasum, 2002

Seed wt/cage 100-seed wt/cage Seed germination Source df F P FPFP Fly species (F) 1 292.13 Ͻ0.0001 8.34 0.0098 0.00 0.9750 Fly density (D) 2 109.64 Ͻ0.0001 2.21 0.1385 0.47 0.0346 F ϫ D 2 44.95 Ͻ0.0001 0.23 0.7951 0.20 0.8201 Linear trend for D 1 219.15 Ͻ0.0001 4.07 0.0589 0.03 0.8620 Nonparallel regression for F 1 88.82 Ͻ0.0001 0.13 0.7268 0.29 0.5945 Error 18 week of October 2002 and 2004, after which they were species on seed yield per cage were signiÞcant (Table hand-threshed with a Vogel thresher (Bills Welding, 1). Seed yield (weight per cage) increased linearly Pullman, WA) and cleaned of nonseed material by (P Ͻ 0.0001) as densities increased for C. vicina (seed using an in-house built laboratory air screen cleaner. yield ϭ 0.982 ϩ 0.682 ⅐ ßy density; r2 ϭ 0.97) and M. Cleaned and dried seed lots were stored at room tem- domestica (seed yield ϭϪ2.680 ϩ 0.135 ⅐ ßy density; perature (Ϸ21ЊC) before weighing seed in November r2 ϭ 0.72). Except for the main effect of ßy species, no and December 2002 and 2004. Data were recorded for other signiÞcant P values were detected for the 100- total seed weight per cage (47 umbels) and for the seed weights (Table 1). This main effect is illustrated weight of a random sample of 100 seeds per cage. In by mean Ϯ SEM 100-seed weights (grams) of 0.44 Ϯ addition, a representative sample of 100 seeds from 0.01 (0 ßy cages), 0.44 Ϯ 0.01 (M. domestica cages), each cage was germinated between sections of moist and 0.41 Ϯ 0.01 (C. vicina cages) (LSD ϭ 0.02). Over- blotter paper in a plastic germination box (11.5 by 11.5 all, mean 100-seed weights ranged narrowly between cm) (50 seeds per box). Seed germination at the “loop 0.4 and 0.5 g (Table 2). stage” (Brewster 1994) was recorded after 21 d at 20ЊC The seed yield data in Table 2 illustrate the positive (photoperiod of 8:16 [L:D] h) in a controlled tem- effect of high ßy densities on pollination and seed perature incubator (Hoffman Manufacturing Inc., Al- production of caged leek plants. In 2002, signiÞcantly bany, OR). This germination procedure is in accor- more seed (x ϭ 340.7 g) was produced by caged plants dance with that established by the Association of that received 500 C. vicina pupae per week compared OfÞcial Seed Analysts (AOSA 1998). with 0, 100, and 250 M. domestica and C. vicina pupae Statistical Analysis. Data (seed weights [total, 100 per week (averages of 2.3Ð175.3 g). Additionally, the seed sample] and germination rates [%] per cage) results for seed weight per cage clearly show that C. from each year were analyzed separately using PROC vicina was superior to M. domestica for cage pollination GLM based on a randomized complete block design and seed production (Table 2). with a two-way treatment structure (SAS Institute The superior performance of C. vicina in 2002 led to 2003). Analysis of 2002 data assessed the main effects its exclusive use in 2004 and for pollination of two leek for ßy species (M. domestica and C. vicina) and density accessions. Analysis of the 2004 data revealed signif- (0, 100, 250, and 500 pupae) per ßy species, and the icant effects of accession, ßy density, and the acces- interactions, on seed yield and germination of PI sion ϫ ßy density interaction on seed yield per cage. 368343. Analysis of 2004 data (seed yield and germi- In addition, this analysis generated a signiÞcant linear nation comparisons, and the interactions) involved PI trend for ßy density and nonparallel regressions for 368343 and PI 168977 and different C. vicina densities accessions (Table 3). Seed yield (weight per cage) (0, 100, 250, and 500 pupae). Orthogonal polynomial increased linearly (P Ͻ 0.0001) for PI 368343 (seed contrasts were used to assess a linear trend in seed yield ϭ 22.52 ϩ 0.716 ⅐ ßy density; r2 ϭ 0.89) and PI yield response due to ßy density and to determine whether the linear trends were parallel for two ßy Table 2. Yields and germination rates of seed produced by species (2002) and two leek accessions (2004). Mean caged leek plants (PI 368343) exposed to different fly pollinators comparisons were made using Fisher least signiÞcant and densities, 2002 difference (LSD) (␣ ϭ 0.05 for all comparisons). Re- Ϯ gression analysis was performed to using the PROC Mean SEM Diptera Fly density 100-seed Seed Seed wt/ REG procedure to assess the linear relationship be- species (no. pupae/wk) wt/cage germination cage (g) tween seed yield and ßy density (SAS Institute 2003). (g) (%) M. domestica 100 11.8 Ϯ 3.6a 0.5 Ϯ 0.0 87.8 Ϯ 2.4 Results and Discussion 250 19.7 Ϯ 3.9a 0.4 Ϯ 0.0 83.5 Ϯ 4.1 500 70.5 Ϯ 15.5b 0.4 Ϯ 0.0 86.0 Ϯ 1.5 C. vicina 100 65.2 Ϯ 3.9b 0.4 Ϯ 0.0 85.5 Ϯ 0.9 For 2002, the effects of ßy species and ßy density on 250 175.3 Ϯ 13.9c 0.4 Ϯ 0.0 84.5 Ϯ 2.9 seed yield per cage of PI 368343 were signiÞcant. Also, 500 340.7 Ϯ 19.1d 0.4 Ϯ 0.0 87.5 Ϯ 2.9 there was a signiÞcant ßy ϫ ßy density interaction on No ßies 0 2.3 Ϯ 0.9a 0.4 Ϯ 0.0 78.3 Ϯ 5.7 seed yield per cage, indicating a differential effect of ϭ ßy density on the pollination efÞciency of M. domes- Means (n 4) in the one column followed by a common letter are not signiÞcantly different (P ϭ 0.05; LSD). Mean comparisons were tica and C. vicina. As well, the linear trend for ßy not computed for the values in the other columns because ßy spe- density and nonparallel regressions for the effect of ßy cies ϫ ßy density interactions were not signiÞcant. 134 JOURNAL OF ECONOMIC ENTOMOLOGY Vol. 100, no. 1

Table 3. Partial ANOVA results for the effects of fly pollinator density on seed yield of two leek accessions (PI 168977 and PI368343), 2004

Seed wt/cage 100-seed wt/cage Seed germination Source df F P FPFP Accession (A) 1 17.50 Ͻ0.0004 1.02 0.3236 19.15 0.0003 Fly density (D) 3 111.87 Ͻ0.0001 6.87 0.0021 6.73 0.0023 A ϫ D 3 8.99 Ͻ0.0005 2.06 0.1359 0.57 0.6402 Linear trend for D 1 334.82 Ͻ0.0001 19.80 0.0002 19.80 0.0002 Nonparallel regression for A 1 20.98 Ͻ0.0002 3.20 0.0881 1.58 0.2231 Error 21

168977 (seed yield ϭ 3.169 ϩ 1.194 ⅐ ßy density; r2 ϭ as revealed by the generally higher mean values for PI 0.94) because C. vicina densities for each of these 168977, and an overall linear trend for ßy density accessions increased from 0 to 500 pupae per week. (Tables 3 and 4). The analysis also revealed a signif- Our analysis of 100-seed weight data detected signif- icant effect for ßy density on the germination rates of icance for ßy density and a linear trend for ßy density each accession (Table 3), which can be discerned (Table 3). The signiÞcance for ßy density (across the from the results in Table 4. two accessions) is illustrated by mean Ϯ SEM 100-seed Three general conclusions can be gathered from our Ϯ Ϯ weights (grams) of 0.37 0.01 (0 ßy cages), 0.36 overall results. First, insect pollination is essential for Ϯ 0.01 (100 C. vicina), 0.34 0.01 (250 C. vicina), and high seed yields of leek accessions in Þeld cages. Sec- Ϯ ϭ 0.33 0.01 (500 C. vicina) (LSD 0.02). The differ- ond, ßies possess desirable attributes for insect polli- ences between mean 100-seed weights across all 2004 nation of leek accessions, with C. vicina superior to treatments were not pronounced (0.3Ð0.4 g) (Table M. domestica for pollination and high seed production 4); however, these means are generally lower than the of caged plants. The 250 and 500 C. vicina treatments 2002 values (Table 2) for reasons not evident to us. produced the highest seed yields (weight per cage) in Caged leeks stocked weekly with 500 C. vicina pu- 2002 and 2004. By contrast, mean 100-seed weights pae produced the highest seed yields in 2004 (Table 4), a pattern also recorded in 2002 (Table 2). Also, among treatments were relatively constant each year. more seed, on average, was produced by PI 168977 Finally, there was a trend for higher germination rates than by PI 368343 in 2004, irrespective of C. vicina of seed lots from “ßy cages” compared with “0ßy density (Table 4). Interestingly, mean seed weights cages.” from 2002 and 2004 caged plants of PI 368343, across A minimally successful WRPIS Allium seedÐregen- all ßy densities, were roughly equivalent (Tables 2 eration cycle requires the production of 10,000 seeds and 4). per accession, which equates to Ϸ36gofA. ampelo- Main and interaction effects for seed germination prasum seed. However, a much larger amount of seed rates in 2002 were not statistically signiÞcant (Table (130 g) is required to Þll an accession storage bag in 1), with mean values from 78.3% (0 ßy cages) to 87.8% the WRPIS gene bank. Our results show that 250 and (100 M. domestica) (Table 2). By contrast, mean ger- 500 C. vicina pupae per week were required for 47 mination rates differed signiÞcantly (P Ͻ 0.05) among caged plants to produce at least 130 g of seed. More- 2004 treatments (Table 4). Regarding these 2004 ger- over, we posit that these high C. vicina densities mination rates, there was a signiÞcant accession effect, effected high rates of cross-pollination. Good cross- pollination can lead to a high degree of heterozy- gosity and the production of high quality seed in a Table 4. Yield and germination rates of seed produced by caged plants of two leek accessions (PI 168977 and PI 368343) leek accession. It also can negate the inßuence of exposed to different densities of the fly pollinator C. vicina, 2004 deleterious alleles responsible for inbreeding de- pression and other genetic defects (Currah 1986, Mean Ϯ SEM Brewster 1994). Fly density 100-seed Seed Accession Seed wt/ Expenses for M. domestica and C. vicina (pupae (no. pupae/wk) wt/cage germination cage (g) (g) (%) and shipping) were $298.20 and $178.83, respec- tively, in 2002 and $210.14 for C. vicina in 2004. Had Ϯ Ϯ Ϯ PI 368343 0 10.7 2.1a 0.4 0.0 69.8 1.4a honey bees been available for this study, the cost for 100 75.5 Ϯ 9.0ad 0.3 Ϯ 0.0 73.3 Ϯ 2.7ab 250 255.1 Ϯ 37.1b 0.3 Ϯ 0.0 78.3 Ϯ 1.0bc a four-frame nucleus of A. mellifera would have 500 357.5 Ϯ 17.0c 0.3 Ϯ 0.0 83.3 Ϯ 2.3cd been $50.00. Thus, the expense of using 24 honey PI 168977 0 20.2 Ϯ 1.1a 0.3 Ϯ 0.0 80.0 Ϯ 1.5bc bee nuclei in each year of this study would have 100 119.3 Ϯ 10.1d 0.4 Ϯ 0.0 83.0 Ϯ 3.0cd 250 273.1 Ϯ 31.5b 0.3 Ϯ 0.0 84.5 Ϯ 2.3cd totaled at least $2,400, an amount that far exceeds 500 615.3 Ϯ 52.9e 0.3 Ϯ 0.0 88.0 Ϯ 3.7d the amount paid for the ßy pupae. In conclusion, our results show that C. vicina (250Ð500 ßies per week Means (n ϭ 4) in the two columns followed by a common letter are per cage) is an efÞcient and cost-effective pollinator not signiÞcantly different (P ϭ 0.05; LSD). Mean comparisons were not computed for the values in the middle column because the for seed production of leek accessions at the WRPIS accession ϫ ßy density interaction was not signiÞcant. gene bank. February 2007 CLEMENT ET AL.: FLY POLLINATION OF A. ampeloprasum 135

Acknowledgments Currah, L., and D. J. Ockendon. 1984. Pollination activity by blowßies and honeybees on onions in breedersÕ cages. We are grateful to Kurt Tetrick and Scott McGee for Ann. Appl. Biol. 105: 167Ð176. expert help and assistance with establishing and maintaining Free, J. B. 1970. Insect Pollination of crops. Academic, New the Þeld plots, and to Jacqueline Cruver for seed cleaning York. operations. Rich Hannan provided valuable input on an early Moffett, J. O. 1965. Pollinating experimental onion varieties. draft of this manuscript. Am. Bee J. 105: 378. McGregor, S. E. 1976. Insect pollination of cultivated crop plants. U.S. Dep. Agric. Handb. 496. References Cited SAS Institute. 2003. SAS userÕs guide: statistics, version 9.1. SAS Institute, Cary, NC. [AOSA] Association of Official Seed Analysts. 1998. Rules Schittenhelm, S., T. Gladis, and V. R. Rao. 1997. EfÞciency of for testing seeds. AOSA, Las Cruces, NM. various insects in germplasm regeneration of carrot, onion Bohart, G. E., W. P. Nye, and L. R. Hawthorn. 1970. Onion and turnip rape accessions. Plant Breed. 116: 369Ð375. pollination as affected by different levels of pollinator Silva, E. M., B. B. Dean, and L. K. Hiller. 2003. Honey bee activity. Utah Agric. Exp. Stn. Bull. 482. (Hymenoptera: Apidae) foraging in response to precon- Brewster, J. L. 1994. Onions and other vegetable alliums. ditioning with onion ßower scent compounds. J. Econ. CABI Publishing, CAB International, Wallingford, United Entomol. 96: 1510Ð1513. Kingdom. Silvertand, B., E. Jacobsen, and A. M. van Harten. 1995. Ge- Clark, R. L., H. L. Shands, P. K. Bretting, and S. A. Eberhart. netic variation and control of plant regeneration in leek 1997. Managing large diverse germplasm collections. (Allium ampeloprasum L.). Plant Breed. 114: 333Ð336. Crop Sci. 37: 1Ð6. van der Meer, Q. P., and P. Hanelt. 1990. Leek (Allium Currah, L. 1986. Leek breeding: a review. J. Hort. Sci. 61: ampeloprasum), pp. 179Ð196. In H. D. Rabinowitch and 407Ð415. J. L. Brewster [eds.], Onions and allied crops vol. 3. CRC, Currah, L., and D. J. Ockendon. 1983. Onion pollination by Boca Raton, FL. blowßies and honeybees in large cages. Ann. Appl. Biol. 103: 497Ð506. Received 18 May 2006; accepted 15 October 2006.