Seasonal Abundance and Phenology of Oebalus pugnax (Hemiptera: Pentatomidae) on Graminaceous Hosts in the Delta Region of Mississippi.

The rice stink bug, Oebalus pugnax (F.), is a graminaceous feeder, and the most injurious insect pest of heading rice, Oryza sativa L., in the United States. Rice growers are aware of the economic importance of host grasses in O. pugnax abundance. However, the need for increased knowledge of host sequence relative to O. pugnax abundance is vital. Densities of O. pugnax on 15 graminaceous hosts were evaluated in the central Mississippi Delta from April through August in 2011 and 2012. Two cultivated and 13 wild host grasses were sampled using a sweep net. Overall, populations of O. pugnax were lower in 2012 than in 2011. Italian ryegrass, Lolium perenne L. ssp. multiflorum (Lambert), was the main host that supported O. pugnax survival and reproduction from overwintering to early summer. Echinochloa spp., Digitaria spp., and Eriochloa spp. maintained greater populations of O. pugnax in the summer. Browntop millet, Urochloa ramosa (L.) Nguyen, and broadleaf signalgrass, U. platyphylla (Munro ex C. Wright) R. D. Webster, were important for populations of O. pugnax populations immediately prior to overwintering. Host switching was also an important factor that contributed to O. pugnax abundance. The evolution of Italian ryegrass resistance to the broad spectrum herbicide glyphosate in the central Mississippi delta has become an important component of O. pugnax population dynamics because of its increased abundance in and around agricultural areas. Cultural control measures on host grasses before flowering could result in less use of insecticides, thereby reducing cost of rice production.

The rice stink bug, Oebalus pugnax (F.), is a pest of cultivated crops such as wheat, Triticum aestivum L. (Lugger 1900, Forbes 1905, Sailer 1944); rice, Oryza sativa L. (Riley 1882, Webb 1920); corn, Zea mays L. (Forbes 1905, Odglen and Warren 1962); millet, Panicum miliaceum L., (Garman 1891); grain sorghum, Sorghum bicolor (L.) Moench, (Dahm 1942, Hall and Teetes 1981); barley, Hordeum vulgare L.; oats, Avena spp.; and rye, Secale cereale L. (Odglen and Warren 1962). O. pugnax is the most injurious insect pest of heading rice in all rice producing states of the United States (Douglas 1939, Douglas and Ingram 1942, Way 2003), except California (Gianessi 2009). O. pugnax relies on a broad range of graminaceous host species for feeding and reproduction throughout the year. These hosts allow the buildup of populations that eventually migrate into rice (Odglen and Warren 1962). The abundance of these host species can influence O. pugnax population dynamics (Velasco and Walter 1993). Although Douglas and Ingram (1942) and Odglen and Warren (1962) documented host grass species of O. pugnax, research on the abundance and distribution of O. pugnax relative to the host phenology of host grass species is limited. Additionally, some of the more important host species have recently developed resistance to glyphosate (Roundup, Monsanto Co., St. Louis, MO), a broad spectrum herbicide that is used to manage multiple weed species in agriculture (Orleff et al. 2009). Unfortunately, these host grass species are in close proximity to rice fields, a source of concern among rice growers.

Italian ryegrass, Lolium perenne L. ssp. multiflorum (Lambert); barnyardgrass, Echinochloa crus-galli (L.) P. Beauv.; and Johnsongrass, Sorghum halepense (L.) Pers., have all been reported to have developed herbicide resistance in Mississippi (Allen et al. 1995, Bond and Eubank 2012). As a result, host plant species have become more abundant than they were in the past (Orloff et al. 2009) and O. pugnax injury to rice can be more severe under this situation (Awuni 2013). Host grasses have, therefore, become an important component of the population dynamics of O. pugnax prior to rice heading. In particular, Italian ryegrass is currently widespread across the agricultural landscape in many of the rice producing areas of the southern United States because it is no longer managed with glyphosate applications during the early spring.

In Mississippi, Italian ryegrass and wheat, Triticum aesticum L., are the most common spring hosts when O. pugnax emerge from overwintering. Italian ryegrass emerges during the fall, begins to flower during early spring, and does not senesce until early summer (Bond and Eubank 2012). During the early summer, O. pugnax will migrate from spring hosts to summer annuals. Summer annuals that have been documented as hosts of O. pugnax include: barnyardgrass, E. crus-galli (L.) Link; Johnsongrass, Sorghum halepense (L.) Link; dallisgrass, Paspalum dilatatum Poir.; and bahiagrass, Paspalum notatum Flueggé (Douglas 1939, Douglas and Ingram 1942). The abundance of these host grass species during summer facilitates dispersal of O. pugnax into rice fields, causing widespread infestations (Douglas 1939, Douglas and Ingram 1942, Swanson and Newson 1962). Previous research has cited rice as the most preferred host for nymphal development and adult reproduction (Naresh and Smith 1984). O. pugnax will, therefore, abandon all other host grass species to feed on rice (Odglen and Warren 1962, McPherson and McPherson 2000). O. pugnax has a unique lifecycle (Nilakhe 1976) facilitated by overwintering conditions, the availability and succession of numerous cultivated and noncultivated host grass species (Panizzi 1997, Naresh and Smith 1984). Rice growers generally depend on insecticide sprays for O. pugnax management, but improper timing of applications and new immigration into fields could result in poor yields and increased production costs. Therefore, alternative control strategies such as alternate host plant management could play an important role in mitigating the impact of O. pugnax populations in rice fields. Relatively, little work has been conducted to determine the importance of host grasses on the population growth of O. pugnax before dispersal into rice fields. Most of the work on O. pugnax host grass relationships simply reported the grass species that host O. pugnax without relating their relevance to O. pugnax population dynamics (Douglas and Ingram 1942, Odglen and Warren 1962, Panizzi 1997).

The goal of this study was to examine the role of host grass species (cultivated and uncultivated) in the population dynamics of O. pugnax in the Mississippi Delta. Specifically, the study was designed to estimate the relative abundance of O. pugnax on the principal host grass species and identify the most important host grass species that contribute to O. pugnax dispersal into Mississippi rice production fields.

Materials and Methods

Study Area and Sampling Procedure

O. pugnax was monitored with a standard sweep net (38 cm) on 15 graminaceous hosts in Washington, Bolivar, and Sunflower counties in the central Delta of Mississippi from 4 May until 18 August 2011, and from 12 April to 15 August 2012. O. pugnax nymphs and adults were counted at sampling sites. Host grass species were sampled at weekly intervals from 9:30 to 11:30 am or 2:30 to 4:30 pm, based on host availability. Ten samples of 10 consecutive sweeps were taken on each potential host during each week when it was at a maturity suitable for stink bugs. There were three people that conducted sampling during each sampling period. Sampling pattern depended on the nature and location of host grass species. S. halepense was the only host grass species that was monitored with the sweep net raised above the shoulders. All other host grass species were sampled with the sweep net below the shoulders. Captured O. pugnax were sorted, counted, and placed in 29 by 29 by 29-cm rearing cages (BugDorm, BioQuip Products, Rancho Dominguez, CA) and transported to the laboratory for further studies. Sampling was initiated when host grasses began reproductive development (panicle emergence), the stage suitable for O. pugnax, and was discontinued when ∼70% of host grasses had senesced.

Host Grass Species and Habitats

Two cultivated crops, T. aestivum and O. sativa, and 13 uncultivated host grass species were identified and monitored at weekly intervals. Host grass species at various reproductive stages (flowering to maturity) were targeted for sampling. This was determined by the physical presence of inflorescences and/or seed forming structures at the time of sampling. Specific sampling locations were selected at random during times when hosts were at a growth stage suitable for O. pugnax sampling. Habitats sampled included roadsides, pasturelands, drainage ditches, and margins of cultivated fields. A habitat was sampled with 10 sets of 10 sweeps. Specimens of each host grass species were collected and transported to the laboratory for identification. Grass specimens were identified, as described by Bryson and DeFelice (2009) and with approval from a resident weed scientist. The relatively short period of panicle development, coupled with chemical and mechanical management of host grasses among some of the sampling locations impacted continued sampling of individual patches over multiple weeks. However, a particular sampling area was determined to be within a 1-km diameter of a central location when sampling was disrupted by management.

Data Analysis

Host grass species were categorized according to genera and analysis conducted. A comprehensive analysis comparing all factors and their interactions could not be conducted because of differences in time of host maturity and number of samples conducted per host genus during each week and year. Additionally, weather conditions such as rainfall and temperature were not the same between years (Table 1). As a result, separate analyses were conducted to determine the impacts of year and month. In the first analysis, the numbers of O. pugnax adults and nymphs per 10 sweeps in 2011 were compared with 2012 averaged across all weeks and hosts. Year was considered a fixed effect in the model and sample by year interaction included as random effect. The second analysis compared mean numbers of O. pugnax for each month within each year. Month was considered a fixed effect in the model and the sample by month interaction included as a random effect. Finally, an additional analysis compared O. pugnax among genera within each Julian week and year. Genus was considered a fixed effect in the model and the sample by week interaction was considered random. The means and SEs were evaluated from analysis of the raw data and the mean separation statistics were evaluated based on analysis of square root transformed data, and reported based on back-transformed means. All analyses were conducted with analysis of variance PROC MIXED (version 9.3; Littell et al. 1996). Degrees of freedom were estimated using the Kenward-Roger method. Means and SEs were calculated using the PROC MEANS statement and means were separated based on the LSMEANS using Fisher’s protected least significant difference with α = 0.05.

Table 1.

Monthly temperature, relative humidity, and rainfall data during sampling period (2011–2012)

Year	Month	Temperature max (oC)	Temperature min. (oC)	RH (max) (%)	RH (min.) (%)	Rainfall (mm)
2011	April	26.83	14.06	91.13	39.93	160.27
	May	28.30	16.83	90.16	41.55	70.00
	June	34.89	23.04	89.60	41.00	40.13
	July	35.41	23.73	93.16	45.35	49.78
	Aug.	35.38	23.04	93.03	40.68	61.21
	Mean	32.18	20.16	91.43	41.72	381.52
2012	April	25.82	13.82	91.80	41.50	106.43
	May	30.88	19.03	90.63	36.87	51.56
	June	31.71	20.35	91.00	41.52	162.31
	July	33.92	23.48	94.32	50.60	116.08
	Aug.	33.78	21.11	94.15	38.57	108.97
	Mean	31.25	19.59	92.37	41.86	545.34

Source: Delta Research and Extension (DREC) weather station.

Results

Sequence of O. pugnax on Graminaceous Host Species

There were 15 host grass species (Table 2) in 10 genera that supported the survival, development, and reproduction of O. pugnax throughout the sampling periods of 2011 and 2012. The importance of a host was determined by O. pugnax density within the week of sampling. Based on this study, L. perenne spp. multiflorum and T. aestivum were the two most important host grass species that supported O. pugnax from winter emergence. L. multiflorum and T. aestivum supported O. pugnax populations beginning during week 17 of 2011 and week 15 of 2012 (Table 2). These two genera accounted for 16% in 2011 and 8% in 2012 of the number samples conducted. Digitaria spp. and Echinochloa colona (L.) Link were the most common and important host grass species during the summer of both years. E. crus-galli was limited and was listed as E. colona. Digitaria spp. and E. colona genera together maintained O. pugnax populations longer into the season, and accounted ∼32% of the samples of O. pugnax each during both years. Digitaria spp. and E. colona supported O. pugnax for 12 wk in 2011 and 14 wk in 2012, respectively (Tables 2). E. colona and E. crus-galli both supported O. pugnax for 11 wk in 2011 and 15 wk in 2012 (Table 2). There were two host grass species (Eriochloa acuminata (J. Presl) Kunth and Eriochloa contracta Hitchc.) represented in Eriochloa in both years. These species together maintained O. pugnax for a period of 11 wk each in both years. S. halepense, Urochloa platyphylla (Munro ex C. Wright) R. D. Webster, and Paspalum dilatatum Poir were also important hosts in both years. All other host grass species maintained O. pugnax populations for ≤5 wk in both years (Table 2). Most of the host grass species maintained O. pugnax populations longer into the season in 2012 than in 2011 (Table 2). Urochloa texana (Buckl.) R. D. Webster was not found during 2012.

Table 2.

Host grass species, Julian week sampled, and number of samples on host grass species during sampling period in and around the Delta Research and Extension Center (DREC), Stoneville, MS, in 2011 and 2012

Host grass species	2011			2012
	Start week	End week	No of weeks	Start week	End week	No. of weeks
Wheat, Triticum aestivum L.	17	20	4	15	18	4
Ryegrass, Lolium perenne L. ssp. multiflorum (Lam.) Husnot	17	23	7	15	18	4
Crabgrass, Digitaria spp. Haller	18	29	12	18	31	14
Johnsongrass, Sorghum halepense (L.) Link	20	26	7	18	27	11
Prairie cupgrass, Eriochloa contracta Hitchc.	21	31	11	21	32	12
Southwestern cupgrass, Eriochloa acuminata (J. Presl) Kunth	21	31	11	21	32	12
Junglerice, Echinochloa colona (L.) Link	22	32	11	18	32	15
Bahiagrass, Paspalum notatum Flugge	23	28	6	18	30	13
Dallisgrass, Paspalum dilatatum Poir	23	28	6	18	30	13
Yellow foxtail, Setaria pumila (Poir) Roem & Schult.	25	27	3	22	29	8
Texas millet, Urochloa texana (Buckl.) R. D. Webster	25	33	9	NA	NA	NA
Browntop millet, Urochloa ramosa (L.) Nguyen	25	33	9	23	32	10
Broadleaf signalgrass, Urochloa platyphylla (Munro ex C. Wright) R. D. Webster	25	33	9	23	32	10
Rice, Oryza sativa L.	26	27	2	26	27	2
			72			92

Mean Annual and Monthly Abundance of O. pugnax

There were no significant difference between mean monthly densities of adult O. pugnax per 10 sweeps across hosts in 2011 (F = 1.38; df = 3, 716; P = 0.25; Table 3). However, more nymphs were collected during August 2011 than the other months of 2011 (F = 358.7; df = 1, 1628; P < 0.01; Table 3). In 2012, the mean monthly densities of O. pugnax per 10 sweeps collected across host grass species increased significantly as the season progressed for adults (F = 37.01; df = 4, 905; P < 0.01) and nymphs (F = 68.86; df = 4, 904; P < 0.01; Table 3). Densities of O. pugnax adults and nymphs were greater during August 2012 compared with any other month during that year. The overall annual average densities of O. pugnax across host grass species during the study period from all counties recorded significantly more adults (F = 358.7; df = 1, 1628; P < 0.01) and nymphs (F = 115.1; df = 1, 1627; P < 0.01) per 10 sweeps in 2011 compared with 2012 (Table 3).

Table 3.

LS Mean (SEM) monthly and annual densities of O. pugnax per 10 sweeps collected around Washington, Sunflower, and Bolivar Counties 2011 and 2012

Month	O. pugnax
	Adults per 10 sweepsa		Nymphs per 10 sweepsa
	2011	2012	2011	2012
April	–	0.6 ± 0.56d	–	0.0 ± 0.36d
May	16.8 ± 1.54a	1.7 ± 0.23d	2.8 ± 0.43c	0.8 ± 0.15c
June	15.3 ± 1.17a	2.3 ± 0.22c	3.9 ± 0.33b	1.0 ± 0.14c
July	13.1 ± 1.13a	4.0 ± 0.20b	3.2 ± 0.32bc	2.1 ± 0.16b
Aug.	14.5 ± 2.03a	5.3 ± 0.27a	7.0 ± 0.57a	4.1 ± 0.17a
Annualb	14.7 ± 0.68A	3.9 ± 0.20B	4.0 ± 0.13A	1.7 ± 0.09B

Means within the same column followed by the same lower case letter are not significantly different (α = 0.05).

Annual means within a row for each O. pugnax stage with the same upper case letter are not significantly different (α = 0.05).

Mean Weekly Abundance of O. pugnax on Graminaceous Hosts

2011

Significant differences in O. pugnax density were observed among host genera during 14 of the 17 wk sampled (Table 4). Lolium and Triticum were the first hosts to have adult O. pugnax. Lolium sustained the greatest densities of adult O. pugnax until week 19, when Digitaria had similar densities of O. pugnax (Table 4). Digitaria and Sorghum recorded the highest densities of O. pugnax during week 20. The first O. pugnax on Eriochloa were collected on week 21, and this was the highest density recorded throughout 2011. O. pugnax densities declined on Eriochloa during the next several weeks. High densities of O. pugnax were observed on Digitaria during weeks 22 to 24 (Table 4). Digitara and Echinochloa sustained O. pugnax longer than any other host. When O. pugnax were found in O. sativa during week 27, densities in rice were higher than in all other hosts. For the rest of the season, adult O. pugnax were collected primarily from Echinochloa and Urochloa (Table 4). The population pattern of O. pugnax nymphs was similar to that of the adults (Table 5). Lolium and Triticum were spring hosts. Digitaria, Echinochloa, and Eriochloa were major summer hosts. Echinochloa and Urochloa supported populations later into the season than all other hosts in 2011 (Tables 4 and 5).

Table 4.

Julian seasonal week and O. pugnax adult numbers (mean ± SEM) relative to host availability per 10 sweeps with a 38-cm-diameter sweep net sampled on 10 genera of host grass species in 2011

Julian Week	Host grass genera
	TRITM	LOLM	DIGTA	SOGM	ERIOLA	ECHIN	PASLM	SETAR	UROCH	ORYZA	P > F
17 (Spring)	3.3 ± 0.47b	8.6 ± 1.59a	–	–	–	–	–	–	–	–	<0.01
18 (Spring)	1.9 ± 0.43c	14.5 ± 1.59a	9.5 ± 1.75b	–	–	–	–	–	–	–	<0.01
19 (Spring)	0.0 ± 0.0b	8.7 ± 2.03a	8.2 ± 1.74a	–	–	–	–	–	–	–	=0.05
20 (Spring)	0.0 ± 0.0c	6.2 ± 0.50b	16.3 ± 3.57ab	24.7 ± 9.34a	–	–	–	–	–	–	=0.04
21 (Spring)	–	4.9 ± 0.50c	4.6 ± 0.50c	41.0 ± 7.51b	83.2 ± 14.2a	–	–	–	–	–	<0.01
22 (Spring)	–	1.1 ± 0.31d	32.2 ± 6.09ab	20.2 ± 1.94b	40.8 ± 9.51a	11.4 ± 1.64c	–	–	–	–	<0.01
23 (Spring)	–	0.1 ± 0.0c	33.5 ± 5.11a	1.3 ± 0.50c	5.5 ± 0.84b	27.6 ± 3.87a	9.6 ± 2.33b	–	–	–	<0.01
24 (Spring)	–	–	23.6 ± 2.08a	2.0 ± 0.33d	3.0 ± 0.39d	16.1 ± 1.58b	12.2 ± 2.75c	–	–	–	<0.01
25 (Summer)	–	–	12.8 ± 2.06b	0.0 ± 0.0d	6.3 ± 0.81c	9.4 ± 1.18bc	8.7 ± 1.13bc	46.1 ± 7.17a	1.8 ± 0.39d	–	<0.01
26 (Summer)	–	–	1.9 ± 0.53bc	0.0 ± 0.0d	4.1 ± 0.91b	20.5 ± 3.08a	0.0 ± 0.0d	13.6 ± 1.85a	1.1 ± 0.31c	–	<0.01
27 (Summer)	–	–	6.0 ± 0.83bc	–	2.8 ± 0.44c	13.1 ± 1.52b	0.0 ± 0.0d	3.0 ± 0.39c	0.0 ± 0.0d	45.0 ± 7.08a	<0.01
28 (Summer)	–	–	3.1 ± 0.53a	–	2.8 ± 0.39a	3.9 ± 0.79a	3.4 ± 0.21a	–	4.6 ± 0.05a	2.3 ± 0.78a	= 0.07
29 (Summer)	–	–	0.0 ± 0.0d	–	1.1 ± 0.31c	12.0 ± 1.78b	–	–	25.4 ± 2.00a	–	<0.01
30 (Summer)	–	–	–	–	5.8 ± 0.44a	6.0 ± 0.82a	–	–	7.2 ± 0.76a	–	=0.31
31 (Summer)	–	–	–	–	5.5 ± 0.84c	7.3 ± 1.13b	–	–	11.6 ± 1.25a	–	<0.04
32 (Summer)	–	–	–	–	–	8.1 ± 1.09b	–	–	15.6 ± 1.85a	–	<0.01
33 (Summer)	–	–	–	–	–	–	–	–	23.3 ± 2.84	–	NA

Means within a week followed by the same lower case letter(s) are not significantly different (α = 0.05). Means and SEs based on back-transformed data. Statistical analysis based on square root-transformed data.

TRITM, wheat, Triticum aestivum; LOLM, ryegrass, Lolium perenne L. ssp. multiflorum; DIGTA, crabgrass, Digitaria spp.; SOGM, Johnsongrass, Sorghum halepense; ERIOLA, southwestern cupgrass, Eriochloa acuminata, and Prairie cupgrass, Eriochloa contracta; ECHIN, junglerice, Echinochloa colona, and Echinochloa crus-galli; PASLM, bahiagrass, Paspalum notatum, and dallisgrass, Paspalum dilatatum; SETAR, yellow foxtail, Setaria pumila. UROCH, Texas millet, Urochloa texana, browntop millet, Urochloa ramosa Nguyen, and broadleaf signalgrass, Urochloa platyphylla; ORYZA, rice, Oryza sativa.

Dash (–), Not sampled.

Table 5.

Julian seasonal week and O. pugnax nymph numbers (mean ± SEM) relative to host availability per 10 sweeps with a 38-cm-diameter sweep net sampled on ten genera of host grass species in 2011.

Julian week	Host grass genera
	TRITM	LOLM	DIGTA	SOGM	ERIOLA	ECHIN	PASLM	SETAR	UROCH	ORYZA	P > F
17 (Spring)	0.4 ± 0.22a	0.1 ± 0.10a	–	–	–	–	–	–	–	–	=0.23
18 (Spring)	1.9 ± 0.62a	1.0 ± 0.33a	1.2 ± 0.36a	–	–	–	–	–	–	–	=0.35
19 (Spring)	0.0 ± 0.0c	1.3 ± 0.30b	4.0 ± 0.56a	–	–	–	–	–	–	–	<0.01
20 (Spring)	0.0 ± 0.0c	4.0 ± 0.83a	0.4 ± 0.22b	0.0 ± 0.0c	–	–	–	–	–	–	<0.01
21 (Spring)	–	1.7 ± 0.55bc	2.2 ± 0.42b	0.4 ± 0.22c	20.3 ± 3.39a	–	–	–	–	–	<0.01
22 (Spring)	–	0.0 ± 0.0d	5.8 ± 1.19b	1.5 ± 0.60c	21.2 ± 4.48a	1.4 ± 0.26c	0.0 ± 0.0d	–	–	–	<0.01
23 (Spring)	–	0.0 ± 0.0c	2.8 ± 0.51a	0.6 ± 0.50b	2.5 ± 0.50a	2.7 ± 0.69a	3.0 ± 0.59a	–	–	–	<0.01
24 (Spring)	–	–	12.0 ± 2.27a	0.4 ± 0.16d	1.3 ± 0.30cd	5.2 ± 0.95b	2.5 ± 0.48c	–	–	–	<0.01
25 (Summer)	–	–	4.5 ± 0.56a	0.0 ± 0.0c	0.7 ± 0.15b	4.9 ± 0.62a	1.6 ± 0.34b	4.4 ± 1.00a	0.0 ± 0.10c	–	<0.01
26 (Summer)	–	–	1.0 ± 0.36b	0.0 ± 0.0c	1.5 ± 0.45b	5.4 ± 0.85a	0.0 ± 0.0c	0.4 ± 0.22b	0.4 ± 0.16b	–	<0.01
27 (Summer)	–	–	6.2 ± 1.56a	–	1.3 ± 0.28b	4.6 ± 1.14a	0.0 ± 0.0c	0.0 ± 0.0c	0.0 ± 0.0c	7.5 ± 1.52a	<0.01
28 (Summer)	–	–	2.2 ± 0.68a	–	0.0 ± 0.0c	0.3 ± 0.30b	0.0 ± 0.0c	–	0.0 ± 0.0c	0.0 ± 0.0c	<0.01
29 (Summer)	–	–	4.3 ± 0.34a	–	4.0 ± 0.56a	5.0 ± 0.73a	–	–	5.1 ± 0.66a	–	=0.93
30 (Summer)	–	–	–	–	1.1 ± 0.21b	2.3 ± 0.63a	–	–	1.1 ± 0.21b	–	=0.04
31 (Summer)	–	–	–	–	3.2 ± 0.42a	3.7 ± 0.77a	–	–	5.8 ± 0.74a	–	=0.08
32 (Summer)	–	–	–	–	–	7.2 ± 0.13a	–	–	8.6 ± 1.11a	–	=0.42
33 (Summer)	–	–	–	–	–	–	–	–	11.5 ± 1.32	–	NA

Means within a week followed by the same lower case letter(s) are not significantly different (α = 0.05). Mean and SEs based on back-transformed data. Statistical analysis based on square root-transformation data.

TRITM, wheat, Triticum aestivum; LOLM, ryegrass, Lolium perenne L. ssp. multiflorum; DIGTA, crabgrass, Digitaria spp.; SOGM, Johnsongrass, Sorghum halepense; ERIOLA, southwestern cupgrass, Eriochloa acuminata, and Prairie cupgrass, Eriochloa contracta; ECHIN, junglerice, Echinochloa colona, and Echinochloa crus-galli; PASLM, bahiagrass, Paspalum notatum, and dallisgrass, Paspalum dilatatum; SETAR, yellow foxtail, Setaria pumila; UROCH, Texas millet, Urochloa texana, browntop millet, Urochloa ramosa Nguyen, and broadleaf signalgrass, Urochloa platyphylla; ORYZA, rice, Oryza sativa.

Dash (–), Not sampled.

2012

Significant differences in O. pugnax density were observed among host genera during 10 of the 18 wk sampled (Table 6). Similar to 2011, Lolium and Triticum were the first host genera to have adult O. pugnax. There were no significant differences between O. pugnax densities on Lolium and Triticum during weeks 15 and 17. Lolium, however, continued to sustain adult O. pugnax into week 18, when Digitaria began to support O. pugnax (Table 6). O. pugnax were found on Digitaria, Echinochloa, Paspalum, and Sorghum from week 18 through to week 24. However, Digitaria and Echinochloa continued to support O. pugnax densities until weeks 30 and 32, respectively. O. pugnax were collected from Oryza only during week 29. O. pugnax density was greater on Oryza than any other host that week. For the remainder of the growing season, adult O. pugnax were collected primarily from Echinochloa and Urochloa (Table 6). As in 2011, host grasses sustained O. pugnax nymphs in a similar pattern to that observed with adults (Table 7).

Table 6.

Julian week and O. pugnax adult numbers (mean ± SEM) relative to host availability per 10 sweeps with a 38-cm-diameter sweep net sampled on 10 genera of host grass species in 2012

Julian week	Host grass genera
	TRITM	LOLM	DIGTA	SOGM	ECHIN	PASLM	ERIOLA	SETAR	UROCH	ORYZA	P > F
15 (Spring)	0.1 ± 0.0a	0.4 ± 0.22a	–	–	–	–	–	–	–	–	=0.09
16 (Spring)	0.0 ± 0.0	0.0 ± 0.0	–	–	–	–	–	–	–	–	NA
17 (Spring)	0.8 ± 0.29a	1.3 ± 0.42a	–	–	–	–	–	–	–	–	=0.34
18 (Spring)	0.0 ± 0.0d	1.2 ± 0.27ab	1.9 ± 0.48a	2.1 ± 0.40a	1.7 ± 0.34ab	0.7 ± 0.26bc	–	–	–	–	<0.01
19 (Spring)	–	–	2.9 ± 0.64a	3.1 ± 1.14a	2.4 ± 0.78a	2.7 ± 0.79a	–	–	–	–	=0.95
20 (Spring)	–	–	1.5 ± 0.54a	1.3 ± 0.50a	2.4 ± 0.60a	1.8 ± 0.40a	–	–	–	–	=0.53
21 (Spring)	–	–	1.6 ± 0.54ab	2.0 ± 0.33a	2.0 ± 0.47a	0.8 ± 0.21bc	0.5 ± 0.22c	–	–	–	<0.01
22 (Spring)	–	–	0.0 ± 0.0	0.0 ± 0.0	4.8 ± 1.25a	0.7 ± 0.33b	4.3 ± 1.53a	3.7 ± 1.25a	–	–	<0.03
23 (Spring)	–	–	2.6 ± 0.47b	3.5 ± 0.73ab	3.6 ± 0.47ab	2.0 ± 0.52b	4.6 ± 0.93a	0.0 ± 0.0c	1.8 ± 0.39b	–	=0.05
24 (Spring)	–	–	1.7 ± 0.38ab	0.7 ± 0.31c	1.6 ± 0.32ab	1.0 ± 0.29a	0.0 ± 0.0d	0.0 ± 0.0d	2.2 ± 0.33a	–	<0.05
25 (Summer)	–	–	1.8 ± 0.49a	0.0 ± 0.0c	0.8 ± 0.29a	0.0 ± 0.0c	0.0 ± 0.0c	0.6 ± 0.31a	0.0 ± 0.0b	–	=0.07
26 (Summer)	–	–	2.2 ± 0.85a	0.0 ± 0.0c	0.0 ± 0.0c	0.6 ± 0.27b	–	0.0 ± 0.0c	1.3 ± 0.30ab	–	=0.13
27 (Summer)	–	–	0.5 ± 0.17cd	0.3 ± 0.21d	1.8 ± 0.47b	0.0 ± 0.0d	0.8 ± 0.20bc	0.0 ± 0.0	6.4 ± 0.82a	–	<0.01
28 (Summer)	–	–	0.0 ± 0.0c	0.0 ± 0.0c	2.5 ± 0.44b	–	3.7 ± 0.56ab	2.9 ± 0.91b	6.2 ± 1.52a	–	=0.02
29 (Summer)	–	–	4.9 ± 0.59b	–	1.4 ± 0.37c	0.0 ± 0.0d	6.7 ± 0.76b	0.0 ± 0.0	6.1 ± 0.80b	19.4 ± 1.96a	<0.01
30 (Summer)	–	–	1.8 ± 0.51b	–	0.0 ± 0.0c	2.2 ± 0.63b	3.3 ± 0.53b	–	6.6 ± 1.36a	0.0 ± 0.0c	<0.01
31 (Summer)	–	–	0.0 ± 0.0c	–	1.0 ± 0.33b	0.0 ± 0.0c	7.0 ± 0.83a	–	8.0 ± 0.97a	–	<0.01
32 (Summer)	–	–	–	–	3.7 ± 0.83a	–	4.4 ± 0.65a	–	4.5 ± 0.86a	–	=0.74

Means within a week followed by the same lower case letter(s) are not significantly different (α = 0.05). Means and SEs based on back-transformed data. Statistical analysis based on square root-transformation data.

TRITM, wheat, Triticum aestivum; LOLM, ryegrass, Lolium perenne L. ssp. multiflorum; DIGTA, crabgrass, Digitaria spp.; SOGM, Johnsongrass, Sorghum halepense; ECHIN, junglerice, Echinochloa colona, and Echinochloa crus-galli; PASLM, bahiagrass, Paspalum notatum, and dallisgrass, Paspalum dilatatum; ERIOLA, southwestern cupgrass, Eriochloa acuminata, and Prairie cupgrass, Eriochloa contracta; SETAR, yellow foxtail, Setaria pumila; UROCH, Texas millet, Urochloa texana, browntop millet, Urochloa ramosa Nguyen, and broadleaf signalgrass, Urochloa platyphylla; ORYZA, rice, Oryza sativa.

Dash (–), Not sampled.

Table 7.

Julian week and O. pugnax nymph numbers (mean ± SEM) relative to host availability per 10 sweeps with a 38-cm-diameter sweep net sampled on 10 genera of host grass species in 2012

Julian week	Host grass genera
	TRITM	LOLM	DIGTA	SOGM	ECHIN	PASPM	ERIOLA	SETAR	UROCH	ORYZA	P > F
15 (Spring)	0.0 ± 0.0	0.0 ± 0.0	–	–	–	–	–	–	–	–	NA
16 (Spring)	0.0 ± 0.0	0.0 ± 0.0	–	–	–	–	–	–	–	–	NA
17 (Spring)	0.0 ± 0.0	0.0 ± 0.0	–	–	–	–	–	–	–	–	NA
18 (Spring)	0.0 ± 0.0	0.8 ± 0.15ab	1.4 ± 0.42a	0.4 ± 0.11bc	0.8 ± 0.17ab	0.4 ± 0.16bc	–	–	–	–	<0.01
19 (Spring)	–	–	1.5 ± 0.40a	0.7 ± 0.33a	1.4 ± 0.40a	0.8 ± 0.33a	–	–	–	–	=0.31
20 (Spring)	–	–	1.1 ± 0.31a	0.4 ± 0.16a	1.1 ± 0.35a	0.6 ± 0.13a	–	–	–	–	=0.08
21 (Spring)	–	–	2.2 ± 0.55a	0.5 ± 0.22bc	1.5 ± 0.54ab	0.4 ± 0.15c	1.0 ± 0.45abc	–	–	–	<0.01
22 (Spring)	–	–	0.0 ± 0.0c	0.0 ± 0.0c	0.7 ± 0.30ab	0.3 0.21b	2.5 ± 1.06a	1.3 ± 0.40ab	–	–	=0.05
23 (Spring)	–	–	1.8 ± 0.48a	0.7 ± 0.26a	1.0 ± 0.37a	0.6 ± 0.22a	2.1 ± 0.46a	0.0 ± 0.0b	0.9 ± 0.41a	–	=0.08
24 (Spring)	–	–	0.6 ± 0.20a	0.3 ± 0.18a	0.4 ± 0.20a	0.7 ± 0.24a	0.0 ± 0.0b	0.0 ± 0.0b	1.1 ± 0.46a	–	=0.30
25 (Summer)	–	–	1.2 ± 0.29a	0.0 ± 0.0c	0.8 ± 0.70b	0.0 ± 0.0c	0.0 ± 0.0c	0.0 ± 0.0c	0.0 ± 0.0c	–	<0.01
26 (Summer)	–	–	0.3 ± 0.15b	0.0 ± 0.0c	0.0 ± 0.0c	0.3 ± 0.21b	–	0.0 ± 0.0c	2.1 ± 0.57a	–	<0.01
27 (Summer)	–	–	0.2 ± 0.20b	0.0 ± 0.0c	1.3 ± 0.60ab	0.0 ± 0.0c	0.2 ± 0.13b	0.0 ± 0.0c	2.1 ± 1.07a	–	=0.02
28 (Summer)	–	–	0.0 ± 0.0d	0.0 ± 0.0d	0.9 ± 0.28c	–	3.6 ± 0.59a	0.8 ± 0.20bc	3.3 ± 0.84ab	–	<0.01
29 (Summer)	–	–	3.6 ± 0.82ab	–	1.2 ± 0.44c	0.0 ± 0.0d	6.2 ± 1.02a	0.0 ± 0.0d	3.0 ± 0.45b	0.8 ± 0.39c	<0.01
30 (Summer)	–	–	3.7 ± 0.42ab	–	0.0 ± 0.0d	1.6 ± 0.52c	3.1 ± 0.64bc	–	5.8 ± 0.66a	0.0 ± 0.0d	<0.01
31 (Summer)	–	–	0.0 ± 0.0c	–	1.5 ± 0.65b	0.0 ± 0.0c	5.4 ± 0.66a	–	4.5 ± 0.20a	–	<0.01
32 (Summer)	–	–	–	–	1.9 ± 0.53b	–	4.4 ± 0.94a	–	3.0 ± 0.47ab	–	=0.05

Means within a week followed by the same lower case letter(s) are not significantly different (α = 0.05). Means and SEs based on back-transformed data. Statistical analysis based on square root transformed data.

TRITM, wheat, Triticum aestivum; LOLM, ryegrass, Lolium perenne L. ssp. multiflorum; DIGTA, crabgrass, Digitaria spp.; SOGM, Johnsongrass, Sorghum halepense; ECHIN, junglerice, Echinochloa colona, and Echinochloa crus-galli; PASLM, bahiagrass, Paspalum notatum, and dallisgrass, Paspalum dilatatum; ERIOLA, southwestern cupgrass, Eriochloa acuminata, and Prairie cupgrass, Eriochloa contracta; SETAR, yellow foxtail, Setaria pumila. UROCH, Texas millet, Urochloa texana, browntop millet, Urochloa ramosa Nguyen, and broadleaf signalgrass, Urochloa platyphylla; ORYZA, rice, Oryza sativa.

Dash (–), Not sampled.

Discussion

The total number of O. pugnax sampled across host grass species was greater in 2011 compared with 2012. Because densities were high throughout 2011, it is likely that overwintering survival was high during the winter of 2010–2011 compared with the winter of 2011–2012. There were differences in weather conditions between the years. It could also be argued that abiotic conditions better favored O. pugnax activities in 2011 than in 2012. The average maximum and minimum temperatures in 2011 were 89.89 and 68.29°F, compared with 88.25 and 67.26°F in 2012, respectively. Temperatures were all time higher during summer in 2011 than in 2012, as indicated in Table 1. Precipitation during the study period was greater in all months for 2012 than for 2011. Constant cold conditions resulting from wet environs can result in reduced insect metabolism. Therefore, 2011 was much a favorable year for arthropods activity than in 2012. Although the relationship between insect outbreak and microclimate has not been well established, the relationship seems compelling (Haile 2000). Environmental factors, primarily extremes in temperature and precipitation, have been documented to cause recurring outbreaks in insect populations of pest species. These conditions may have accounted for the larger numbers of O. pugnax population recorded in 2011 compared with 2012.

Although the densities of O. pugnax collected from host grass species varied between years, the periods during which grasses were used as hosts were similar in both years. Most stink bugs, including O. pugnax, are polyphagous and feed on a broad range of cultivated and uncultivated host plants (Panizzi 1997, Jones and Sullivan 1982). In the current study, the general dynamics of both adult and nymph O. pugnax populations indicated that host grass species can be categorized into four groups. The first category includes Lolium and Triticum that supported O. pugnax survival, development, and reproduction early in the spring. These host grasses are winter annuals that germinate in the fall and bear fruiting structures during the early spring in both years. Densities on Triticum were lower compared with Lolium during the early spring, but both species are important in supporting O. pugnax from overwintering. L. multiflorum has become one of the most dominant host grass along roadsides, ditches, and fallow fields from late winter to early summer. This may be the result of its difficulty to control owing to resistance to the broad-spectrum herbicide glyphosate (Bond and Eubank 2012).

The second group includes Digitaria, Echinochloa, and Eriochloa genera that sustained O. pugnax populations during the summer. Digitaria is an important transitional host because it supported O. pugnax just after Lolium and Triticum and at least 2 wk earlier than the other summer annuals. E. crus-galli was rarely found, and if found was in low densities along drainage habitats and field margins. E. colona was more abundant in and around Mississippi rice production fields than E. crus-galli. Perhaps, the most important role of E. colona is its ability to effectively compete with rice for resources in rice fields, and the ability to mimic rice at the seedling stage. When not detected early in rice fields, E. colona can attract O. pugnax into rice fields even before rice panicles begin to head because they are early maturing than rice. An important role of Eriochloa was its support for nymphal development as nymphal densities on Eriochloa were often as high as or higher than on any other host.

The third group of host genera included Paspalum, Sorghum, and Setaria that sporadically hosted O. pugnax, so may not be critical factors impacting O. pugnax population dynamics. Although large numbers of O. pugnax adults were captured on S. halepense, it was the least supportive to O. pugnax nymphs throughout the 2-yr study. This may be an indication of non-suitability for nymphal development. Garman (1891) observed O. pugnax feeding on several species of Setaria and other grass spp. consistent with the current study. The fourth group includes three species of Urochloa; U. texana, U. ramose, and U. platyphylla. These were generally observed in senescing corn fields or along abandoned cropped fields. This group appeared to serve as transitional hosts for O. pugnax populations prior to overwintering. Hall and Teetes (1981) previously reported Urochloa fusca (Swartz), and U. texana as hosts of O. pugnax during the months of June and July. But, in this study, these host grasses appear to support O. pugnax populations that overwinter. Because of foliar insecticide sprays for O. pugnax control in rice fields, limited sampling was conducted in rice, but other research showed that rice is preferred for feeding (Awuni et al. 2014).

The sequence of host grass availability was important because the preferred feeding sites of O. pugnax change over time, so a succession of hosts is required for successive generations during any crop season (Panizzi 1997, Borges et al. 2011). Douglas (1939) listed seven host grass species utilized by O. pugnax in Louisiana, but noted that the list was incomplete. In a later development, Douglas and Ingram (1942) identified 10 host grass species of O. pugnax, while Odglen and Warren (1962) listed 7 cultivated and 10 uncultivated host species in Arkansas. In this study, 2 cultivated and 13 uncultivated host grass species representing 10 genera were identified as important for O. pugnax survival and reproduction in Mississippi. Admittedly, this may not be an exhaustive list of host grass species fed on by O. pugnax in Mississippi. Odglen and Warren (1962) observed that these host grasses could support two to three generations of O. pugnax before rice starts to head.

The preference of O. pugnax among host grass species remain a subject of debate. Douglas and Ingram (1942) reported vaseygrass, Paspalum urvillei Steud., as the most preferred host of O. pugnax amongst wild host grass species. In contrast, Odglen and Warren (1962) reported E. crus-galli as the most preferred host of O. pugnax amongst the uncultivated host grass species. In a related study, Awuni et al. (2014) reported E. colona as the most preferred of the 13 host grass species in a preference and suitability tested.

The current control of O. pugnax largely depends on in-crop treatment with broad-spectrum insecticide sprays that may increase costs of production and harm the environment (Sudarsono et al. 1992). Manipulation of host grass species could reduce production costs if implemented in a timely manner. Any management strategy aimed at disrupting or reducing host abundance may impact O. pugnax populations. This can be accomplished through the use of integrated pest management program, which involves multidisciplinary approach to minimizing rice field infestations and reducing treatment costs. Although Douglas (1939) argued that mowing host grasses could increase O. pugnax infestations in adjoining rice fields, Webb (1920) noted that mowing grasses around rice fields in a timely manner could reduce rice injury from O. pugnax infestation. The key is mowing host grasses before flowering grass panicle starts to head. This means host management could be an important component of O. pugnax management. Destruction of spring and early summer hosts of O. pugnax could reduce O. pugnax populations later in rice. Late-season habitats that can increase O. pugnax populations prior to overwintering could be destroyed similarly, reducing the number of O. pugnax adults that overwinter.

In conclusion, rice growers are well aware of the economic significance of O. pugnax infestation in rice and the role of host grass species in the abundance of O. pugnax. There is, however, need for improved awareness of the relationship between O. pugnax and host grasses. This is particularly important because of reported herbicide resistance of some principal host grasses such as E. colona, L. perenne, and S. halepense that are not only prevalent in the landscape, but utilized by O. pugnax in Mississippi.