Moths as potential pollinators in avocado (Persea americana) orchards in temperate regions

ABSTRACT

Avocado (Persea americana) flowers are primarily diurnal, however low minimum overnight temperatures can delay in the opening of female-phase flowers such that the flowers open late in the day and remain open overnight. To determine whether moths are important visitors to avocado flowers at night, we trapped moths in the Bay of Plenty, New Zealand, and examined the pollen moths carried. 11.5% of moths carried avocado pollen grains, and 50.4% of all moths were found to carry pollen grains from other plant species. Moth species diversity was highly variable between orchards, but Ichneutica mutans, Ichneutica ustistriga, Epyaxa rosearia, Rhapsa scotosialis, Phrissogonus laticostatus were observed at all sites and Ichneutica steropastis at three sites, with avocado pollen grains found on 10%–33% of these individual species. We suggest that these species may be the most important pollinating moth species because of their abundance, proportion of individuals carrying avocado pollen, and the number of pollen grains per individual. This exploratory study has shown that moths are capable of carrying avocado pollen grains and may be providing a background pollination service to complement that of honey bees (Apis mellifera). To better understand moth contribution to pollination, fruit set needs to be assessed.

KEYWORDS:

• Interactions
• nocturnal
• crop
• biodiversity
• agroecology
• pollen

Introduction

Globally, animals pollinate around 85% of angiosperm plant species and ∼33% of crops produced, with insects, particularly bees, playing a critical role in this process (Potts et al. 2010; Hahn and Brühl 2016). Declines in both wild and managed pollinators are correlated with declines in pollination services to plants, which has in turn led to uncertainty around the stability of ecosystems and food security (Potts et al. 2010; Ollerton et al. 2011; Hahn and Brühl 2016). As a result, there has been research focusing on alternative insect pollinators to European honey bees (Apis mellifera), particularly in crops, including: bumble bees (Bombus spp.; Rao and Stephen 2009; McBrydie et al. 2017; Cutting et al. 2018) and solitary bees (Broussard et al. 2011; Rader et al. 2012; Howlett et al. 2015), but also non-bee taxa such as flies (Howlett 2012; Rader et al. 2016; Stavert et al. 2018) beetles (Gottsberger 1989; Listabarth 1996), butterflies (da Silva Santos et al. 2020), and moths (Manning and Cutler 2013; Buxton et al. 2018).

Avocado (Persea americana, Lauraceae) is a crop grown commercially from low-latitude temperate regions through to tropical regions globally (Knight 2002). In its native Mexican range, the flowers are visited by bees, native wasps (Brachigastra mellifica), and flies (Chrysomya megacephala) (Pérez-Balam et al. 2012). Avocado has synchronous protogynous flowers that open first as functionally female before closing and opening on the following day as functionally male, preventing self-pollination within a flower (Evans et al. 2010; Pattemore et al. 2018). Avocado flowers are small, green, and contain a single stigma surrounded by 6 anthers, and offer both nectar and pollen as rewards. However, low overnight temperatures are associated with a delay in the start of the female flower phase, resulting in flowers remaining open overnight (Pattemore et al. 2018). Nocturnal pollination may therefore be more important in avocado orchards than previously considered, particularly in temperate regions where cooler temperatures during spring flowering are more common. Moths are potential pollinators of these late-opening flowers because of their predominantly nocturnal or crepuscular lifecycles (Pattemore et al. 2018). Recent work by Walton et al. (2020) has also demonstrated that nocturnal moth pollen transport networks can be more complex than diurnal species in agricultural landscapes. Despite some recent publications on the role of moths as pollinators, studies on non-Sphingid moths remain critically low on a global scale and are almost completely lacking from agricultural settings (Martins and Johnson 2009; MacGregor et al. 2014; Hahn and Brühl 2016; Buxton et al. 2018; Lu et al. 2021).

In 2017 alone, 1,089,222 tonnes of avocado were harvested from 133,040 ha of land in Australia, southern Brazil, Chile, New Zealand, South Africa and the USA (California), which all occur within the temperate zone (FAOSTAT 2019). The value of the avocado crop grown in these countries was US$1.1 billion in 2016 (gross production value, FAOSTAT 2019). Increasing avocado production in temperate regions may pose a challenge for pollination as more flowers would be open as functionally female at night, restricting access to the flowers by diurnal insects (Pattemore et al. 2018). Studies assessing nocturnal pollination in avocado orchards in temperate regions are thus required.

Typical fruit set in avocado orchards is 0.3% (Evans et al. 2010), for which 1–20 pollen grains are required per flower, while fruit set of 5% is achieved in flowers receiving 50 or more deposited pollen grains (Evans et al. 2010; Pattemore et al. 2020). As the pollen requirements for fruit set are not high, flowers may benefit from even infrequent or inefficient pollinating taxa that might carry fewer pollen grains than other taxa utilised for pollination like honey bees.

This study aims to address two questions: do moths carry avocado pollen? And which moth species are likely to be more important in avocado orchards for their role in pollination?

Materials and methods

Two Heath moth traps were positioned in each of four avocado orchards near Katikati, New Zealand, in spring 2017. Orchards were spatially separated by over 1 km from other avocado orchards included in the study. Automated heath moth traps with fluorescent bulbs which emit actinic light were used, turning on at sunset and off at sunrise. However, because of mechanical failure, some traps failed to catch moths on some nights. Egg cartons were placed within the traps to avoid pollen cross contamination between species. Traps were placed at least 25 m apart and in areas where light would not be obscured by vegetation to maximise the area sampled while minimising between-trap interference. Traps were set up in each orchard on 18, 24 and 31 October, and on 7 and 21 November 2017 as heavy nocturnal flowering was expected because of low overnight temperatures the previous night (Pattemore et al. 2018). Each trap was cleared the following morning and moths were euthanised in a –20°C freezer.

Each moth was swabbed with a 3 mm³ cube of gelatine-fuchsin (Kearns and Inouye 1993), by rubbing the cube along the antennae, patting each eye and each side of the proboscis, two dabs on the thorax and three dabs on the abdomen to ensure as much of the ventral side of moth bodies came into contact with the gel. One cube was used to swab one entire moth individual, and the cube was placed on a slide, melted, and examined under a microscope at 40× magnification for the presence of pollen. Avocado pollen grains were counted and identified visually with the aid of a reference avocado pollen slide and the total number of avocado pollen grains was averaged across individuals and individuals within a specific species. The presence of non-avocado pollen was noted but no attempts to count or identify these grains were made (i.e. if a moth was found to carry non-avocado pollen, this was recorded as ‘1’ regardless of how many pollen grains and pollen species were carried).

Moths were identified to the lowest taxonomic level possible. To examine if we adequately sampled nocturnal visitors to avocado orchards, we ran a sample-based rarefaction method (Chiarucci et al. 2008) using the function specaccum within the R package vegan (Oksanen et al. 2010). We used non-metric multidimensional scaling (NMDS) to determine the difference in moth communities between sites with the function metaMDS in the R-package vegan. Significance was determined using the function envfit with 1000 permutations.

To test whether different species of moths were more likely to carry avocado pollen or pollen from other plant species, and whether there were site-based effects, we ran two general linearised mixed models (GLMM) using glmer in R package lme4 (Bates et al. 2014), with P-values generated using the Satterthwaite method of denominator synthesis, implemented within the package lmerTest (Kuznetsova et al. 2015; this method results in non-integer degrees of freedom). For both models, the fixed effects were moth species and the orchard, with sample date as a random effect. The presence/absence of avocado pollen was the response variable in the first model, and the presence/absence of non-avocado pollen was the response variable in the second model. We used binomial error distributions for both models. Model selection was done by a stepwise removal of terms from the model and comparing AIC values. Both models were checked for over-dispersion.

Results

Moth community

A total of 929 individual moths were caught over the five trapping nights, encompassing 51 species (Table 1). Orchards 1 and 3 had the greatest abundance and the greatest species diversity (309 and 294 individuals from 33 and 32 species, respectively) while Orchards 2 and 4 had the lowest abundance and species diversity (194 and 132 individuals from 24 and 18 species, respectively). 22 species (43%) were caught only once, and 36 (70%) of the species were caught less than 10 times. The six most abundant species were Ichneutica mutans (formerly Graphania mutans (Hoare 2019), n = 273), Epyaxa rosearia (n = 179), Pseudocoremia suavis (n = 55), Izatha churtoni (n = 40), Ichneutica insignis (formerly Graphania insignis (Hoare 2019), n = 38) and Eudonia submarginalis (n = 36) (Table 2).

Table 1. Species abundance, diversity and number of avocado pollen grains carried at each orchard.

Orchard	Number of moths caught	Number of moth species	Number of avocado pollen grains carried
Orchard 1	309	33	84
Orchard 2	194	24	60
Orchard 3	294	32	72
Orchard 4	132	18	19
Total	929	51	235

Table 2. Summary details for moth species suggested as being the most important for the pollination of avocado flowers.

Moth species	Moth family	Total abundance	Total number of avocado pollen grains	Percent individuals carrying avocado pollen	Average number of grains per individual	Standard error
Ichneutica mutans^	Noctuidae	273	91	12.09	0.33	0.089
Epyaxa rosearia^	Geometridae	179	31	10.06	0.17	0.064
Rhapsa scotosialis^#	Erebidae	34	19	14.71	0.56	0.316
Ichneutica steropastis^#	Noctuidae	27	14	18.52	0.52	0.294
Ichneutica ustistriga*	Noctuidae	9	4	33.33	0.44	0.242
Phrissogonus laticostatus*	Geometridae	8	7	25	0.88	0.743

The * symbol indicates moth species we suggest as being potentially the most effective pollinating moth species at the individual level (percentage of individuals carrying pollen and average number of grains carried). The ^ symbol indicates moth species we suggest as being important pollinating species collectively as a taxa (abundance and total pollen grains). The ^# symbol indicates moth species that may also be key pollinating species because of their relative abundance and average number of grains per individual.

The results of the NMDS showed that the moth communities at each orchard were significantly different (P = 0.002, Figure 1) and the rarefaction method determined that the moth community in the four orchards collectively was likely to be sufficiently sampled, but not at each site individually (Figure 2).

Figure 1. Non-metric multidimensional scaling plot of moth communities at each sample point. Crosshairs represent the community composition at different sites, while the coloured rings are standard error ellipses from the centroid for each site. The axes are arbitrary.

Figure 2. Rarefaction method on moths sampled at each orchard combined (A), at Orchard 1 (B), Orchard 2 (C), Orchard 3 (D) and Orchard 4 (E). Graph A shows further sampling efforts are unlikely to result in the trapping of previously un-trapped species.

Avocado pollen

A total of 235 avocado pollen grains was removed from moth bodies (Table 1). We found no effect of site or species on the number of pollen grains each moth was predicted to carry, but this may be because of the low power of our analysis as a result of the small sample size for some taxa. Between 9% and 13% of all moths caught carried avocado pollen across the 4 orchards, and 19%–23% of moths carrying any pollen (Figure 3). Of the moths carrying any pollen, between 15% and 18% carried both avocado and other pollen and between 2% and 6% of moths carried only avocado pollen. For individuals carrying avocado pollen, an average of between 1.6 and 2.8 avocado pollen grains were carried in the four orchards, with 2.2 grains being carried on average overall. Twelve moths carried 5 or more avocado pollen grains. There was no significant difference in the number of avocado pollen grains carried per individual per species and as such the results of the model are not presented here. While not significant, Phrissogonus laticostatus, Rhapsa scotosialis and Ichneutica steropastis (formerly Tmetolophota steropastis (Hoare 2019)) carried the highest number of avocado pollen per individual (0.88, 0.56 and 0.52 grains per individual, respectively).

Figure 3. The proportion of moths carrying avocado pollen (blue), the proportion of moths carrying non-avocado pollen grains (black), and the average number of avocado pollen grains removed from moths that carried avocado pollen at each orchard for moths that carried at least one pollen grain.

Non-avocado pollen

Between 45.3% and 61.6% of individuals caught carried non-avocado pollen grains in the four orchards, with 50.4% of individuals carrying non-avocado pollen in total (Figure 3). The results of the model showed that five moth species carried significantly more non-avocado pollen and that Orchard 2 had a higher proportion of moths carrying non-avocado pollen than the other orchards (Table 3).

Table 3. General linearised mixed models summary statistics for the presence of non-avocado pollen grains on moth species at the four orchards examined.

	Estimate	Standard error	Z-value	P-value
(Intercept)	0.73566	0.44191	0.665	0.09597***
Bityla defigurata	−0.94728	0.76965	−1.231	0.2184
Chloroclystis filata	0.23311	0.15295	0.069	0.28483
Declana floccosa	−0.41605	0.55942	−0.744	0.45704
Epyaxa lucidata	−1.75556	0.73183	−2.399	0.01645*
Epyaxa rosearia	−0.57014	0.45863	−1.243	0.21383
Eudonia octophora	−2.34339	0.18115	−1.984	0.04726*
Eudonia submarginalis	−1.23184	0.55527	−2.218	0.02652*
Ichneutica insignis	−0.95228	0.55389	−1.719	0.08557***
Ichneutica lignana	−0.01247	0.80592	−0.016	0.98766
Ichneutica mutans	−0.72426	0.44816	−1.616	0.10608
Ichneutica plena	−0.90359	0.58929	−1.533	0.12519
Ichneutica ustistriga	−0.17161	0.83507	−0.206	0.83717
indet.	−0.82141	0.81709	−1.005	0.31476
Izatha churtoni	−1.55143	0.55816	−2.78	0.00544**
Leucania stenographa	−1.09317	0.01962	−1.072	0.28366
Meterana exquisita	−1.22059	0.62877	−1.941	0.05223***
Phaeosaces coarctatella	−1.78055	0.7276	−2.447	0.0144*
Phrissogonus laticostatus	−1.50798	0.86358	−1.746	0.08078***
Pseudocoremia suavis	−0.63993	0.51837	−1.234	0.21701
Rhapsa scotosialis	−0.45238	0.56593	−0.799	0.42409
Ichneutica steropastis	−1.01099	0.59458	−1.7	0.08907.
Orchard1	0.03327	0.20861	0.16	0.87329
Orchard2	0.52088	0.18868	2.761	0.00577**
Orchard3	−0.26722	0.23217	−1.151	0.24974

Note: The intercept condition is Agrotis ipsilon at Orchard 4. * = p < 0.05, ** = p < 0.01, *** = p < 0.1.

Discussion

Pattemore et al. (2018) showed that nocturnal pollination could be important for avocado flowers grown in temperate regions with variable climatic conditions during the flowering period. Pattemore et al. (2018) found some evidence that Lepidoptera may play an important role in the pollination of avocado flowers, but suggested coleopteran, dipteran and neuropteran individuals are likely the most important. The results of our study conforms to Pattemore et al. (2018) and suggests that moths may indeed contribute to pollination within avocado orchards, as 11.5% of individuals caught carried an average of 2.2 avocado pollen grains with 1 pollen grain being sufficient to achieve fertilisation. While the amount of pollen on moth bodies was low, the pollen loads of honey bees in avocado orchards are also low with the majority of individuals having no pollen at all (Pattemore et al. 2020). It is important to note that the pollen loads recorded on honey bees are also likely to be an overestimation as not all pollen on a bee is available for pollination due to location on the body (Ish-Am and Eisikowitch 1993; Pattemore et al. 2020). In comparison, the pollen loads on moth bodies are likely an under-representation as swabbing is not always effective at removing pollen grains from moth bodies due to pollen being often located on curled mouthparts and in amongst dense scales (MacGregor et al. 2019). In addition to being beneficial to the avocado orchard itself, moths could be important for the plants in and around avocado orchards; Hahn and Brühl (2016) suggested that while moths may not be essential to crop pollination, they can be important for non-crop plants in orchards and our results suggest moths may be contributing to the pollination of wild plants in and around avocado orchards. While the presence of pollen on insect bodies can be used as an indicator of a plant-pollinator relationship, and while this is useful information it should not be used as empirical proof that pollination actually occurs and be interpreted with care (King et al. 2013; Buxton et al. 2018).

The rate of floral visitation is an important aspect of pollination success and in assessing how an individual taxa contributes to pollination, but the presence and amount of pollen on an individual’s body is also important when making these assessments (Vázquez et al. 2005; Rader et al. 2009). Both of these two approaches are commonly used when assessing moths as pollinators (MacGregor et al. 2014; Buxton et al. 2018). When looking at the proportion of individuals that carried avocado pollen and the average number of pollen grains carried per individual, species such as I. ustistriga and P. laticostatus may be the most important because they had the highest percentage of individuals carrying pollen and a relatively high number of grains carried per individual. Species such as R. scotosialis and I. steropastis may also be important pollinating moth species in avocado orchards because of their relative abundance, relatively higher proportion of individuals carrying avocado pollen, and relatively higher number of pollen grains per individual (see Table 1). Unsurprisingly, the most abundant moth species also carried the greatest collective number of avocado pollen grains. It is important to note that the most abundant moth species are likely to be the most important pollinating taxa within the orchard space; I. mutans and E. rosearia as aggregate species carried the most pollen. There was no statistical difference between species because of very high variance and small sample sizes for some taxa, however, we propose that the most important pollinating moth taxa may be I. mutans, I. ustistriga, E. rosearia, R. scotosialis, I. steropastis, and P. laticostatus based on a combination of their abundance, the proportion of individuals carrying avocado pollen, and the number of pollen grains carried per individual.

Avocado is a single-seed crop, which means that the deposition of one pollen grain is sufficient for fertilisation and fruit to develop, although flowers can benefit when receiving up to 20 or even 50 avocado pollen grains (Evans et al. 2010; Pattemore et al. 2020). Honey bees in New Zealand and Australian avocado orchards had poor pollen loads (47.1% and 73.7% of bees had fewer than 5 pollen grains respectively) and these poor pollen loads have been attributed to poor fruit set (Pattemore et al. 2020). While pollen loads for moths was also low, for species such as R. scotosialis and I. steropastis, one pollen grain can theoretically be deposited every second visit. Moth visitation rates to avocado flowers are unknown, but a study on how moths visit flowers of Silene stellata found that moths visit multiple flowers per plant at a rate of 0.93 ± 20 plants an hour (Reynolds et al. 2009). In our theoretical scenario presented here one individual visiting 10 flowers on one panicle per hour over an 8-hour period (40 flowers per night) is not unreasonable and could see one individual fertilising 20 flowers per night. For the purpose of estimations, we can assume that the traps were sampling the moth community within a 25 m² radius before the light gets obstructed by vegetation, however, the trapping radius of moth traps is often estimated to be less than 25 m so our estimations here are being overly cautious (Van Grunsven et al. 2014). Assuming there is one R. scotosialis or I. steropastis individual per 25 m² fertilising 20 flowers in a 20 ha orchard this could result in 800–1600 flowers being fertilised per night. Given that the minimum nightly catch per trap was 10 moths and the maximum was 113, this could represent 16,000–181,000 fertilised avocado flowers per night. This amount could potentially be greater if the moth species has a faster flower handling time–some moth species spend as little as 5–20 sec handling flowers (e.g. hawkmoths visiting Nicotiana attenuata, Haverkamp et al. 2018), quite a bit less than our conservative 12 min over an 8-hour night estimate. While we have limited data and the above scenario is a theoretical example, our study suggests that moths may be providing valuable supplemental pollination when flowers are open at night, especially in the absence of competition with honey bee foragers.

Quantifying pollination services from moths at an ecosystem or community level is not yet possible to do in New Zealand. This is because of the lack of knowledge about the moth community in many parts of New Zealand, the fact that many taxa are still to be described, and the lack of knowledge about the effects of agricultural practices on New Zealand moth species (e.g. insecticide use). Moths are not typically considered to be primary pollinators in agro-ecosystems and are instead believed to act as crop co-pollinators and as pollinators for other plants within these agro-ecosystems (Hahn and Brühl 2016). Our findings agree in part–moths may be playing a role as co-pollinators for avocado and in the pollination of other plants within the orchard, but we also suggest that moths may have a greater contribution to crop pollination in general than is currently considered (see Devoto et al. 2011; Cutler et al. 2012 for further examples) and that further research on their contribution to the pollination of crop plants is necessary.

Conclusion

Moths are capable of carrying avocado pollen, and as such could be providing an important service in early-spring flowering in temperate climates such as New Zealand’s Bay of Plenty where this work was conducted. Moths are also likely contributing to the pollen movement of other plant taxa, which may be beneficial to native bush remnants and plantings within orchards. While different species vary in their importance as pollinating species, we propose that I. mutans, I. ustistriga, E. rosearia, R. scotosialis, I. steropastis, P. laticostatus could be the most important moth species in avocado orchards because of either their relative abundance, greater pollen loads per individual, or a combination of the two factors. Furthermore, the pollen found on moth bodies is likely to be an under-representation of what they actually carry and, as such, future work should assess pollen deposition by common and efficient moth taxa identified here, and compared with other nocturnal flower visiting taxa.

Acknowledgements

We would like to thank James Sainsbury and Milena Janke for their contribution to planning and lab work, and the avocado growers for allowing this research to occur on their property.

Disclosure statement

No potential conflict of interest was reported by the author(s).

Additional information

Funding

This work was funded through the Blue Skies programme of The New Zealand Institute for Plant and Food Research Limited.