The small avocado seed weevil, Conotrachelus perseae Barber (Coleoptera: Curculionidae), is reported from Guatemala, Honduras, Costa Rica, and Mexico (Barber 1919, Whitehead 1979). In Mexico, the distribution of C. perseae is restricted to an altitudinal range of 1,270 to 1,954 m above sea level (masl), in the biogeographic provinces of the central and southern parts of the country including the states of Mexico, Puebla, Morelos, Guanajuato, Hidalgo, Veracruz, Oaxaca, and Chiapas. These regions produce avocado commercially and also have native populations of other species of Persea and botanical varieties of avocado such as P. americana var. drymifolia, P. americana var. guatemalensis, and P. americana var. americana that can serve as host plants for this curculionid (Castañeda-Vildózola et al. 2015, Luna et al. 2017, Vázquez et al. 2015).

According to Peterson and Orden (2008), C. perseae is classified as a quarantine pest whose presence in commercial orchards presents a risk for the export of Mexican avocado to the United States. This initiated several studies aimed at an improved knowledge of the biology of this species (Llanderal and Ortega 1990, Payán-Arzapalo 2014, Salinas-Castro 1999). Castañeda-Vildózola et al. (2015) documented that C. perseae is nocturnal and that females perforate growing fruits to oviposit. The number of perforations ranges from one to six, with one to eight eggs deposited in each perforation. Larvae hatch from eggs, on average, in 7.0 d and feed in the seed, with as many as 24 larvae per seed. The entire larval stage lasts 25.0 d and, once concluded, last instars abandon the fruit and pupate in the soil. The prepupal and pupal stages last 24.4 and 16.33 d, respectively.

Integrated pest management programs require knowledge of the biology, ecology, and behavior of the target pest (Gold et al. 1999). Enumeration and characterization of the larval instars of the insect is an important component, but for spermatophagous species, such as C. perseae, the direct observation of larval molts is difficult, thus forcing reliance on indirect methods such as the measurement of head capsule width (Dyar 1890). Computer algorithms also have been developed for determination of larval instar based on head capsule widths (Logan et al. 1998, Merville et al. 2014).

Use of Dyar's (1890) rule is useful in determining larval instar of the curculionids Lissorhoptrus kuscheli O'Brien (Pantoja et al. 1999), Anthonomus grandis Boheman (Reardon et al. 2002), Conotrachelus psidii Marshall (Bailez et al. 2003), Pissodes castaneus (De Geer) (Panzavolta 2007), Rhyssomatus subtilis Fiedler (Cazado et al. 2014), Curculio chinensis Chevrolat (Zhi-Wen et al. 2015), and Heilipus lauri Boheman (Castañeda-Vildózola et al. 2016). Its application to weevils with intraspecific variation in the number of larval instars, such as Cosmopolites sordidus Germar (Gold et al. 1999), Rhynchophorus ferrugineus Olivier (Jaya et al. 2000), and Scyphophorus acupunctatus Gyllenhal (Valdés-Estrada et al. 2010), is much less reliable. Such variation is attributable to biological polymorphism, food availability and quality, and environmental factors like temperature, humidity, and photoperiod (Esperk et al. 2007).

Several studies of C. perseae in Mexico have reported number of larval instars observed, but the topic remains controversial. Payán-Arzapalo (2014) reported four larval instars, while Salinas-Castro (1999) and Cancino and Ponce (2000) observed five instars, and Llanderal and Ortega (1990) and Coria-Ávalos (1999) reported six. Our study was undertaken to clarify the number of larval instars occurring in a population of C. perseae collected from Persea americana Mill. cv. Fuerte avocado fruits cultivated in the state of Mexico.

Materials and Methods

Infested avocado fruits were collected from three orchards located in the community of Meyuca de Morelos, in the municipality of Coatepec Harinas, state of Mexico (N18.55°142′, W99.78°143′; 1,948 masl) at weekly intervals from January through October of 2015. In total, 1,756 infested avocados weighing a total of 311.23 kg were collected and transported to the laboratory to extract eggs and larvae. This is a zone of transition between temperate subhumid and warm subhumid climates. The mean annual temperature is 28°C (range, 11–37°C), with a mean annual precipitation of 1,321 mm.

In the laboratory, each oviposition hole was dissected using a camel hair brush to extract eggs; eggs were individually placed in petri dishes (9.0 × 1.0 cm) with absorbent paper moistened with distilled water as an incubation substrate. Eggs were monitored every 12 h, and newly emerged larvae were collected and preserved in 70% ethanol. Larvae also were extracted from seeds, placed in hot water (96°C), separated by size, and preserved in 70% ethanol.

Head capsule widths of the larvae were measured using computerized image analysis. Digital microphotographs were taken of the head capsules in frontal view with a PAXcam 3 Digital Microscope Camera (PAX-it, Villa Park, IL) mounted on a Tessovar III photomicroscope (Carl Zeiss, Obekochen, Germany). Images were saved in JPG format. The system was calibrated using an image of a ruler marked in gradations of hundredths of millimeters. The head capsules in the images were measured in micrometers from the outermost lateral edges (gena), using the program Image Tool 3.0 (Wilcox et al. 2002).

Head capsule width measurements were analyzed using the HCAP program developed by Logan et al. (1998), which is a generalization of the technique previously described by McClellan and Logan (1994). The precision of Dyar's rule was analyzed using linear regression by fitting the number of larval instars obtained with the natural logarithm of the mean head capsule width of each instar, following the method proposed by Gaines and Campbell (1935). A one-way analysis of variance using InfoStat (2008) software was conducted to test the statistical significance of the slope of the regression line.

Results and Discussion

Head capsules were measured on 1,148 larvae, including measurements of 56 neonates from eggs removed from fruits and incubated in petri dishes. Head capsule widths ranged from 212 to 1,562 µm which, according to the frequency distribution analysis, formed five well-defined modal groups and suggests five larval instars (Fig. 1). The descriptive statistics, including mean, standard deviation, and number of larvae observed in each stage, are included in Table 1.

View Figure

• Download as Powerpoint
• Download Figure

Table 1 Mean, range, and probability of misclassification for the measurements (width) of head capsules of the larval instars of C. perseae calculated using the HCAP program.

View Fullscreen

The minimum and maximum values for each larval instar assigned by HCAP reduced the probability of overlap attributed to misclassification, which was below 3% (Table 1). The value of Dyar's constant was between 1.39 and 1.58, which suggests a geometric progression of growth of the head capsule of each larval instar as a signal of Dyar's (1890) rule. Gaines and Campbell (1935) concluded that a perfect geometric progression in head capsule growth could be represented by a simple linear regression, plotting the natural logarithm of head capsule width and the number of larval stages. In our study, the linear equation of the linear regression of head capsule width on the increase in each larval stage was highly significant and indicates that all head capsule measurements were satisfactorily classified as one of the five instars (P = <0.0001, R = 0.992, R² = 0.988) (Fig. 2). The excellent fit of the model indicates no overlapping instars. The significant difference observed between head capsule width for each larval instar (F = 13.188.28; df = 3, 11; P = <0.0001) confirms the results obtained from the linear regression.

View Figure

• Download as Powerpoint
• Download Figure

The results of our study differ from those reported by Llanderal and Ortega (1990), Coria-Ávalos (1999), and Payán-Arzapalo (2014), but are consistent with those of Salinas-Castro (1999) and Cancino and Ponce (2000). Payán-Arzapalo (2014) measured only 357 larvae extracted from a mix of Hass, Fuerte, and criollo avocados (P. americana var. drymifolia); therefore, it is possible that the small sample size may have limited their conclusions of the occurrence of four larval instars. They also noted that their definition of instars 1 and 2 was uncertain. Llanderal and Ortega (1990) and Coria-Ávalos (1999) reported six larval instars in larvae collected from the Hass variety of avocado Coria-Ávalos (1999) also failed to include methods of measurement and analysis of head capsule measurements. While Llanderal and Ortega (1990) used similar methods of measurement and analysis as reported herein, their six modal groups had very narrow ranges among the instars, indicating the possibility of erroneous assignments of head capsule widths during their frequency distribution analysis.

Salinas-Castro (1999) and Cancino and Ponce (2000) suggested five larval instars in C. perseae collected from the Hass variety of avocado. Both studies used the same measurement methods but different techniques for analysis of head capsule widths. The results of those two studies suggest a slight overlap between the second and third instars. There was an observed 7% probability of misclassification in the Salinas-Castro (1999) study, while our study yielded a 3% probability of misclassification. The lower level of error in our study is attributed to a larger sample size of 1,148 larvae as compared with 841 in Salinas-Castro (1999).

Research in other species of Conotrachelus indicates interspecific variation in number of larval instars, but no intraspecific variation, as demonstrated by Bodenham et al. (1976), Raccette et al. (1992), Mendes et al. (1997), and Bailez et al. (2003), whose results showed four larval instars in C. nenuphar (Herbst), C. humeropictus Fiedler, C. neomexicanus Fall, and C. psidii Marshall, respectively. On the other hand, Muñiz and González (1982) and Tedders and Payne (1986) determined five larval instars for C. dimidiatus Champion and C. schoofi Papp.

Our study adds one more species to the list of Conotrachelus with five larval instars and is supported by the results of Salinas-Castro (1999) and Cancino and Ponce (2000). In addition, these results suggest that the genotype of the avocados used as food by the larvae (Hass cultivated in Michoacán; Fuerte cultivated in the state of México) does not lead to variation in the number of larval instars. As such, the discrepancies among different studies (i.e., Llanderal and Ortega 1990 and Payán-Arzapalo 2014) is more likely attributed to the small number of larvae measured and errors in the interpretation of the frequency distribution analysis of the head capsule width of C. perseae larvae.