Materials and Methods

Experiments were conducted at the Agricultural Research Station of the U.S. Department of Agriculture (USDA), Salinas, CA. We screened spinach collections from the USDA (seeds were provided by the North Central Regional Plant Introduction Station, Iowa State University, Ames, IA) for leafminer resistance in a preliminary study. The collection includes 332 accessions of cultivated spinach, four accessions of Spinacia turkestanica (Ames 23666, PI 494751, PI 604792, and PI 608713), and two accessions of S. tetrandra (Ames 23664 and PI 608712). Sixteen seeds from each accession were planted in a plastic pot (10 × 10 × 10 cm) with 2 sand:1 soil (by volume) in a greenhouse on 12 Aug. 2001, and seedlings were thinned to 10 plants per pot. Plants were moved into an outdoor insect cage (2 m high × 4 m wide × 8 m deep) made of polypropylene shadecloth for resistance screening 5 weeks after planting. Lettuce leaves with leafminer mines were collected from newly harvested fields around Salinas and hung in the shade to allow leafminer larvae to emerge from the leaves and pupate. Pupae were collected and put in plastic containers to allow adult flies to emerge. Approximately 3500 flies were then released in the outdoor cage to feed on the spinach plants. After 10 d, number of stings per unit area was counted on the leaf with most leafminer stings on each plant using an optical glass binocular magnifier (OptiVisor; Donegan Optical Co., Lenexa, KS), and the number of plants with mines for each accession was recorded.

From the preliminary screening, 22 accessions with the fewest stings and 10 accessions with the most stings per unit leaf area were selected for further testing. These genotypes plus six commercial hybrid cultivars and two Chinese local varieties were planted in plastic pots (10 × 10 × 10 cm) with soil on 15 Aug. 2002 and were thinned to five plants per pot. Plants were placed in the outdoor insect cage 6 weeks after planting and were arranged in a randomized complete block with a single pot as the experimental unit and eight replications. Approximately 4300 leafminer flies were released in the cage. Stings per unit area were counted on the leaf with the highest sting density on each plant 10 d after the introduction of flies in the cage. Mines per plant and the fresh weight of the plant excluding roots were also recorded. The same 40 genotypes were planted in the field in a randomized complete block design with eight replications on 29 Aug. 2002 and on 3 Aug. 2004. Each plot consisted of five plants of a genotype with 30 cm between plants and 35 cm between rows on 1-m wide double-row beds. Leafminer stings were counted in a 20-cm² leaf area with the highest sting density on each plant 5 weeks after planting. Mines per plant and plant weight excluding roots were also recorded.

Per-plant values were averaged and analysis was conducted on the basis of pot or plot means. Data were analyzed by analysis of variance using the general linear model procedure of JMP v. 5 (SAS Institute, Cary, NC). Genotype was considered the fixed effect, and replication was considered the random effect. For comparisons between genotypes, least significant differences were calculated with an error rate of P = 0.05. A correlation matrix for genotype means was calculated for all variables using the multivariate platform of JMP. To test differences in rank order among the genotypes grown in different environments, Spearman's rank correlations were calculated (Steel and Torrie, 1980).

Results and Discussion

None of the genotypes tested was immune to leafminers, because all genotypes had at least a few stings or mines. Significant genotypic differences were found for leafminer stings per unit leaf area, mines per plant, mines per 100 g plant weight, and plant weight both in the outdoor insect cage and in the field (Table 1). Leafminer stings per square centimeter leaf area ranged from 1.8 to 7.1 in the cage, from 10.5 to 30.8 in 2002 in the field, and from 9.9 to 23.8 in 2004 in the field for the genotypes tested. Mines per plant varied from 0.4 to 3.1 in the cage, from 6.3 to 38.4 in 2002, and from 14.8 to 79.5 in 2004 in the field. Because plant sizes differed among plants, we divided mines per plant by plant weight to standardize the result. Mines per 100 g fresh weight ranged from 5.6 to 85.1 in the insect cage, from 3.5 to 48.1 in 2002 in the field, and from 9.5 to 55.5 in 2004 in the field. These results suggest that there is substantial genetic variation among spinach genotypes in their suitability as a leafminer host.

Table 1. Means and least significant differences (lsd) at P ≤ 0.05 for traits of 40 spinach genotypes tested in an outdoor insect cage and in the field in Salinas, CA, in 2002 and 2004.

View Fullscreen

None of the four accessions of Spinacia turkestanica and two accessions of S. tetrandra had low leafminer sting density in the preliminary screening and they were not included in further cage and field tests. PI 274065 (cultivar Wisemona) had the lowest sting density (10.5 and 9.9 sting/cm² in 2002 and 2004) in the field where the six commercial cultivars averaged 20.4 and 17.3 stings per square centimeter leaf area in 2002 and 2004, respectively. Two accessions from Turkey showed the fewest mines per 100 g fresh weight (3.5 and 9.5 for PI 174385 and 4.4 and 10.5 for PI 169673 in 2002 and 2004, respectively) among genotypes in the field as compared with 15.3 and 26.5 for the six commercial cultivars in 2002 and 2004, respectively. These results showed a great potential to improve the level of leafminer resistance in the current spinach cultivars.

Rank order of stings per square centimeter leaf area did not significantly change for the genotypes in the cage, 2002 field test, and 2004 field test as indicated by the high Spearman's rank correlation (Table 2). The sting results from different tests were also highly correlated (Table 3). The consistency in performance demonstrated that differences in sting density were stable and a cage test can be used to screen germplasms for fewer leafminer stings. It also suggests that sting per unit leaf area is a reliable trait for the selection of leafminer feeding nonpreference. Similar results were found in a study of leafminer resistance in lettuce (Mou and Liu, 2003).

Table 2. Spearman's rank correlation coefficients of leafminer-resistant traits calculated from means of 40 spinach genotypes tested in an outdoor insect cage and in the field in Salinas, CA, in 2002 and 2004.

View Fullscreen

Table 3. Phenotypic correlation coefficients of traits calculated from means of 40 spinach genotypes tested in an outdoor insect cage and in the field in Salinas, CA, in 2002 and 2004.

View Fullscreen

Although the rank correlation for mines per plant or per 100 g plant weight was highly significant between 2002 and 2004 field tests, the correlation was only moderate or nonsignificant between cage and field experiments (Table 2). This may be partly because the level of mine damage in the cage was low (approximately one mine per plant on average; Table 1). These results suggest that mines per plant or per unit plant weight are also relatively stable over different years in the field, but a cage test may not be a good method of germplasm screening for resistance to mines.

The majority of stings are caused by feeding activities of leafminer adults. It is interesting to know whether leafminer flies tend to lay fewer eggs on a spinach plant if they do not like to feed on that plant. Leafminer eggs are difficult to count because they are tiny and are laid within leaf tissue. We had to use number of mines as an indirect measure of eggs laid in the plant. Stings per unit leaf area were not correlated with mines per plant or per 100 g plant weight either in the insect cage or in the field (Table 3). This suggests that feeding nonpreference does not necessarily mean oviposition–nonpreference for a spinach genotype, and these two traits can be improved independently.

In these choice tests, fewer stings or mines suggest host nonpreference (antixenotic resistance). Resistance based on antixenosis would be desirable because even the losses in photosynthetic capacity and appearance caused by adult feeding and oviposition would be reduced (Trumble et al., 1985). This is especially important for spinach because more than 95% of the value of spinach crop in the United States is from fresh market use (National Agricultural Statistics Service, 2007), and quality standards for fresh market spinach are extremely high. Antixenosis could prompt leafminer movements to weeds or crops tolerant to insect damage. For example, broccoli and cauliflower (both important crops in central California) with six or more leaves are rarely damaged by leafminers regardless of population numbers (University of California, 1992).

Sting per square centimeter leaf area and mine per 100 g plant weight had little correlation with plant weight in cage and field tests (Table 3). This suggests that leafminer sting and mine densities are not associated with plant biomass, and it is possible to improve and combine leafminer resistance and yield traits in a spinach cultivar.

Chemical control of leafminers usually lasts only a short period of time, and adult control with contact insecticides is especially unsatisfactory because flies can easily move around, and the treated field is subject to reinfestation from adjacent untreated crops and weeds (LeStrange et al., 1999). Many studies have shown that leafminers can develop a high degree of resistance to a broad range of insecticides (Keil and Parrella, 1990; Mason et al., 1987; Parrella and Trumble, 1989). In California, chemical control is often not an option for spinach. Fresh market “baby leaves” are harvested at ≈24 d after planting (at the four- to five-leaf stage) with “junior leaves” harvested 4 d later. Many systemic insecticides for larval control have a requirement of 14-d preharvest spray interval (the period with no chemical sprays before harvest). This means that fields have to be sprayed ≈10 d after planting or earlier when plants are still small and most of the sprays hits the ground and is wasted. Consequently, some growers try to avoid the leafminers by planting spinach in fields where the insect pressure is low, but growers may not have that option and leafminer infestation is often unpredictable. As a result, spinach is often tainted with the stipples of adult feeding or tunnels (mines) from larva feeding and has reduced quality, appearance, and value. Therefore, it is essential to develop alternative management strategies for leafminers.

Resistant varieties remain the most economical means of insect control. Their use may reduce the costs of chemicals, energy, and labor associated with pesticide spray and minimize potential adverse effects of pesticide use. However, commercial spinach cultivars with high levels of resistance to leafminers are not currently available. In our study, a wide range of genetic variation in traits related to leafminer resistance was found in spinach germplasm. Some genotypes had much lower levels of leafminer stings and mines than commercial cultivars. From these findings, genetic improvement of spinach for leafminer resistance seems feasible.