The effects of relative humidity and diet moisture content level were determined for the coffee berry borer, Hypothenemus hampei Ferrari (Coleoptera: Curculionidae), when reared on an artificial diet (Cenibroca) at a constant temperature of 25 °C. Three relative humidity levels (i.e., 65, 75, and 85%) at 3 moisture content levels (i.e., 50, 60, and 70%) were evaluated. This artificial diet has been used routinely for over 5 yr to rear coffee berry borer. The following biological parameters were measured for each treatment diet, i.e., preoviposition period, oviposition and feeding behavior, total progeny production, and reproductive potential. The highest reproductive rate and intrinsic rate of increase were obtained with Cenibroca diets containing 50% moisture content level at 85% relative humidity, and 60% moisture content level at 75% relative humidity. The lowest reproductive and intrinsic rate of increase occurred using a diet containing 70% moisture content level at 85% relative humidity.
Se examinó el efecto de la humedad relativa y niveles de contenido de humedad en la broca del café, Hypothenemus hampei Ferrari (Coleoptera: Curculionidae) cuando se reprodujo en una dieta artificial (Cenibroca) a una temperatura constante de 25 °C. Se evaluó la respuesta de la broca del café bajo 3 niveles de humedades relativas (i.e., 65, 75, y 85%) y 3 niveles de contenido de humedad (i.e., 50, 60, y 70%). Esta dieta artificial se ha venido utilizando rutinariamente para reproducir la broca el café por generaciones sucesivas por más de 5 años. En cada tratamiento se estimaron los siguientes parámetros biológicos, i.e., período de pre-oviposición, comportamiento de oviposición y alimentación, progenie total por hembra y potencial reproductivo. Los valores mas altos de la tasa reproductiva y la tasa intrínseca de crecimiento se obtuvieron en los tratamientos con la dieta Cenibroca que contenía 50% niveles de contenido de humedad a 85% humedad relativa, y dieta con un 60% niveles de contenido de humedad a 75% humedad relativa. Los valores mas bajos de la tasa reproductiva se encontraron en el tratamiento 70% niveles de contenido de humedad a 85% humedad relativa.
The coffee berry borer, Hypothenemus hampei Ferrari (Coleoptera: Curculionidae), is a major insect pest of coffee, Coffea arabica L. (Rubiaceae), throughout the world. This species is native to Africa and has spread to the Middle East and Asia, as well as the Central, South, and North American continents (Hargreaves 1926; Bergamin 1943; Le Pelley 1968; Baker 1984; Portilla et al. 2000, 2014; Aristizabal et al. 2017). The most recent introduction was reported in Oceania (Papua New Guinea) in May 2017 (Portilla et al. 2021). The borer has caused substantial production losses where it occurs. Losses are due essentially to berry weight reduction upon harvesting, as well as shedding of borer-damaged fruits (Le Pelley 1968; Reid & Mansingh 1985; Smith & Bellotti 1996; Borbon 2004; Ruiz-Cardenas & Baker 2010; Aristizabal et al. 2016). In addition to the above economic loss, increased labor costs associated with sorting product and depreciation of infested marketable products have been reported (Waterhouse 1998; Aristizabal et al. 2017). Currently, coffee berry borer is a problem for US coffee growers (e.g., Puerto Rico and Hawaii), and adoption of an integrated pest management program to control this pest is crucial. Aristizabal et al. (2017) mentioned that during 2011 to 2013, coffee berry borer in Hawaii substantially increased the cost of coffee production by 10 to 15%.
The coffee berry borer resides within coffee berries where it is relatively well-protected from pesticides. Management has principally relied upon cultural and biological control strategies (Bustillo et al. 1996; Aristizabal et al. 2017). Fertile coffee berry borer females directly damage the fruit by tunneling into coffee berries to feed and lay eggs. Pre-oviposition lasts from 4 to 20 d where the females oviposit batches of 2 to 3 eggs daily over a 20-d period (Abraham et al. 1990). Egg hatch occurs in 7.6 d on average with 13.8 d for larval development, 2.0 d for prepupae, and 6.4 d for pupation at 27 °C (Le Pelley 1968). Development from egg to adult averages 27.5 d at 24.5 °C with adult longevity ranging from 28 d (Baker et al. 1994) to 282 d (Bergamin 1943) on coffee berries, 25 to 105 d on parchment coffee, and 38 to 325 d on Cenibroca artificial diet (Portilla & Streett 2008). The ratio of females to males is variable, but females always predominate with ratios of 10 females to 1 male under field conditions (Le Pelley 1968). Typically, adult males cannot fly but remain in the coffee bean for sibling mating (Waterhouse 1998).
Although the coffee berry borer is 1 of the most extensively studied tropical insects, there are only a few reports on abiotic factors that affect mass-production under laboratory conditions. Baker et al. (1994) studied coffee berry borer emergence from infested coffee berries under different temperature and relative humidity regimens. Coffee berry borer emergence from coffee berries held individually was found to increase steadily between 90 to 100% relative humidity at 20 °C to 25 °C. Few coffee berry borers were found emerging at temperatures < 20 °C with no significant increase in adult emergence between 25 °C to 30 °C. Apparently, coffee berry borers can exit berries under humidity < 50%, which was attributed to an escape response by the coffee berry borer seeking higher relative humidity levels (Baker et al. 1994). This behavior may be a sub-social task such as maternal sanitation as mentioned by Vega et al. (2016). Baker et al. (1994) reported that female coffee berry borer survived < 6 d at < 50% humidity levels, and adult survival was greatest at 20 °C with a mean survival time of almost 28 d at 93.5% relative humidity. Adults reared under this regime produced an average of 6.2 and 5.1 individuals per d in infested coffee berries over a 44-d period maintained at 90% and 93.5% relative humidity, respectively. The number of immature developmental stages was dependent on ambient humidity with a 2 to 5-fold increase of immature stages found at 93.5% relative humidity versus 78% relative humidity (Baker et al. 1994).
Mass production of coffee berry borer's parasitoids is linked directly to coffee berry borer host production. To date, only 3 successful large-scale rearing of coffee berry borer parasitoids have been developed, 2 in Colombia by Portilla and Bustillo (1995) and Orozco (2002) using its natural host (parchment coffee), and 1 in the US by Portilla and Streett (2008) using Cenibroca artificial diet. The use of artificial diet has been crucial for coffee berry borer studies where its natural host is not available. Therefore, our study was designed to identify the optimal combination of relative humidity and moisture content level for coffee berry borer mass-production using the Cenibroca artificial diet (Portilla & Streett 2006). Moreover, effects of these abiotic factors on coffee berry borer preoviposition period, oviposition behavior, production, and reproductive potential were studied. The information provided by these efforts will allow for increased efficiency when mass-rearing coffee berry borer in an effort to produce large numbers of coffee berry borer parasitoids such as Cephalonomia stephanoderis Betrem (Hymenoptera: Bethylidae), Prorops nasuta Waterson (Hymenoptera: Bethylidae), and Phymastichus coffea LaSalle (Hymenoptera: Eulophidae).
Materials and Methods
INSECTS
Insects used in this study were obtained from a laboratory colony of coffee berry borers maintained at the National Research Center of Coffee, Chinchina, Colombia. Initially, colonies were shipped to the USDA-ARS Biological Control of Pest Research Unit quarantine facility in Stoneville, Mississippi, USA, in 2000 and reared for 4 generations before being transported to the Robert T. Gast facility in Starkville, Mississippi, USA. Coffee bean borers were cultured on Cenibroca diet in 14.6 × 55.1 cm 32 cell rearing trays (Portilla & Streett 2008) previously molded at the Gast facility using clear polyvinyl chloride plastic (American Mirrex Corp®, New York, New York, USA) (Tillman et al. 1977). Each cell measured 2.5 × 2.5 cm with a depth of 1.1 cm.
EXPERIMENTAL DESIGN
Cenibroca diet was added to 63 rearing trays using an automated dispenser pump drive (Master Flex®, Radnor, Pennsylvania, USA) calibrated to dispense 1.5 mL of material per cell (1,890 pellets were used per replication, i.e., 1,260 for preoviposition and feeding behavior, 315 for reproduction, and 315 per moisture content level). Trays with cooled diet were separated into 3 groups of 21 trays each then heated to 50 °C using an Econotherm® laboratory oven (Model PR305225M, Haverhill, Massachusetts, USA). Moisture content level in the diet was adjusted to 50%, 60%, or 70% for each treatment group. One diet pellet per tray per group was randomly removed from the oven to quantify diet moisture content level using an IR-50 Moisture Analyzer (P/N 901569-1 REV-A®, Denver Instrument Company, Denver, Colorado, USA). For each treatment group, all trays were stored for 2 d in a plastic container with a vented lid at 22 °C. After storage, 1 recently emerged coffee berry borer female (previously disinfected with 1% benzalkonium chloride solution BZK1000; www.LabAlley.com) was introduced into each cell. After introduction, diet trays were completely covered and sealed with clear self-laminate paper ( www.avery.com) with three 0.5 mm ventilation holes for each cell. Seven trays of each moisture content level treatment were randomly placed in a growth chamber maintained at 65%, 75%, or 85% relative humidity and 25 °C under continuous light. The 9 treatment combinations of moisture content level and relative humidity were repeated 3 times over time with random tray placement within the growth chamber during each repetition.
DEVELOPMENT AND OVIPOSITION ACTIVITY
Adult developmental behavior was observed daily for each treatment from the first d after infestation to the oviposition period. Seven randomly selected diet pellets across trays with adult feeding and oviposition activity were removed daily and dissected. From each dissected pellet the following was recorded: internal and external feeding by the adult, presence or absence of an entry hole into the pellet, presence or absence of eggs, and whether eggs were oviposited inside (internal reproduction) or on the outer surface of the pellet (external reproduction). Reproduction rates were calculated for each pellet noting whether it was internal, external, or both. Image-Pro Plus 7.0.1. photographic software (Media Cybernetics, Rockville, Maryland, USA) was used to verify internal and external oviposition.
REPRODUCTIVE POTENTIAL AND PROGENY PRODUCTION
Total offspring and stage of each coffee berry borer were recorded for each of 7 randomly selected diet pellets from each treatment every 10 d for 50 d. Diet pellets were dissected and the rate of coffee berry borer reproduction calculated. Eggs found at 50 d after infestation were recorded but were not included in life table calculations. An additional 7 diet pellets also were randomly removed from each treatment at the same time to measure moisture content level using an IR-50 Moisture Analyzer (P/N 901569-1 REV-A®, Denver Instrument Company, Arvada, Colorado, USA). After 50 d after infestation, evaluations of all remaining diet pellets were observed until adult emergence.
STATISTICAL ANALYSIS
The study design featured a 3 × 3 factorial (3 relative humidity and 3 moisture content levels) in a randomized complete block configuration. Sub-sampling was accomplished daily for the oviposition experi ment and every 10 d for the developmental rate study. The impact of moisture content level upon coffee berry borer reproductive rate was determined by developing a life table according to the methodology described by Ruiz-Cardenas and Baker (2010) and Portilla et al. (2014). The following parameters were estimated based on age (x), age-specific survival rate (lx), and age-specific fecundity rate (mx): net reproductive rate (Ro = 
lxmx ), mean generation time (T = lxmx/ xlxmx), doubling time (DT = ln(2)/rm), intrinsic rate of increase (rm) ( e-r(x+0.5)lxmx = 1), and finite rate of increase (λ = erm) (Carey 1993; Krebs 2001). To estimate the daily rate, the number of individuals recorded at the beginning of oviposition for each treatment replicate was divided by the number of d from the beginning of oviposition until the date of the first sampling (i.e., 10 d after infestation). Subsequently, the number of individuals observed for the second sample (i.e., 20 d after infestation) was subtracted from the number of individuals counted in the previous sample. This calculation was done iteratively for all subsequent samples (30, 40, and 50 d after infestation) as described in Portilla at al. (2014). The mx value for each treatment was determined based on the sex ratio obtained when reproduction ceased at 50 d after infestation. For those treatments with no males or with 1 or less females per male due to late oviposition, sex ratios from the longest preoviposition treatments were used. No foundress mortality was found; therefore, lx was 1 for all evaluations. All additional statistics were performed using SAS system software (SAS 2013). Mean data of progeny per pellet, preoviposition period, percentage of internal and external feeding, and developmental time were analyzed using t PROC GLM (ANOVA) followed by Tukey's HSD (P < 0.05) to detect differences between treatments.
Results
DEVELOPMENT, PREOVIPOSITION, AND OVIPOSITION ACTIVITY
There were statistically significant differences in oviposition: F(8,91) = 584.59; P = 0.0001; internal: F(8, 2) = 2379.08; P = 0.0001 and external: F(8, 2) = 551.60; P = 0.0001 feeding; and internal: F(8, 2) = 17726.30; P = 0.0001 and external: F(8, 2) = 16617.40; P = 0.0001 development (Table 1). The shortest preoviposition period occurred for females fed on diet pellets with 50% moisture content level and 65% relative humidity. The preoviposition period for coffee berry borer females continued to increase with corresponding increase in diet moisture content level and relative humidity (Table 1). Mature females reared on the 70% moisture content level diet pellets exhibited the longest preoviposition period for all 3 relative humidity regimens. All mature coffee berry borer females did not tunnel into 60% moisture content level diet pellets maintained at 75% and 85% relative humidity, or 70% moisture content level diet pellets at all 3 relative humidity regimens. Typically, eggs were deposited inside diet pellets at 50% moisture content level for all 3 relative humidity regimens. Females tunneled into the 50% moisture content level diet and deposited eggs as a mass at the end of the tunnel (internal reproduction). Females and larvae remained in the 50% moisture content level and 85% relative humidity diet pellet until emergence of F1 adults. Ovipositional behavior changed with diets containing 60% moisture content level at 65% relative humidity, with internal and external reproduction (Table 1). All foundress females reared on the higher moisture content level diet pellets (60% and 70%) deposited eggs on the outer surface of the pellets (i.e., external reproduction). Larvae emerging from eggs deposited on the external surface continued to feed externally and did not penetrate the pellet or feed internally. Foundress females tunneled and fed within the first h after they were placed onto the diet pellets with 50% moisture content level at all 3 relative humidity regimens. Feeding and minor burrowing behavior also was observed for coffee berry borer females reared on diet pellets with 60% moisture content level at the lowest relative humidity. In addition to internal feeding, the coffee berry borer females exhibited external feeding when reared on 60% moisture content level diet pellets maintained at 65% relative humidity. Coffee berry borer females were observed crawling on 70% moisture content level pellets for several h before tunneling and feeding.
Table 1.
Mean (± SE) preoviposition, feeding, and developmental behavior of coffee berry borer on Cenibroca artificial diet with 50, 60, and 70% moisture content at 65, 75, and 85% relative humidity.
A second generation (F2) was not observed for females reared on diet pellets at 50% moisture content level at 65% relative humidity (Table 1; Figs. 1 & 2). However, a F2 generation was observed for those borers reared on diets containing 50% moisture content level at 75% and 85%, as well as 60% moisture content level at 75% and 85% relative humidity (Table 1). The F2 generation produced at the higher moisture content diet levels occurred when the moisture content level in the diet pellets exceeded > 40% (Fig. 2).
PROGENY PRODUCTION AND REPRODUCTIVE POTENTIAL
Diets with 50% and 60% moisture content level at 65% relative humidity did not provide sufficient moisture for coffee berry borer reproduction and feeding. Consequently, high mortality resulting in a dramatic reduction in adult production was observed for those treatments at 40 and 50 d after infestation, respectively (Figs. 1 & 2). Table 2 shows the live progeny produced per diet pellet at different d after infestation for each treatment. Significant differences were detected among all treatments at 10 d: F(8, 2) = 58.59; P = 0.0001; 20 d: F(8, 2) = 64.94; P = 0.0001; 30 d: F(8, 2) = 84.83; P = 0.0001; 40 d: F(8, 2) = 69.39; P = 0.0001; and 50 d: F(8, 2) = 98.28; P = 0.0001. Females that fed on all treatments, with the exception of 50% moisture content level at 65% relative humidity, demonstrated similar ovipositional patterns. However, there were highly significant differences among treatments when cumulative mean number of offspring per female was considered (Table 2). Higher coffee berry borer progeny production levels were attributed to the moisture content level of pellets remaining between 40 and 50% (Figs. 1 & 2). The greatest mean progeny production was obtained with pellets containing 50% moisture content level at 85% relative humidity, and 60% moisture content level at 75% relative humidity. Conversely, low numbers of progeny per pellet were observed at 10 and 20 d after infestation for females reared on 60 and 70% moisture content level at 85% relative humidity pellets (Table 2) (Fig. 3A). This was attributed to the initially high moisture content level of the diet (Fig. 2). Females did not begin to oviposit in greater numbers until 30 d after infestation for those treatments (Fig. 1 & 3B). The number of progeny per pellet continued to increase at 40 and 50 d after infestation (Fig. 1 & 3C, D) as t moisture content levels of the pellets continued to decrease (Fig. 2).
Fig. 1.
Mean brood production of coffee berry borer per Cenibroca diet pellet with 50, 60, and 70% moisture content level maintained at 65, 75, and 85% relative humidity at different time periods. (95% confidence limits of the mean; n = 7 per evaluation time per treatment).
Fig. 2.
Loss of moisture content level percentage on a Cenibroca diet pellet with a 50, 60, and 70% moisture content level maintained at 65, 75, and 85% relative humidity. (95% confidence limits of the mean; n = 7 per evaluation time per treatment).
Diets that maintained a moisture content level > 40% following 50 d after infestation produced a second generation (Figs. 1 & 2; Table 2). This probably was due to the softness of the diet and the relative humidity being suitable for coffee berry borer reproduction. Consequently, coffee berry borer maintained on diets with 50% moisture content level at 75 and 85% relative humidity, and 60% moisture content level at 65, 75, and 85% relative humidity produced a second generation. Diet with 70% moisture content level could have produced a second generation; however, the moist conditions of the diet maintained at all 3 relative humidity regimens often led to Aspergillus Micheli (Eurotiales: Trichocomaceae) contamination.
The reproductive potential of coffee berry borer was influenced by relative humidity and the moisture content level of the diet (Table 3). The highest intrinsic rate of increase rm occurred with diet containing 60% moisture content level at 75% relative humidity. No significant differences were found among diets containing 50% moisture content level at 75 and 85% relative humidity. The finite rate of increase (i.e., λ, daughters per female per d) was significantly different among treatments (Table 3). Table 2 shows that sex ratio (although varied) averaged about 10 females per male.
Discussion
Hypothenemus hampei preoviposition and oviposition behavior on Cenibroca diet is comparable in some respects to that observed on its natural host, the coffee berry. The texture and moisture content level of the artificial diet is closely related to that of a coffee berry and thereby provides optimum conditions for establishment and reproduction of female borers (Ruiz-Cardenas & Baker 2010). We observed ever-ascending trends in the number of borers that built oviposition galleries on diets with low moisture content level, whereas longer preoviposition period and postponement of oviposition occurred on diets with high moisture content level. These results are in agreement with those of Abraham et al. (1990), who recorded longer preoviposition behavior by coffee berry borer adults in coffee berries where the endosperm was still soft. Further work by Gaviria et al. (1995) revealed that H. hampei postponed oviposition until the berry was desiccated.
We also observed that oviposition capacity increased in relation to the moisture content level in all treatments. No significant differences in number of eggs occurred among treatments at 30 d after infestation but there were significant differences in the other stages between treatments. Most of the treatments at that time lost moisture content principally for those exposed to 65% and 75% relative humidity, which provided an optimum media for reproduction and development (Fig. 2). Gaviria et al. (1995) found that the number of eggs laid was strongly influenced the physiological stage of the coffee berries, which agreed with Ruiz-Cardenas and Baker (2010), who recorded low numbers of eggs on berries with high dry weight. This could explain, in our study, the high mortality of immature stages at 40 d after infestation for the 50% moisture content level at 65% relative humidity, where the density of diet made it less suitable for development. In general, 65% relative humidity was not sufficient to hold enough moisture content for the initial brood or a build-up for a second generation. The 60% moisture content level at 75% relative humidity provided a better media for harvesting immature stages and adults due to external oviposition and development.
Table 2.
Mean number (± SE) live progenies of Hypothenemus hampei per tray cell reared on Cenibroca artificial diet with various moisture content levels (50, 60, and 70%) and maintained at different relative humidities (65, 75, and 85%)
Fig. 3.
Coffee berry borer external feeding and reproduction behavior on Cenibroca artificial diet with 60% moisture content at 75% relative humidity. Notice how the external reproduction offers a simple separation of the coffee berry borer immature and adults from the diet for coffee berry borer parasitoid production. (A) Diet pellet 10 to 15 d after infestation showing external feeding and development of first-generation offspring, arrows showing eggs of first generation. (B) Diet pellet 25 to 30 d after infestation showing external feeding and development of first-generation offspring. Stages suitable for the African ectoparasitoids (Cephalonomia stephanoderis and Prorops nasuta) reproduction. (C) Diet pellet 35 to 40 d after infestation showing external feeding and development of first-generation offspring. Notice the presence of mature and teneral females, arrows showing oviposition of second generation. (D) Diet pellet > 50 d after infestation showing female production, suitable for the reproduction of the African ectoparasitoid P. nasuta.
Table 3.
Mean number (± SE) reproductive parameters of Hypothenemus hampei reared on Cenibroca artificial diet with various moisture content levels (50, 60, and 70%) and maintained at different relative humidities (65, 75, and 85%). Data was generated 50 d after infestation.
In summary, the developmental parameters values obtained in our study clearly demonstrated that Cenibroca diet with 60% moisture content at 75% relative humidity was the best medium for H. hampei survivorship and external reproduction. We suggest that the average number of coffee berry offspring (second instars, prepupae, and pupae) obtained per diet pellet at 30 d after infestation should be suitable for the development of ectoparasitoids C. stephanoderis, and P. nasuta (Portilla et al. 2014). If rearing the endoparasitoid P. coffea female, H. hampei produced on diet pellets 40 to 50 d after infestation will be suitable for this species (Portilla & Grodowitz 2018). Although microbial contamination was present in diets at 85% relative humidity, it generally did not appear until after 40 d after infestation. This may not be a serious issue because coffee berry borer immatures typically will be collected at 35 to 40 d after infestation depending on the parasitoid rearing system. In summary, the external development of H. hampei on Cenibroca artificial diet with 60% moisture content level at 75% relative humidity offers a simple technique for separation of coffee berry borer immatures from the diet for coffee berry borer parasitoid production (Portilla & Streett 2008).
Acknowledgments
The authors would like to express their appreciation to Luis Carlos Jojoa, Department of Entomology and Plant Pathology, Mississippi State University, Mississippi, USA, for technical assistance. We also are grateful to Juan Morales-Ramos and Michael Grodowitz (NBCL-ARS-USDA) for critically reviewing an early version of this manuscript.
