Selection and Suitability of an Artificial Diet for Tuta absoluta (Lepidoptera: Gelechiidae) Based on Physical and Chemical Characteristics.

Tuta absoluta (Meyrick, 1917) is a key tomato pest in South America and, recently, in Europe and Africa. To develop efficient control methods for this pest, adequate rearing protocols are desirable. As an alternative to tomato leaves (natural diet), we evaluated four artificial diets. Biological traits including larval and pupal viability and development time, pupal weight and deformations were assessed. Additionally, the optimum container size and larval density were evaluated. The diet based on casein, wheat germ and cellulose allowed the best development of T. absoluta, showing higher viability and no negative effects on larval instars and pupal weight. The best container was a glass tube measuring Ø 1 × h 6 cm, topped with waterproof cotton, with a density of three larvae. To evaluate the suitability of this diet, T. absoluta was reared during eight generations and life-table parameters were estimated for the F1, F3, F6, and F8 generations. The total viability (egg-adult) increased over the generations, reaching 75% in the eighth generation. Based on life-table estimations no differences among generations were found. The net reproductive rate (Ro) was higher than 40, the intrinsic rate of increase (rm) ranged between 0.08 and 0.11, the finite rate of increase (λ) was 1.1, the mean generation time (T) have a maximum of 44 d and doubling time ranged from 5.89-8.32 generations. These results indicated that a diet based on casein, wheat germ and cellulose was suitable for T. absoluta rearing in laboratory conditions.

Tuta absoluta (Meyrick, 1917) (Lepidoptera: Gelechiidae) It is one of the most devastating tomato pests because it feeds on all aerial structures at all stages of plant growth, causing losses of 80–100% (Desneux et al. 2010; Escobar and Lee 2009; Estay 2009; Urbaneja et al. 2012, 2013) As well as on tomatoes, this pest has been reported from other economically important solanaceous plants including Solanum tuberosum (L.) (potato), Solanum melongena (eggplant) and weed plants like Solanum nigrum (European black nightshade) L., Atropa belladonna L (belladonna), Datura ferox (L.) (fierce thorn–apple), Nicotiana longiflora Cav. (longflower tobacco), Nicotiana glauca Graham (tree tobacco), Salpichroa origanifolia (Lam.) Baill. (Lily of the valley vine), Solanum americanum Mill. (American black nightshade), Solanum sisymbriifolium Lam. (sticky nightshade), Brugmansia arborea (L.) (angel’s trumpets), Solanum chenopodioides Lam. (goosefoot nightshade), Solanum sarrachoides Sendtner (hoe nightshade), Solanum pygmaeum and Datura stramonium (L.) (Jimson weed) (Bawin et al. 2015, Salas Gervassio et al. 2016) This lepidopteran was first reported in 1914 in Peru, and currently is a common pest found in South America (Jham et al. 2001) Since 2006, T. absoluta had invaded Europe and Africa, where it has caused significant economic losses (Desneux et al. 2011, Urbaneja et al. 2012, 2013).

The failure to control this pest may have a strong economic impact, and its recent history of introductions has increased the need for studies to develop strategies for its biological control, by the use of natural enemies, mainly the larval parasitoids such as Apanteles gelechiidivoris Marsh (Bajonero et al. 2008) or Pseudapanteles dignus (Muesebeck) (Luna et al. 2007), and microorganisms such as granulovirus PhopGV (Mascarin et al. 2010). For these studies it is necessary the production of large numbers of insects preferably with low costs and high performance (Zou et al. 2015). The rearing of T. absoluta in natural diet needs extensive manual labor and space to keep plants capable to support a significant pest population (Cely et al. 2010). The most used tomato varieties are “Santa Clara” and “IPA6” (Neves et al. 2003, Oliveira et al. 2009) in South America and “Marmande” (Chailleux et al. 2013, Han et al. 2015) and “Money maker” (Cuthbertson et al. 2013) in Europe.

Artificial diets have been essential to the development of knowledge in biology and ecology of many species of insect pests (Sorensen et al. 2012). Due to the possibility of producing a large number of insects, it has been possible to study nutritional processes, biochemical, behavioral and biological control methods of many insect pests (Parra 2012; Vanderzant 1974). Some artificial diets for T. absoluta have been tested. Giustolin et al. (1995) evaluated 10 diets for T. absoluta with different protein contents, and determined that a diet based on “Carioca” white beans, wheat germ, casein, yeast extract and soy protein was the most appropriate. Later, Mihsfeldt and Parra (1999) developed a diet based on white beans plus powdered tomato leaves that showed promise for rearing T. absoluta in laboratory conditions. However, the total viability obtained was <75%, which is proposed by Singh (1983) as the minimum for the use of a diet in the production of insects in the laboratory.

To understand the effect of external factors, in this case an artificial diet, in the development, survival and reproduction at the population level, life tables are a very important tool (Chi and Su 2006, Wittmeyer and Coudron 2001). They provided useful information to analyze the mortality factors in insect population and evaluate the relative importance of the numerous independent variables derived from field or laboratory experimentation (Kakde et al. 2014).

To improve rearing techniques for T. absoluta, the aim of this study was to select an artificial diet for T. absoluta based on physical and nutritional characteristics, assessing the insect’s biological development, the most suitable container and density in the selected diet, and evaluating insect development for eight successive generations.

Material and Methods

Rearing Stock of T. absoluta

Individuals used in this experiment were obtained from the colony of T. absoluta maintained in the Laboratory of Insect Biology of the Department of Entomology and Acarology, Luiz de Queiroz College of Agriculture, University of São Paulo (ESALQ-USP), Piracicaba, Brazil. Rearing was initiated with a population collected in the state of São Paulo, Brazil, in 2012. T. absoluta was maintained in rearing cages 70- × 50- × 50-cm (height × width × depth) covered with voile containing tomato plants var. “Santa Clara” in a climate-controlled room (temperature 25 ± 2 °C, relative humidity (RH) 70 ± 10% and a photoperiod of 14:10 (L:D) h. The plants were infested with adults, allowing oviposition for 24 h, after that, they were removed and the larvae allowed to develop until pupae formation. Obtained adults were transferred to new cages and were fed with honey droplets placed on the walls of the cage. Eggs were collected daily on individual tomato leaflets kept in a container with water and fastened at the top of the cage. Subsequently, the eggs were removed from the leaflets with a brush inside a glass recipient with distilled water, and then collected by filtration on polyester fabric. The eggs were placed in a Petri dish and kept in an incubator (temperature 25 ± 2 °C, RH 70 ± 10% and a photoperiod of 14:10 (L:D) h until the larvae hatching.

Diets Evaluated

Four diets were compared (Table 1): D1:the diet of Mihsfeldt and Parra (1999) for rearing T. absoluta, based on the diet of Greene et al. (1976) ”developed for Anticarsia gemmatalis Hübner; D2: the diet of Hensley and Hammond (1968) which is recommended for rearing Diatraea saccharalis (F.) and other lepidopterans; D3: the diet of Berger (1963) which was initially used for Heliothis virescens (F.) and contains cellulose, a component that does not provide nutrients but confers structural characteristics on the diet that may benefit the insect feeding; and D4: Berger’s (1963) diet, with the addition of lyophilized powdered tomato leaves . The tomato-leaf powder was obtained by lyophilizing 7-week-old leaves of the variety Santa Clara for 3 d in a lyophilizer (model Novalyphe-NL150, Savant Instruments, Holbrook, NY), and then ground to a fine powder. Because of the small size of T. absoluta, the dry ingredients with larger particles were milled to homogenize each diet. The artificial diets were prepared as described by Parra (2013). Fresh tomato leaves were used as control (D5).

Table 1.

Composition of artificial diets for rearing T. absoluta

Component	D1a	D2b	D3c	D4d
White beans	37.5 g	–	–	–
Wheat germ	30.0 g	27.0 g	22.5 g	22.5 g
Cellulose	–	–	3.7 g	3.7 g
Soy protein	15.0 g	–	–	–
Saccharose	–	45.0 g	26.2 g	26.2 g
Choline chloride	–	0.9 g	0.7 g	0.7 g
Wesson salts	–	9.0 g	7.5 g	7.5 g
Casein	15.0 g	27.0 g	26.2 g	26.2 g
Brewer’s yeast	18.7 g	–	–	–
Ascorbic acid	1.8 g	3.6 g	3.0 g	3.0 g
Sorbic acid	0.9 g	–	–	–
Methyl parahydroxybenzoate (Nipagin)	1.5 g	1.3 g	1.5 g	1.5 g
Tetracycline	56.5 mg	0.2 mg	0.5 mg	0.5 mg
Formaldehyde 37%	1.8 ml	0.4g	1.0 ml	1.0 ml
Vitamin solutione	4.5 ml	9.0 ml	7.5 ml	7.5 ml
Agar	11.5 g	18.0 g	9.0 g	9.0 g
Water	600.0 ml	780.0 ml	630.0 ml	630.0 ml
Powdered tomato leaves (lyophilized)	15 g	–	–	4.0 g

aMihsfeldt and Parra (1999), based on Greene et al. (1976).

bHensley and Hammond (1968).

cBerger (1963) modified.

dBerger (1963) modified + powdered tomato leaves.

eNiacinamide: 1.0 g; calcium pantothenate: 1.0 g; riboflavin: 0.50 g; thiamin: 0.5 g; pyridoxine: 0.25 g; folic acid: 0.10 g; biotin: 0.02 mg; vitamin B12 (1,000 mg/ml): 2.00 ml.

Development of T. absoluta on Artificial and Natural Diets

For each diet, 100 repetitions with three larvae each were used. The container used was an Ø1.0 × h 6.0 cm flat-bottom glass tube capped with water-repellent cotton. In each container about 4 ml of the diet tested was placed, taking care to form a thin film of diet on the walls of the container, increasing the area covered and available for the larvae to feed and simulating the natural conditions, where the larvae feeds on the leaf.

For the natural diet, the experimental unit was a plastic cup (Ø 8.5 cm × h 14), covered with thin veil to allow gas exchange. Each cup contained a tomato leaflet with the petiole submerged in water. In each experimental unit, three newly hatched larvae were placed, as for the artificial diet.

For the artificial diets, the mortality at 7 d after inoculation was evaluated, and for natural and artificial diets duration and viability of larval and pupal stages, pupal weight after 24 h, sex ratio (Butt and Cantu 1962, Sannino and Espinosa 2010) and percentage of deformations were estimated. With the adults obtained from each diet, 15 pairs were formed and placed in individual cages similar to those used to assess the natural diet. The egg viability was evaluated by collecting 20 eggs from each pair from the second day of oviposition, totaling 300 eggs per diet; these eggs were placed on filter paper in glass Petri dishes. All assay units were kept in a laboratory at a temperature of 25 ± 2 °C, RH 70 ± 10% and a photoperiod of 14 h.

Determination of Instar Number of T. absoluta on the Selected Diet (D3)

After the results obtained in the previous experiments, the D3 diet was selected for the subsequent assays. To assess whether the diet that produced the highest viability affects the development of T. absoluta, the number of instars was determined. Eighty tubes of diet D3 were each inoculated with one larva. The head-capsule width was measured until pupation, using an ocular micrometer attached to a stereomicroscope (20× magnifications). The number of instars and the growth rate were determined by the mathematical model based on Dyar (1890), as recommended by Parra and Haddad (1989).

Container Size for Rearing T. absoluta

In order to assess whether the type of container affected the development of T. absoluta, three sizes and types of container were compared. The first container “A” was a large glass tube of Ø 2.0 × h 8.5 cm covered with waterproof cotton, which is commonly used for rearing Lepidoptera in Brazil. The second was similar to the first, but smaller, measuring Ø 1.0 × h 6.0 cm, it was chosen due to the small size of T. absoluta. The third container was a plastic cup measuring Ø 4.0 × h 6.0 cm, with plastic cover, which is used for rearing non cannibalistic larva. It was used thirty containers with one larva per treatment. Larvae were allowed to develop to the pupal stage and their viability was compared. Water loss from each container was also determined, using 10 containers of each type with no larvae, each type of container were dried at 50 °C (due to the plastic material of the containers) to constant weight. The water content was calculated as a percentage (Colestino 2010):

Optimum Number of T. absoluta larvae per container

Densities of 1, 3, 6, and 12 larvae in each tube were tested. The experimental unit was a flat-bottom glass tube Ø 1.0 × h 6.0 cm capped with water-repellent cotton containing the diet that gave the best results (D3). For each treatment, 30 experimental units were evaluated for the parameter of larval viability.

Biological Development of T. absoluta on the Selected Diet for Eight Generations

T. absoluta development reared on the most appropriate diet, D3, was evaluated throughout the F1, F3, F6, and F8 generations. For each generation, 100 tubes were inoculated with three larvae each one. The viability in 7 d, larval and pupal viability and duration were evaluated. From the emerged adults, 25 pairs were formed and kept in identical cages to those described on “Development of Tuta absoluta on Artificial and Natural Diets” and fed with pure honey droplets (Bogorni and Carvalho 2006). The pre-oviposition period, longevity and number of eggs laid daily per female were recorded. For each pair, 20 eggs were separated from the second day and the egg viability was calculated. All tests were performed at a temperature of 25 ± 2 °C, RH 70 ± 10% and a photoperiod of 14 h.

Data Analysis

Data were analyzed by generalized linear models (GLMs) (Nelder and Wedderburn 1972) of the binomial type for viability and sex ratio data, and the Poisson type for duration, longevity, and number of eggs laid per female. The data were evaluated using a standard half-probability graph with simulated envelope (Demétrio and Hinde 1997, Hinde and Demetrio 1998) . In case of significant differences, the Tukey multiple comparisons test was applied at 5% significance, using the glht function multicomp package with adjusted P.

With the collected data for the D3 diet, fertility life tables were constructed and the net reproductive rate (Ro), mean generation time (T), intrinsic rate of increase (rm), finite rate of increase (λ), and doubling time (DT) were calculated. These parameters were compared by a bootstrap analysis (Meyer et al. 1986) and all analyses were done in the statistical program R version 3.2.3.

Results

Comparing the performance of the four artificial diets at 7 d after inoculation showed that diet D3 yielded approximately 77% of viability the highest value, followed by diet D2 (F = 11.972; df = 3, 895; P < 0.001). Larvae fed on diets containing lyophilized tomato leaves (D1 and D4) did not reach a viability level of 60% 7 days after inoculation (Table 2).

Table 2.

Effects of five different diets on T. absoluta viability within 7 d after feeding (mean % ±EP), developmental stage duration (mean days ± EP), and egg to adult viability (mean % ± EP).

Parameter		Diet
		D1a	D2b	D3c	D4d	D5e
Viability with 7 df		44.67 ± 4.72 d	65.33 ± 54.36 b	76.63 ± 2.85 a	55.33 ± 5.18 c	–
Duration (days)g	Larva	15.83 ± 4.25 ab	17.64 ± 2.07 bc	22.10 ± 1.24 d	16.12 ± 3.87 b	12.99 ± 1.72 a
	Pupa	8.34 ± 0.13 c	8.65 ± 0,127 bc	9.20 ± 0, 175 a	8.22 ± 0.19 c	9.02 ± 1.16 ab
Viabilityf(%)	Egg	87.00 ± 1.94 b	86.33 ± 1.98 b	94.10 ± 1.37 a	90.00 ± 1.74 ab	93.53 ± 1.44 a
	Larva	35.33 ± 3.91 d	57.33 ± 4.05 c	74.83 ± 2.55 b	48.00 ± 4.09 cd	91.39 ± 2.30 a
	Pupa	84.90 ± 4.96 ab	81.40 ± 4.22 b	97.74 ± 1.86 a	74.64 ± 5.19 b	93.43 ± 2.12 a
	Total	26.36	40.00	69.71	30.80	80.29

aMihsfeldt and Parra (1999), based on Greene et al. (1976).

bHensley and Hammond (1968).

cBerger (1963) modified.

dBerger (1963) modified + powdered tomato leaves.

eTomato leaves.

fMeans followed by the same letter in the line do not differ significantly (GLM with quasi-binomial distribution, followed by post hoc Tukey test; P < 0.05).

gMeans followed by the same letter in the line do not differ significantly (GLM with quasi-Poisson distribution, followed by post hoc Tukey test; P < 0.05).

The larval period lasted from 16 to 22 d on artificial diets, a longer period compared with the natural diet (F = 366.33; df = 4, 554; P < 0.001). Among the artificial diets, shorter larval and pupal periods were obtained in the diets containing lyophilized tomato leaves. The longest durations of larval (22 d) and pupal (9 d) stages were obtained with diet D3 (Pupa: F = 12.63; df = 4, 442; P < 0.001) (Table 2).

Larval viability was highest with the natural diet (91%); among the artificial diets, D3 allowed the highest viability (74%) (F = 35.155; df = 4, 895; P < 0.001). The pupal viability was >70% in all diets, being higher in the natural diet and D3 (F = 5.1308; df = 4, 560; P < 0.001). The egg viability exceeded 80% with all four diets; however, those registered with diets D3 and D4 reached 90%, and were equal to that obtained with the natural diet (F = 4.2583; df = 4, 1499; P < 0.001), (Table 2).

A 1:1 Sex ratio was registered in all treatments (χ2 = 6.33; df = 4; P = 0.1753) (Table 3). Pupal weight was about 3.5 mg when T. absoluta was fed with diets D3–D5, being heavier than those fed on diets D1 and D2. The deformation rates were <5% for all diets (Table 3).

Table 3.

Sex ratio and weight of pupae of T. absoluta reared on five diets

Diet	Sex ratioa	Pupal weight (mg)b		Deformations%
		Females	Males
D1c	0.55	2.50 ± 0.18 b	2.32 ± 0.107 b	3.76
D2d	0.44	2.51 ± 0.15 b	2.20 ± 0.12 b	4.00
D3e	0.40	3.53 ± 0.07 a	3.34 ± 0.104 a	3.33
D4f	0.53	3.58 ± 0.09 a	3.44 ± 0.069 a	4.33
D5g	0.50	3.64 ± 0.10 a	3.48 ± 0.047 a	–

aNo difference in sex ratio (Chi2 = 6.33, df = 4, P = 0.1753).

bMeans followed by the same letter in the column do not differ significantly (GLM with quasi-Poisson distribution, followed by post hoc Tukey test; P < 0.05).

cMihsfeldt and Parra (1999), based on Greene et al. (1976).

dHensley and Hammond (1968).

eBerger (1963) modified.

fBerger (1963) modified + powdered tomato leaves.

gTomato leaves.

Larvae of T. absoluta fed on diet D3 showed four instars which are the same observed when the insect develops on a natural diet. The widths of the head capsule were 0.13 ± 0.014, 0.19 ± 0.025, 0.34 ± 0.051, 0.49 ± 0.046, for the first to fourth instar, respectively.

Larval viability of tomato pinworm measured in the small glass tube resulted in higher viability (66.7%) compared with the other two types of containers tested (F = 5.906; df = 2, 87; P = 0.027) (Table 4). Water loss was greater in the glass tube container (Table. 4).

Table 4.

Larval viability (%) of T. absoluta reared in three types of containers and water loss of the diet per container

Container type	Larval viability (%)a	Water loss (%)
Large glass tube (2.0 × 8.5 cm)	43.33 ± 9.20 b	15.75
Small glass tube (1.0 × 6.0 cm)	66.66 ± 8.75 a	20.40
Plastic cup (4.0 × 6.0 cm)	23.33 ± 7.85 c	13.23

aMeans followed by the same letter in the column do not differ significantly (GLM with quasi-binomial distribution, followed by post hoc Tukey test; P < 0.05).

Optimum Number of T. absoluta Larvae per container

Larval viability was greater for a density of 3 T. absoluta larvae (87%), and decreased at lower and higher densities (F = 23.397; df = 4, 75; P = 0.0445). The pupal viability (80–90%) was not affected by the density (F = 0.2137; df = 4, 70; P = 0.091) (Table 5).

Table 5.

Larval and pupal viability of T. absoluta reared at five densities per rearing container

Number of larvae per container	(%)
	Larval viability	Pupal viability
1	68.75 ± 4.78 ab	90.9 ± 10.16 a
3	87.5 ± 5.70 a	93.75 ± 5.84 a
6	77.08 ± 7.76 ab	78.12 ± 10.30 a
9	76.39 ± 4.60 ab	90.40 ± 4.83 a
12	61.45 ± 1.35 b	98.43 ± 1.46 a

aMeans followed by the same letter in the column do not differ significantly (GLM with quasi-binomial distribution, followed by post hoc Tukey test; P < 0.05).

Biological Development of T. absoluta on the Selected Diet for Eight Generations

A viability of ca. 93% was reached during the first 7 d after feeding with the artificial diet in the eighth generation, being greater than in the previous generations (F = 18.492; df = 3, 1196; P < 0.001) (Table 6). The developmental period of eggs was longer in the first generation, decreasing in subsequent generations (F = 3.66; df = 3, 39; P < 0.001) (Table 6). The larval stage was longer in F1, then significantly shorter in F6 and F8 and the shortest in F4 (F = 262.63; df = 3, 918; P < 0.001). Generation F4 showed a longer pupal stage (a mean of 11.5 d) than the other generations assessed (F = 83.24; df = 3, 747; P < 0.001). The T. absoluta developmental time, considering from egg to adult were 34.7, 30.7, 33.0, and 31.8 in the first, fourth, sixth, and eighth generations, respectively.

Table 6.

Viability with 7 d, duration and viability of T. absoluta on diet D3 (casein, wheat germ and cellulose) in the first, fourth, sixth, and eighth generations

Parameter		Generation
		F1	F4	F6	F8
Viability with 7 d a		77.77 ± 2.52 b	87.66 ± 1.62 ab	83.00 ± 2.22 ab	93.33 ± 1.44 a
Duration (days) b	Egg	3.70 ± 0.221 a	3.00 ± 0.00 c	3.40 ± 0.16 b	3.30 ± 0. 15 b
	Larva	21.93 ± 1.53 a	16.24 ± 1.39 c	20.50 ± 1.73 b	20.36 ± 2.04 b
	Pupa	9.25 ± 1.15 b	11.55 ± 1.31 a	9.36 ± 1.10 b	9.05 ± 1.67 b
Viability (%) a	Egg	98.75 ± 0.05 a	94.50 ± 0.11 b	97.25 ± 0.80 ab	93.00 ± 0.10 b
	Larva	76.00 ± 3.85 a	83.00 ± 3.63 a	79.33 ± 2.34 a	88.33 ± 2.99 a
	Pupa	91.28 ± 1.91 a	89.75 ± 2.00 a	93.59 ± 1.80 a	89.64 ± 1.92 a
	Total	68.87 ± 0.42	69.97 ± 0.91	71.18 ± 0.70	74.66 ± 0.94

aMeans followed by the same letter in the line do not differ significantly (GLM with quasi-binomial distribution, followed by post hoc Tukey test; P < 0.05).

bMeans followed by the same letter in the line do not differ significantly (GLM with quasi-Poisson distribution, followed by post hoc Tukey test; P < 0.05).

The viability of the egg stage was highest in the first generation and decreased in subsequent generations ranging from 93 to 99% The larval and pupal stages showed no difference in the viability across the generations (larva: F = 3.491; df = 3.118; P = 0.0058; pupa: F = 0.4962; df = 3.952; P = 0.685). The total viability was >70% in the four generations studied.

T. absoluta female longevity ranged from 12 to 16 d throughout the generations studied when fed on the artificial diet. Fecundity was lowest in the first generation, and increased in subsequent generations up to 130 eggs per female (Table 7).

Table 7.

Longevity and fecundity of T. absoluta reared for eight generations on a diet based on casein, wheat germ and cellulose

Generation	Longevity (days) a	Fecundity (egss per female)a
F1	12.80 ± 1.11 b	88.06 ± 13.89 b
F4	11.77 ± 0.79 b	117.25 ± 10.73 a
F6	15.97 ± 0.78 a	130.72 ± 12.32 a
F8	13.96 ± 0.63 b	115.88 ± 8.33 a

aMeans followed by the same letter in the column do not differ significantly (GLM with quasi-Poisson distribution, followed by post hoc Tukey test; P < 0.05).

Life-table analyses for generations F1, F4, F6, and F8 (Table 8) showed no differences among generations. The net reproductive rate (Ro) was between 40 and 45, the intrinsic rate of increase (rm) ranged between 0.008 and 0.11, the finite rate of increase (λ) was 1.1 and the mean generation time (T) was 32–44 d. The DT was also similar, ranging from 5.89 to 8.32 generations.

Table 8.

Life-table parameters for T. absoluta reared for eight generations on diet D3, based on casein, wheat germ, and cellulose (I. C. 95%)

Generation	Roa	Rmb	λc	Td	DTe
F1	45.019 (21.06–82.97)	0.117 (0.08–0.14)	1.124 (1.09–1.15)	32.405 30.26–34.92)	5.89 (4.80–7.89)
F4	45.164 (25.24–6.37)	0.099 (0.07–0.12)	1.104 (1.08–1.12)	38.395 (35.88–41.59)	6.984 (5.78–8.79)
F6	45.170 (21.41–93.11)	0.109 (0.08–0.15)	1.116 1.08–1.16)	34.689 (32.08–37.92)	6.310 (4.56–8.59)
F8	40.930 (17.40–83.02)	0.0833 (0.06–0.10)	1.086 1.06–1.10)	44.558 (42.43–46.96)	8.320 (6.75–12.01)

aNet reproductive rate.

bIntrinsic rate of increase.

cFinite growth rate.

dMean duration of one generation.

eDuration time. Parameters compared by bootstrap analysis (10,000 repetitions).

Discussion

In this study, we observed the development of T. absoluta from egg to adult on four artificial diets. The larval stage was the most critical in rearing T. absoluta, especially in the first days of development, as observed by Mihsfeldt and Parra (1999). High mortality was observed during the first 7 d after inoculation, this is frequently found on leaf–mining Lepidoptera, that usually present a mortality >32% in the first instar (Zalucki et al. 2002). Many factors influence this mortality when insects are reared on an artificial diet, such as the proportions of the different diets components and the physical conditions, as well as the water content, particle size of the mixture, and microbial agents, among others (Mihsfeldt and Parra 1999, Cohen 2004).

Diets containing lyophilized tomato leaves (D1 and D4) yielded low viability of young larvae (up to 7-d old). This ingredient was added as phagostimulant as stated previously by Mihsfeldt and Parra (1999). Although the tomato-leaf powder contains sugars and amino acids, which could stimulate the start of feeding (Awmack and Leather 2002), it is possible that secondary compounds that hinder the development of caterpillars were preserved by the lyophilization. For example some studies have not shown any benefits of adding these components compared with diets without plant material (Wang et al. 2014). The use of smaller amounts of the lyophilized product should be evaluated, in order to increase viability. Sugar, other of the essential component to the insect’s metabolism, can be used as a feeding stimulant (Cohen 2004, Glendinning et al. 2007) In this study, the diet with a higher sugar content (D2) did not result in the highest survival of this moth.

T. absoluta fed on a diet that contains cellulose (as in D3) provided the highest viability in young larvae (7 d after egg hatching) (Table 2). Cellulose is a polysaccharide that comprises 20–40% of the dry matter of herbaceous plants (Mcginnis and Kasting 1967). This ingredient does not provide nutrients to insects, but modifies the physical structure of the diet by making them consuming more dry matter in a way of a fibrous matrix. This was observed in Spodoptera eridania (Cramer) (Lepidoptera: Noctuidae), that increased its food intake on a diet rich in cellulose. The presence of cellulose in artificial diets for lepidopterans has been pointed out as producing a slower growth and lengthening the larval stage (Peterson et al. 1988, Timmins et al. 1988). This effect was observed for T. absoluta too, since when it developed in the D3 diet extended the larval period in 10 d compared with the natural diet. T. absoluta experimented an increased larval period (3–4 d) (Table 2) when fed on other artificial diets, which often occurs in insects reared on an artificial medium (Mendoza et al. 2016, Shen et al. 2006) and was previously observed by Mihsfeldt and Parra (1999) for T. absoluta, with a larval duration of 19 d in artificial diet. The pupal period of the insects developing on diet D3 was similar to that observed in the natural diet.

From this study, we could figure out that the best larval and pupal viability was obtained with a diet containing casein, wheat germ and cellulose. That fact suggests that this diet can supply the nutritional requirements of T. absoluta. The fertility was >85% for all the diets, and highest for D3 and D4, which were similar to the fertility rate obtained with the natural diet. High egg viabilities have previously been reported for this species (Bogorni and Carvalho 2006, Gonçalves-Gervásio et al. 1999 reared on both natural and artificial diets (Mihsfeldt and Parra 1999).

Interestingly, the diet that resulted better for T. absoluta larval and pupal viability, and fertility, yielded similar pupal weight to those from the D4 and natural diets, for both males and females, an indication that the nutritional conditions of the artificial diet produced similar patterns to the natural diet. Other positive aspect found in this study for the selected artificial diet D3 is that T. absoluta developed four larval instars on this diets, the same number as the insect has when it is reared on tomato leaves (Erdoğan and Babaroğlu 2014, Giustolin et al. 2002, Haji et al. 1988, Mihsfeldt and Parra 1999).

In the insect rearing process, a relevant aspect to be taken into account is the container’s features. The 1.0 × 6.0 cm flat glass tube, capped with water-repellent cotton, proved to be the most suitable for the development of T. absoluta. Small containers have been found better for rearing other Lepidoptera species as Helicoverpa armigera (Hübner) (Lepidoptera: Noctuidae) (Assemi et al. 2012). The smaller the container the higher surface-to-volume ratio (Garcia et al. 2006), and consequently, a greater water loss is expected. Water content is a determining factor in the quality of an artificial diet. In the case of T. absoluta, the dry surface favored its development, probably due to its feeding habits and a reduction of free water which limited the presence of pathogens. The physical characteristics of the diet are clearly important for T. absoluta, as reported by Mihsfeldt and Parra (1999).

The larval density per rearing container influenced the survival of T. absoluta. Thus, three larvae per tube yielded, larval and pupal viabilities >80 and 90%, respectively, meanwhile lower (one larva per tube) or greater (>3 larvae per tube) densities reduced, larval viability although without affecting the pupal viability (Table 5). The higher mortality at the higher density may be related to the increased humidity provided by larval excrement.

When assessing the biological parameters of T. absoluta in the selected diet, in generations F1, F4, F6, and F8, the length of the egg-adult period decreased with the generations (Table 6), from 34.67 d for the F1 to 31.79 d for the F8 which shows an adaptation to the artificial diet with the passing of the generations. Although the life cycle of this species is reported to be 22–30 d with the natural diet (Barrientos et al. 1998), the results obtained with this artificial diet are not far from these values. The major difference occurred in the larval period which lasted from 20 to 22 d, while in the natural diet lasts 14 d (Desneux et al. 2011, França et al. 2000, Urbaneja et al. 2013).

The viability 7 d after inoculation increased with the generations; while the larval and pupal viabilities were equal over the generations, with a total viability of 75% in the eighth generation, a higher viability than that found for diets previously used (Mihsfeldt and Parra 1999) (Table 6) and in the eighth generation, viability was 75%, which according to Singh (1983) can be considered suitable for rearing an insect in the laboratory.

The longevity of females in the first generation was 12 d, increasing to 15 d in the F6 generation. On a natural diet, this parameter ranged from 10 to 15 d (Estay 2000, Desneux et al. 2010), reaching up to 18 d (Estay 2000), 23 d (Bogorni and Carvalho 2006), or 31 d (Imenes et al. 1994). These values are the result of abiotic differences in experimental conditions of the various studies. Although in this study, the recorded longevity with the D3 diet were lower than some values obtained by these authors, it is evident that the insects adapted to laboratory conditions on the artificial diet selected.

The fecundity was higher than that reported when T. absoluta was reared on other artificial diets; Mihsfeldt and Parra (1999) obtained a mean number of 43 eggs per female. The number obtained in this study, from the fourth generation, was 117 eggs per female (Table 7). Under natural conditions the tomato pinworm can lay up to 260 eggs per female (Uchoâ-Fernandes et al. 1995) but there are records ranging from 55 to 130 eggs (Coelho and Franca 1987, Haji et al. 1988) up to 141 eggs per female (Erdoğan and Babaroğlu 2014).

The life-table parameters of insects on the D3 diet (Table 8) did not differ through the generations. The intrinsic rate of increase (rm) was lower in this study than rates found by other authors using a natural diet (Erdoğan and Babaroğlu 2014, Pereyra and Sánchez 2006). The net reproductive rate (Ro) and the finite rate of increase (λ) were similar to those reported by Erdoğan and Babaroğlu (2014) and Pereyra and Sánchez (2006) using a natural diet. This result indicates that the population growth is similar in both dietary conditions. The mean generation time (T) was similar to that reported by Mihsfeldt and Parra (1999). The DT was shorter for insects on the D3 diet compared with the time obtained by Mihsfeldt and Parra (1999), who reported a population DT of 11, while in this study the population doubled in five to eight generations.

In conclusion, an artificial diet based on wheat germ, casein and cellulose provides optimal nutritional conditions for rearing T. absoluta with a total viability of 75%, without affecting the development and reproduction of the insects for eight successive generations, based on life table parameters. A smaller container (a Ø1.0 × h 6.0 cm, flat glass tube) and a larval density of three larvae per rearing tube provided favorable physical conditions for the development of the tomato pinworm.