Roses (Rosa spp. L.) are important ornamental hosts of chilli thrips, Scirtothrips dorsalis Hood (Thysanoptera: Thripidae). The knowledge of how these thrips affect 8 cultivars of landscape roses popular in Florida (‘Angel Face’, ‘Don Juan’, ‘Pink Summer Snow’, ‘Radeon’, ‘Radrazz’, ‘Radsunny’, ‘St. Patrick’, and ‘Sun Flare’) would help in developing techniques for integrated pest management of S. dorsalis. The effects of 3 rates of fertilizer and cultivars on population densities of chilli thrips and on host plant damage were evaluated. Fertilization rate, plant organ, and cultivar were important in determining S. dorsalis population density. Differences in total numbers of S. dorsalis, damage rating, and in numbers of flowers and buds produced were observed among different fertilizer rates. The higher rates recommended for accelerated floral growth resulted in more S. dorsalis damage, but not in more flowers than the rates suggested for maintenance. Among parts of the rose plant, buds had the highest density of S. dorsalis, followed by flowers and leaves, which had similar low densities. Larger flowers had more S. dorsalis than small flowers, but population densities were similar. Different cultivars of Knock-Out® rose were similar in their susceptibility to S. dorsalis, but ‘Radeon’, ‘Don Juan’, and ‘Sun Flare’ had more damage with lower S. dorsalis abundance and density than other the cultivars.
Scirtothrips dorsalis Hood (Thysanoptera: Thripidae), chilli thrips, is a problematic pest native to southern Asia (Dev 1964; Kumar 2012). It has rapidly expanded its range over the last 20 yr into much of the world's tropical and subtropical regions. In 2005, S. dorsalis was first found established in Florida (Coolidge 2005; Silagyi & Dixon 2006) and Texas (Holtz 2006) and shows considerable potential for expansion into the remainder of North America (Venette & Davis 2004; Meissner et al. 2005; Nietschke et al. 2008). Scirtothrips dorsalis attacks more than 112 plant species in about 40 different families (CABI/EPPO 1997; CABI 2003), and it affects fruit and vegetable hosts of economic significance including cotton (Gossypium spp. L., Malvales: Malvaceae), peanut (Arachis hypogaea L., Fabales: Fabaceae), soybean (Glycine max (L.) Merr, Fabales: Fabaceae), strawberry (Fragaria × annanasa Duchesne, Rosales: Rosaceae), pepper (Capsicum spp. L., Solanales: Solanaceae), mango (Mangifera spp. L., Sapindales: Anacardiaceae), and citrus (Citrus spp. L., Sapindales: Rutaceae) (Venette & Davis 2004; Meissner et al. 2005; Nietschke et al. 2008). Scirtothrips dorsalis also attacks ornamental plants and is a major pest of rose for the cut-flower industry in India where it reduces the quality, number, size, and appearance of flowers (Onkarappa & Mallik 1998; Duraimurugan & Jagadish 2004a). Symptoms of S. dorsalis feeding damage include stunted, scarred, deformed growth, leaf drop, and the leaves, sepals, and petals develop a bronzed, scorched scarring (Dev 1964; Mound & Palmer 1981; Chandrasekaran 2005). Some researchers suggest that a protease in the saliva may contribute to the damage (Dev 1964; Raizada 1965). Scirtothrips dorsalis destroys plant epidermal and mesophyll cells when it feeds (Heming 1993; Kirk 1995). Hence, higher densities of S. dorsalis may cause more damage than host plants can replace and can kill them (Mound & Palmer 1981), and plants damaged by S. dorsalis become unattractive, thus unsalable.
Gahukar (2003) reported that there was high variation in S. dorsalis population abundance among rose cultivars and from year to year. ‘Gladiator’had the highest abundance of S. dorsalis in flowers of different sizes and the highest infestation rates. Also, flower color affected S. dorsalis population distribution with red flowers having more S. dorsalis than orange or yellow flowers (Gahukar 2003).
The Knock-Out® rose, Rosa ‘Radrazz’ (‘Carefree Beauty’ × ‘Razzie Dazzle’), has single red flowers, and a round, shrubby growth habit (Radier 2001), and has become popular in Florida because of its abundant flowers, low maintenance, and strong resistance to black spot disease, Diplocarpon rosae Wolf (Ascomycota) in Florida's humid environment (Radier 2001, Brown 2007). Unfortunately, R. “Radrazz” has also become an important host of S. dorsalis (Silagyi & Dixon 2006), which has affected its sales in Florida nurseries. The Sunny Knock-Out®, R. ‘Radsunny’ (‘Radbrite’ × ‘Radsweet’) also has single flowers, a round, shrubby growth habit, and excellent resistance to blackspot disease. Unlike R. ‘Radrazz’, R. ‘Radsunny’ produces bright yellow flowers that fade to white as they age, and the petioles release a strong “sweetbriar” odor (Radler 2001, 2008). The susceptibility of R. ‘Radsunny’ to S. dorsalis has yet to be determined. Pink Knock-Out®, Rosa ‘Radeon’ (a mutation of ‘Radrazz’) has single, pink flowers, compact, mounding growth habit, and excellent resistance to blackspot disease (Montesino 2004). Rosa ‘Angel Face’[(‘Circus’ × ‘Lavender Pinocchio’) ×Sterling Silver’] is a lavender-flowered hybrid tea rose with an upright, bushy, compact growth habit (Wikipedia 2012). Rosa ‘Don Juan’ (‘New Dawn’ × ‘New Yorker’) is a dark-red-flowered, climbing rose (Jackson & Perkins 2012) and ‘Pink Summer Snow’ (a sport of ‘Summer Snow’) is considered a cluster-flowered type rose, which includes floribunda and grandiflora roses, has pink flowers, and is a shrub (Dave's Garden 2012a). Rosa ‘St. Patrick’ (‘Brandy’ × ‘Gold Medal’) is a hybrid tea rose with double yellow flowers and a bushy growth habit (Dave's Garden 2012b), and ‘Sun Flare’(‘Sunsprite’ × unnamed variety) is a floribunda class rose with yellow flowers, which grows into a 0.6-to-1.2-m tall shrub (Wikipedia 2010). Seven of these 8 varieties have been patented with the United States Patent and Trademark Office: all except ‘Pink Summer Snow’ (Radier 2001, 2008; Montesino 2004; Wikipedia 2010, 2012; Dave's Garden 2012a, b; Jackson & Perkins 2012; USPTO 2012).
Plants provided with high nitrogen fertilization are often more attractive to insect pests than those provided with less nitrogen (Slansky & Rodriquez 1987). Thus, a common nursery practice of adding nutrients to promote or “push” the growth of ornamental plants may be causing S. dorsalis populations to increase. Heavily fertilized plants have more free amino acids in their tissues allowing for more rapid growth of the plant and of populations of insects that feed on them (Mattson 1980). Fertilized plants may have larger plant heights and produce more and larger flowers, which provide additional attractive cues to pests. Elevating quantities of fertilizer increased the population growth rate of Frankliniella occidentalis (Pergande) (Thysanoptera: Thripidae) compared to lower fertilization rates (Chau et al. 2005). However, higher insect growth rates with higher host plant fertilization is mediated by specialized secondary compounds produced by resistant cultivars (Schuch et al. 1998) or by other factors such as season (Salguero-Navas et al. 1991; Cho et al. 2000; Reitz 2002) or host plant phenology. For example younger plants may support higher population densities than older plants (Reitz 2002). Nevertheless, Duraimurugan & Jagadish (2004b) observed larger populations of S. dorsalis with higher fertilization rates on rose, and Varghese & Giraddi (2005) suggested fertilizing chili peppers with only 50 percent of the recommended rate combined with the use of azadirachtin insecticide to control S. dorsalis.
Because S. dorsalis may thrive in both agricultural and ornamental landscapes, growers and homeowners should apply an integrated pest management program to prevent development of pesticide resistance. Knowledge of the effects of S. dorsalis on other rose varieties would be helpful in developing integrated pest management techniques. The purpose of the present study was to evaluate effects of fertilization rate and eultivar on the population density of S. dorsalis and resulting host plant damage.
MATERIALS AND METHODS
All tests were performed at the University of Florida, Tropical Research and Educational Center (TREC), Homestead, during 2008. All test plants were exposed to ambient environmental conditions throughout the experiments and irrigated with 93 ± 15 ml of tap water twice daily at 6:00 AM and 2:00 PM EST by an overhead sprinkler system with a timer and following local water restrictions and extension office suggestions. Total numbers of flowers and vegetative buds on each plant were counted biweekly to estimate the reproductive output of each plant. A damage estimate similar to that described by Kumar et al. (1996) for pepper was developed during preliminary observations and was used weekly to rate plant damage from 0 for no symptoms to 5 for defoliated plants near death. For consistency, all samples of leaves, buds, and flowers were removed with pruning shears and promptly sealed in small plastic containers with 2 drops of 95% ethanol. Samples that could not be immediately processed were frozen at -6 °C until processing; only mature leaves, buds, and flowers of a similar age were sampled. Plant samples were dried overnight and the area of each dried sample was determined using a leaf area meter (LI-3000, LiCor, Lincoln, Nebraska) to allow calculation of the density of S. dorsalis in each sample (S. dorsalis per cm2). The quantity and life stages of S. dorsalis in a sample were determined by washing samples with 75% ethanol and pouring the runoff through a 25-µm mesh as described by Seal & Baranowski (1993) and the resulting debris was examined under a dissection microscope at 6X12X magnification. Funderburk et al. (2007) was a primary aid for identification using morphological characters. Insect voucher specimens were collected and sent to the Florida Department of Agriculture and Consumer Services, Division of Plant Industry, Gainesville, for preservation and to confirm species identification throughout the experiment.
Test 1. Fertilizer Rates With ‘Radsunny’ and ‘Radrazz’ Cultivars
From Aug to Oct 2008, a comparison was made on the effects of fertilizer rates on ‘Radrazz’ and ‘Radsunny’ rose cultivars. Twenty-four bare-root rose plants each of ‘Radrazz’ and ‘Radsunny’ provided by the Conard-Pyle Company (Wilmington, Delaware) were grown in a greenhouse free from S. dorsalis. Plants were grown individually in 11-liter containers using a standard potting medium: 50% Canadian sphagnum peat moss, 25% processed pine bark, and 25% a mixture of perlite and vermiculite (Fafard 3B Mix, Conrad Fafard Inc., Agawam, Massachusetts) for 3 mo then transferred to an outdoor growing area. The plants were then exposed to “Radrazz” roses infested with 3 S. dorsalis per cm2 on the flowers. Four replications each with 1 of each treatment and cultivar were randomly placed within a 6-by8-plant grid that provided 33 cm separation between plants. Eight plants of each cultivar were randomly assigned to 1 of 3 fertilizer treatments. The “high rate” treatment involved 7.9 g of 2020–20 liquid fertilizer (Peter's Professional, Scotts Company, Marysville, Ohio) every 2 wk and 13 g of 15–9–12 granular fertilizer (Osmocote Plus, Scotts Co., Marysville, Ohio) every 2 mo. The maintenance rate treatment used 4.0 g of liquid fertilizer once a mo and 7.5 g of granular fertilizer once every 2 mo. Control plants received no solid fertilizers and only water when plants in the other treatments received liquid fertilizer. Plant samples were collected every 2 wk from 1 Sep to 22 Oct 2008. Estimates of S. dorsalis populations were made by randomly sampling a bud or flower from each plant every 2 wk at 3:00 PM EST as described by Duraimurugan & Jagadish (2004a).
Test 2. Susceptibility of 6 Cultivars
The second experiment (Sep to Nov 2008) compared populations of S. dorsalis on 6 rose cultivars ‘Angel Face,’ ‘Don Juan’, ‘Radeon’, ‘Pink Summer Snow’, ‘St. Patrick’, and ‘Sun Flare’. Five plants of each of the 6 cultivars were donated in 11-L containers by Nelson's Florida Roses, Apopka, Florida, then transported to TREC on 9 Sep 2008 and exposed to naturally occurring S. dorsalis populations. Five replications with 1 plant from each of the 6 cultivars were randomly placed within a 5-by-6-plant grid that provided 33 cm of separation between plants. One bud, 1 flower and 1 leaf were sampled from each plant at approximately 3:00 PM EST once every 2 wk to determine within-plant pest distributions based on methods of Onkarappa & Mallik (1998). Plant samples were collected on 9 Sep 2008, and then biweekly from 11 Sep to 11 Nov 2008.
A two-way ANOVA was initially used to test for interaction between cultivar and fertilizer rate in the first test and cultivar and time in the second test. For both tests, the numbers and densities of S. dorsalis were transformed by applying the square root to homogenize variance before analysis. The number and size of flowers, number and density of S. dorsalis, their proportion at each life-stage, and damage levels for each eultivar and fertilizer treatment were compared among treatments by one-way ANOVA using JMP statistical software (SAS Institute 2007). Means from the first experiment were separated using pairwise T-test comparisons, and by Tukey-Kramer HSD in the second experiment. Least-square means were also used to fit the general linear model and determine the strength of relationships between different factors in each experiment.
Test 1. Fertilizer Rates With ‘Radsunny’ and ‘Radrazz’ Cultivars
In the first experiment, no significant interactions were found between cultivar and fertilizer rate for any of the variables tested. Thus, fertilizer treatments were pooled to compare cultivars and cultivars were pooled to compare fertilizer treatments. No significant differences were found between ‘Radrazz’ and ‘Radsunny’ cultivars for abundance of S. dorsalis at any life stage or sex, densities, damage rating, floral area, or the number of flowers and buds produced (data not shown). However, both the maintenance and high rates of fertilizer resulted in a significantly higher weekly total number of S. dorsalis than the control (F = 3.46; df = 2, 232; P = 0.0330) (Fig. 1). Significant differences also occurred for weekly number of larvae (F = 3.99; df = 2, 232; P = 0.0197) and of adults (F = 3.24; df = 2, 232; P = 0.0411). Damage rating was also significantly higher for the high fertilizer rate than for the control (F = 5.07; df = 2, 232; P = 0.0070) (Fig. 2). Further analysis by week showed that differences in total numbers of S. dorsalis increased and became significant for wk 5 (F = 3.32; df = 2, 44; P = 0.0455) and 7(F = 4.53; df = 2, 44; P = 0.0162) (Fig. 3). There were no significant differences between fertilizer treatments in mean density of S. dorsalis, in total of flowers or of buds, or in mean area of flowers (data not shown).
Test 2. Susceptibility of 6 Rose Cultivars to S. dorsalis.
No significant interaction was observed between eultivar and time for S. dorsalis abundance, density, or damage. Scirtothrips dorsalis density generally increased on all hosts for the first 3–5 wk, then tended to remain constant (Fig. 4). After 5 wk, S. dorsalis was found on all plants of all rose cultivars tested. Although no significant interaction was found, eultivar (F = 3.55, df = 5, P = 0.0037) and plant organ (F = 96.09, df = 2, P < 0.0001) each appeared to significantly affect S. dorsalis densities. When weekly data were pooled, there was a significant difference in total S. dorsalis density between cultivars (F = 2.23; df = 5, 519; P = 0.0496) (Fig. 5). ‘Radeon’ numerically had the highest S. dorsalis density and was significantly higher than ‘Sun Flare’ or ‘Angel Face’. Plant organs also showed significant differences in S. dorsalis density (F = 89.14; df = 5, 519; P < 0.0001) with buds having significantly higher densities than flower petals or leaves (Fig. 6). While no plant had a damage rating greater than 1 during the experimental period, a significant difference was found in mean damage rating per cultivar (F = 5.42; df = 5, 519; P < 0.0001) with ‘Radeon’, ‘Don Juan’, and ‘Sun Flare’ each significantly higher than ‘Angel Face’ (Fig. 7).
For container production, the “high rate” fertilizer treatment was recommended by local growers to accelerate plant growth, while the “maintenance rate” was recommended to “maintain” plants. Plants fertilized at the highest rate showed greater damage than the control treatment, and this may have been partly due to larger populations of S. dorsalis found on these plants. Scirtothrips dorsalis was more abundant on hosts that were fertilized than on unfertilized hosts but its density was similar among the 3 fertilizer rates, which may have allowed S. dorsalis to exploit resources with minimum competition. With higher S. dorsalis populations and pest pressure, crowding may begin to occur, and there may be an upper limit on S. dorsalis population density or on the number that can feed simultaneously on a host plant. The recommendations by Varghese & Giraddi (2005) to reduce fertilizer use on chili by 50 percent may also help to limit S. dorsalis populations and their damage to cultivars such as R. ‘Radrazz’”. This fits with our observation of roses in urban landscapes. Populations are much lower in poorly maintained roses compared to those that are fertilized and watered regularly.
Damage to plants in both experiments did not exceed a rating of 2 on a 0-to-5 scale. New growth was wrinkled, and scarring occurred along the vein bases, most sepals, and the outer petals and most leaves were often stippled. A damage rating of 1 denoted damage to some sepals, but few petals and leaves were stippled. The limited range in feeding damage observed made it difficult to draw strong conclusions about relationships between S. dorsalis numbers or density and damage, hence, these relationships may have been better described by a different damage rating scale. Therefore, our scale helped to predict numbers of larvae and adults detected in flowers, but could not account for much of the variation present in populations. The low plant damage ratings may not be unusual for S. dorsalis behavior during the fall, when this experiment was conducted in south Florida. Also, S. dorsalis density and plant damage may increase on all cultivars during favorable conditions for the thrips, thereby causing more pest pressure and increasing differences in cultivar responses to the pest. In addition, S. dorsalis may not perform the same when multiple cultivars are present as when only 1 cultivar is present. In such a no-choice situation, some cultivars (when tested individually) may show higher levels of S. dorsalis damage, abundance, or density than other cultivars. Although a predictive relationship of S. dorsalis numbers or density to damage could not be described, the damage seemed more predictive of larval than of adult populations and of S. dorsalis numbers than density. While the ‘Radrazz’ (red-flowered) and ‘Radsunny’ (yellow-flowered) rose cultivars can be distinguished by foliar appearance, petal color, and floral scent, S. dorsalis populations seemed to be evenly distributed between both cultivars. This seems surprising because S. dorsalis has been captured in greater frequencies on yellow (Tsuchiya et al. 1995; Chu et al. 2006) or yellow and green (Tsuchiya et al. 1995) sticky traps than traps of other colors such as blue or white. However, Rani & Sridhar (2003) and Gahukar (2003) recovered more S. dorsalis adults on red or orange than on yellow rose petals in a small-arena choice-test. While these results may appear to conflict, S. dorsalis may use different cues and selection criteria to choose hosts at different spatial scales. The choice tests of Rani & Sridhar (2003) and Gahukar (2003) occurred in small arenas within the lab, while Tsuchiya et al. (1995) and Chu et al. (2006) performed choice tests in the field. Comparing these observations to results of the second experiment in the present study suggests that at a range of less than 1 meter, S. dorsalis may not choose between the 2 rose cultivars using criteria obvious to humans, such as color or scent, and other factors may be more important to local S. dorsalis dispersal and distribution.
There was no interaction between plant organ and cultivar suggesting that S. dorsalis will exploit all roses in a similar fashion, although they were found on some plant organs and on some varieties in greater densities than on others. Shibao et al. (1990) observed that populations of S. dorsalis were similar between 2 varieties of grape, but there were internal distributional differences. While S. dorsalis densities are valuable for comparing different rose species or cultivars, total numbers of S. dorsalis may be important to study population expansion and control. Larger populations are better able to persist through catastrophic ecological events and increase pressure on surrounding environments (Memmott et al. 1997; Fagan et al. 2002) especially if the larger roses have a larger floral area or produce large numbers of buds. Scirtothrips dorsalis was found in higher densities on buds than on flowers or leaves of roses. In addition, many thrips appear to prefer the upper part of the plant canopy and the outer extremities of their hosts (Reitz 2002; Hansen et al. 2003). However, removing growth terminals in citrus caused an increased migration of S. dorsalis away from de-budded plants (Shibao et al. 1993; Shibao 1997). Scirtothrips dorsalis may prefer these meristematic regions (buds) because nutrients flow towards them in plants (Lewis 1997). This is generally supported by research about S. dorsalis on Capsicum annuum L. (chili pepper) (Seal et al. 2006) and Camellia spp. L. (Ericales: Theaceae) (Dev 1964) in which thrips preferred younger plants and newer growth over older plants and growth. Also, S. dorsalis females prefer to oviposit inside buds and young leaves at apical meristems, but as populations increase, will oviposit within surfaces of mature leaves (Dev 1964; Raizada 1965). After hatching, larvae migrate from older leaves to the newer growth near apical meristems (Onkarappa & Mallik 1998). Although S. dorsalis was not present in high densities on leaves, the distinctive feeding damage of larvae provided evidence of their presence; more importantly, S. dorsalis occurred on leaves and flower petals at similar densities. Scirtothrips dorsalis is often found in flowers and terminal shoots in vegetables and ornamental plants (Shibao et al. 1993), and within flowers, it feeds on pollen (Saxena et al. 1996). In an olfactometer test, Saxena et al. (1996) showed that a greater number of S. dorsalis were attracted to pollen scents than to those of nectar or other plant tissue. However, plants are mostly non-floral leaves, stems, and buds, and most of the infestation symptoms of stunted, bronzed, deformed growth, and leaf drop are foliar (Dev 1964; Mound & Palmer 1981; Chandrasekaran 2005). Hence, S. dorsalis may be described as either a flower or foliar thrips. Scirtothrips dorsalis may thrive equally well on both substrates, and other factors such as precipitation, predation, or competition may largely determine the distribution of populations within host plants.
In the second test, ‘Radeon’, ‘Don Juan’, and ‘Sun Flare’each had a higher damage rating than ‘Angel Face’, which numerically had the lowest rating, and ‘Radeon’ also was higher in S. dorsalis density than ‘Angel Face’. ‘Radeon’ is a spontaneous mutation of “Radrazz” (Montesino 2004), thus, more closely related to “Radrazz” than to the other varieties in the test, which are hybrids or sports of other varieties. Hence, these findings support the initial belief by extension agents that R. ‘Radrazz’ is an important host to consider when surveying for S. dorsalis (Silagyi & Dixon 2006). Numerically, ‘Radeon’ had the highest S. dorsalis density and was significantly higher than ‘Sun Flare’ or ‘Angel Face’. With S. dorsalis density, the small-flowered varieties, ‘Radeon’ and ‘Pink Summer Snow’, were not different from the largeflowered varieties, ‘St. Patrick’ and ‘Don Juan’. Gahukar (2003) found that large, open flowers can support larger numbers of S. dorsalis, but smaller, more compact flowers can have a relatively high density of S. dorsalis. These hosts may provide similar environments for S. dorsalis, and vegetative size should determine total S. dorsalis population. Other varietal characteristics of roses may also affect S. dorsalis density: for example, in the second test, ‘Angel Face’ was numerically the lowest in S. dorsalis damage rating and population density among 6 rose varieties. This may represent host plant resistance preventing S. dorsalis growth, or a repellence of S. dorsalis based on the structure and chemistry of a particular eultivar and perhaps merits further testing. With this low rate of infestation and damage, ‘Angel Face’ may be the best choice among the cultivars we tested for the home landscape; similarly, ‘Radeon’ numerically had the highest population density and damage rating and may be 1 of the worst choices.
We would like to thank the Conard-Pyle Company for providing the Knock-Out® roses used in this study, and Nelson's Roses of Florida for donating the other cultivars of roses used.