Transporting Tenebrio molitor Eggs: The Effect of Temperature, Humidity and Time on the Hatch Rate

: Insect farming can be an important tool in the realization of a more sustainable future. With a growing insect industry, animal transportation between and within farms is expected to increase. For Tenebrio molitor , using eggs may be convenient as it eliminates the risk of cannibalism, food shortages and has a low risk of asphyxiation. However, there are at least three important variables during regular transport: time, temperature and relative humidity. For each one, as well as their interactions, there is a need to assess the effect on the hatch rate and establish lower and upper boundaries at which a good hatch rate of the eggs is possible. In this study, a total of 20 temperature/RH combinations were assessed (between 5–40 ◦ C and 40–80% RH), with an exposure time ranging between 1 and 6 days for each combination. The results indicated that between 15 and 30 ◦ C no negative effects were observed at any of the assessed RH or exposure times. Higher temperatures did result in a lower hatch rate; no eggs hatched at 40 ◦ C, even after one day. Lower temperatures of 5 and 10 ◦ C can be endured without pronounced adverse effects but only for a limited time (resp. 1 or 2 days). Including relative humidity in the model did improve the overall ﬁt, but the effect is limited (compared to temperature or exposure time) with a slightly better hatch rate in dryer conditions at the extreme temperatures.


Introduction
It is expected that the insect industry will grow in the coming years. Furthermore, the industry is likely to evolve in a similar way as other farmed animal sectors (e.g., poultry and pigs) with specialized breeders and fatteners. This means that the frequency of transport, of insects in early life stages, will increase between or within companies. Except when using (expensive) climatized transport, the temperature and humidity may fluctuate significantly during transport depending on the season, location and weather conditions (e.g., irradiation of the sun on a car rooftop, [1]). It is therefore important to know that the environmental envelope (temperature and humidity) insects can survive without adverse effects. Furthermore, it is important to know how long the animals can survive in these conditions as they are frequently time dependent.
This study was focused on the yellow mealworm (Tenebrio molitor L. 1758) and specifically the transportation and storage of their eggs. Transporting eggs may be more advantageous than transporting larvae as they do not need to feed, reducing the risk of starvation or cannibalism. Furthermore, they most likely produce less emissions (e.g., CO 2 ) and consume less O 2 when compared to other life stages. This may be important as Greenberg and Ar [2], observed that prolonged exposure to 10% O 2 strongly decreases mealworm larvae survival. However, this is only hypothetical as there are no data available on the emissions of mealworm eggs. Finally, eggs cannot escape from a plastic container or chew their way out, making them a practical option for bulk transport between breeders and fatteners.
Overall, the current available knowledge of the influence of the environment on the hatch rate is limited and inadequate to predict their survival during transport. Punzo and Mutchmor [3] observed that eggs cannot complete their development at 10 • C at both low and high relative humidity (12 and 98% RH). Furthermore, at 25 and 35 • C the eggs are prone to dehydration when exposed to prolonged periods of low RH (12%) with mortality occurring after 11 to 14 days. More information is available on the other life stages of Tenebrio molitor and there the intricate relation between time, temperature and humidity is evident. Punzo and Mutchmor [3] observed that there was no significant change in survival at 25 • C, with a changing relative humidity between 12 and 98% for both young (<30 mg) and old mealworms (>100 mg). However, deviating from 25 • C (10-35 • C) resulted in significant mortality, especially when combined with dry or wet conditions. The timetemperature relation was also observed, with 50% of the mealworm larvae dying at 5 • C in 40 h, at −9 • C the time decreased to 10 h or less, and at −10 • C mortality occurred within 2 h only [4].
This study aimed to simulate a wide range of real-world transportation scenarios from express to delayed. Specifically, the eggs were subjected to various combinations of temperature (5-40 • C) and relative humidity (40-80% RH) for 1 to 6 days.

Colony Information
The mealworms used in this study have been bred at the Inagro Insect Research Centre since 2013. They are kept in 60 × 40 cm plastic crates (with an inner surface area of 2000 cm 2 ) at a temperature of 27 • C ± 1 • C SD, 60% ± 3% SD relative humidity and in the dark except during maintenance. The animals are fed ad libitum with INSECTUS Mealworm Grow (Mijten nv, Belgium) and fermented chopped chicory roots. The CO 2 concentration is monitored and kept below 1500 ppm. To collect enough eggs for the experiments, several crates are set-up with 250 g of beetles in each crate. The beetles, who were on average 4 weeks old, could lay eggs for 24 h in wheat flour (<0.5 mm). Thereafter, the eggs were collected by sieving on a 0.5 mm sieve.

Egg Transport Experiments
Fifteen plastic containers (10 × 8 × 3 cm lxwxh, with no lid), each with 2 g of freshly harvested eggs (originating from the same parental batch) and 20 g of wheat bran were made. Three containers were directly placed in the standard rearing conditions in the climate room (27 • C, 60% RH) to assess the baseline hatch rate (hereafter our control). The other 12 containers were placed in a climate chamber at the selected climate condition. After 1, 2, 3 and 6 days, three random replicates were removed from the climate chamber and placed in the main climate room at the standard rearing conditions together with the control (27 • C, 60 RH). The different exposure times were chosen based on estimated EU transport times and previous experiences: 1 day = national express transport, 2 days = international express transport, 3 days = regular transport, 6 days = delayed transport (within the EU). Longer times were not included as the eggs could start to hatch during the exposure.
The assessed climate conditions varied between 5 • C and 40 • C (in 5 • C increments) and 40 and 80% relative humidity (in 20% increments). In total, 20 temperature and humidity combinations were assessed ( Figure 1), with 4 exposure times and a control per combination and 3 replicates per treatment. The original goal was to assess all combinations in a full factorial design, but due to the limitations of the equipment, not all temperatures at 80% RH could be assessed. Additionally, these limitations occurred at the 5 • C and 40% RH combination. Finally, as only one climate chamber was available, the different temperature and humidity combinations were assessed in a staggered manner over time. A few combinations were assessed twice to assess changes over time. and 40% RH combination. Finally, as only one climate chamber was available, the differ ent temperature and humidity combinations were assessed in a staggered manner ove time. A few combinations were assessed twice to assess changes over time. The containers were visually checked every workday to assess if part of the eggs ha hatched. When no larvae were observed three weeks after they were transferred to th control conditions this temperature/humidity/time combination was considered 100% le thal to the eggs. If larvae were observed within the three-week period, the hatch rate wa not assessed until two weeks after the first larvae were seen to ensure all viable eggs ha the opportunity to hatch.
The hatch rate was determined by estimating the number of larvae in each crate vi subsampling. This was carried out by gently homogenizing the content of the crate an taking three samples of approximately one gram. The total number of larvae was counte in each sample. A minimum of 50 larvae was arbitrarily chosen as a cut-off value, if to few mealworms were present the subsample size was increased (the average sample siz was 74 larvae).

Statistical Analysis
A direct comparison between the number of larvae and the original number of egg is near impossible, as when the eggs are counted there is a high risk of damaging them and the average weight of the eggs is not always equal due to the accumulation of fras and flour on their sticky shell. The number of larvae was therefore compared to the contro of the same treatment and not between treatments.
Therefore, in this study, the hatch rate is defined as: the number of larvae per gram of eggs in the treatment divided by the number of larvae per gram of eggs in the contro Hence, a hatch rate of 100 implies that the treatment is performing as equally well as th control. Yet, it is important to note that, in this study, a hatch rate of 100 does not mea 100% of the eggs hatched just that they were equal to the average hatch rate of the contro The 95% confidence interval of the control was between 81 and 119. This indicates tha there is a ±19% (natural) variation in hatch rate between replicates of the control with th current experimental method. Therefore, a hatch rate between 80 and 120 was considere equivalent to the control.
The statistical analyses were performed using R 4.0.0. statistical software (www.r project.org). Multiple linear regression was not possible due to the non-linear nature o the response surface when all combinations are assessed simultaneously. Therefore, gen eralized additive modelling (GAM) was used to assess the results as this technique doe not assume an a priori linear relationship including the three parameters and their inter actions. This full model was reduced via backward selection to find the model with th The containers were visually checked every workday to assess if part of the eggs had hatched. When no larvae were observed three weeks after they were transferred to the control conditions this temperature/humidity/time combination was considered 100% lethal to the eggs. If larvae were observed within the three-week period, the hatch rate was not assessed until two weeks after the first larvae were seen to ensure all viable eggs had the opportunity to hatch.
The hatch rate was determined by estimating the number of larvae in each crate via subsampling. This was carried out by gently homogenizing the content of the crate and taking three samples of approximately one gram. The total number of larvae was counted in each sample. A minimum of 50 larvae was arbitrarily chosen as a cut-off value, if too few mealworms were present the subsample size was increased (the average sample size was 74 larvae).

Statistical Analysis
A direct comparison between the number of larvae and the original number of eggs is near impossible, as when the eggs are counted there is a high risk of damaging them and the average weight of the eggs is not always equal due to the accumulation of frass and flour on their sticky shell. The number of larvae was therefore compared to the control of the same treatment and not between treatments.
Therefore, in this study, the hatch rate is defined as: the number of larvae per gram of eggs in the treatment divided by the number of larvae per gram of eggs in the control. Hence, a hatch rate of 100 implies that the treatment is performing as equally well as the control. Yet, it is important to note that, in this study, a hatch rate of 100 does not mean 100% of the eggs hatched just that they were equal to the average hatch rate of the control. The 95% confidence interval of the control was between 81 and 119. This indicates that there is a ±19% (natural) variation in hatch rate between replicates of the control with the current experimental method. Therefore, a hatch rate between 80 and 120 was considered equivalent to the control.
The statistical analyses were performed using R 4.0.0. statistical software (www.rproject.org). Multiple linear regression was not possible due to the non-linear nature of the response surface when all combinations are assessed simultaneously. Therefore, generalized additive modelling (GAM) was used to assess the results as this technique does not assume an a priori linear relationship including the three parameters and their interactions. This full model was reduced via backward selection to find the model with the lowest Akaike information criterion (AIC). The latter measures the goodness of fit and model complexity.

Results
The GAM model could predict the variability reasonably accurately (Figure 2). The final model, including all predictor variables (temperature, relative humidity, time and interactions), had the lowest AIC indicating that all parameters had a significant effect on the hatch rate.
lowest Akaike information criterion (AIC). The latter measure model complexity.

Results
The GAM model could predict the variability reasonably final model, including all predictor variables (temperature, rel interactions), had the lowest AIC indicating that all parameters the hatch rate. The GAM model indicates that there is a broad climate env rate of the eggs is not particularly influenced by the temperatu 30 °C and 40-80% RH, Figure 3). Near the extremes (≤10 °C or ≥3 rate is observed with a strong interaction between time and lower temperatures. A short exposure to 5 °C (1 day) or 10 °C (2 hatch rate (especially in dry conditions), but prolonged exposur the hatch rate at these temperatures. However, there are still so days at 5 °C. On the other side of the spectrum, the first adver and at 40 °C none of the eggs hatched at any of the assessed expo of humidity is not as clear as the temperature relationship. Wit exposure, a lower relative humidity seems to improve the hatch peratures. It is however clear that, compared to temperature an The GAM model indicates that there is a broad climate envelope in which the hatch rate of the eggs is not particularly influenced by the temperature, humidity or time (15-30 • C and 40-80% RH, Figure 3). Near the extremes (≤10 • C or ≥35 • C), a decrease in hatch rate is observed with a strong interaction between time and temperature especially at lower temperatures. A short exposure to 5 • C (1 day) or 10 • C (2 days) does not reduce the hatch rate (especially in dry conditions), but prolonged exposure (≥3 days) does decrease the hatch rate at these temperatures. However, there are still some eggs hatching after 6 days at 5 • C. On the other side of the spectrum, the first adverse effects appear at 35 • C and at 40 • C none of the eggs hatched at any of the assessed exposure times. The influence of humidity is not as clear as the temperature relationship. With the exception of 6 days' exposure, a lower relative humidity seems to improve the hatch rate at more extreme temperatures. It is however clear that, compared to temperature and exposure time, the relative humidity has a much lower influence within the assessed range and experimental procedure (Figure 4).

Discussion
Using insects as food or feed may be a good way to improve the sustainability of the food/feed chain due to the lower use of natural resources compared to other livestock (e.g., land use) [5]. The results of this study may help when determining if it is possible to transport (or store) eggs at a certain climate and for how long. The effect of temperature is very clear. Between 15 and 30 degrees there is no issue in transporting the larvae for up to 6 days. According to Punzo and Mutchmor [3], mealworm eggs cannot complete their development at 10 °C. These results are confirmed here, but if only briefly exposed (1 or 2 days) to such a low temperature and thereafter returned to the control conditions, it is not detrimental for the further development of the eggs. If prolonged exposure to lower temperatures (<15 °C) is expected, transporting larvae may be more beneficial as they can be stored at 10 °C for up to 120 days [4] or express postal services should be used whether or not they are combined with a heat pack/insulation. It is unsurprising that the eggs do not survive at 40 °C, it is very similar to previously observed results for the larvae and beetles. Previous research indicates that larvae die at 42 °C after only 22.5 min (first instar) and mortality takes up to 104 min at instar 27 [3,6]. Allen et al. [7] observed that the upper and lower temperature limit, above and below which beetles were unable to move or perform coordinated movement, was 41.5 °C and 6.5 °C, respectively. The gradual decrease in hatch rate at the lower temperatures but abrupt changes at high temperatures are similar to what is frequently observed in insect development studies [8]. In general, for insects, the influence of temperature on the development time is not a normal distribution, but a logan (or similar distribution) with a tail towards the lower temperatures and a hard(er) cut-off at higher temperatures. The influence of humidity on the hatch rate remains unclear but also seems to be of lesser importance when compared to the other parameters.

Discussion
Using insects as food or feed may be a good way to improve the sustainability of the food/feed chain due to the lower use of natural resources compared to other livestock (e.g., land use) [5]. The results of this study may help when determining if it is possible to transport (or store) eggs at a certain climate and for how long. The effect of temperature is very clear. Between 15 and 30 degrees there is no issue in transporting the larvae for up to 6 days. According to Punzo and Mutchmor [3], mealworm eggs cannot complete their development at 10 • C. These results are confirmed here, but if only briefly exposed (1 or 2 days) to such a low temperature and thereafter returned to the control conditions, it is not detrimental for the further development of the eggs. If prolonged exposure to lower temperatures (<15 • C) is expected, transporting larvae may be more beneficial as they can be stored at 10 • C for up to 120 days [4] or express postal services should be used whether or not they are combined with a heat pack/insulation. It is unsurprising that the eggs do not survive at 40 • C, it is very similar to previously observed results for the larvae and beetles. Previous research indicates that larvae die at 42 • C after only 22.5 min (first instar) and mortality takes up to 104 min at instar 27 [3,6]. Allen et al. [7] observed that the upper and lower temperature limit, above and below which beetles were unable to move or perform coordinated movement, was 41.5 • C and 6.5 • C, respectively. The gradual decrease in hatch rate at the lower temperatures but abrupt changes at high temperatures are similar to what is frequently observed in insect development studies [8]. In general, for insects, the influence of temperature on the development time is not a normal distribution, but a logan (or similar distribution) with a tail towards the lower temperatures and a hard(er) cut-off at higher temperatures. The influence of humidity on the hatch rate remains unclear but also seems to be of lesser importance when compared to the other parameters.
Based on the overall data in this study, it can be concluded that shipping mealworm eggs without climate control is possible between 15 and 30 • C without compromising the hatch rate up to at least 6 days. Lower temperatures are possible, but only for a limited time and higher temperatures should be avoided. It has to be stressed that any potential sublethal effects (e.g., reduced growth rate, reduced fertility, etc.) were not assessed in this study but may limit the environmental envelope. Furthermore, quick variations in temperature or humidity were not assessed (day/night, sun vs. shade) and may possibly affect the hatch rate. Finally, although the experiment focussed on answering the environmental limits for transporting eggs of T. molitor, the results can be useful for other parts of the industry. For example, temperature and humidity fluctuations may be much larger outside the tightly controlled climate rooms used in research, for example, within the entomoponics concept (e.g., 14-36 • C, [9]).