Notes on the Nest Architecture and Colony Composition in Winter of the Yellow-Legged Asian Hornet, Vespa velutina Lepeletier 1836 (Hym.: Vespidae), in Its Introduced Habitat in Galicia (NW Spain)

Fifteen years ago, at least one multimated female yellow-legged Asian hornet (Vespa velutina Lepeletier 1836) arrived in France, giving rise to a pan-European invasion, altering the environment, affecting ecosystem processes, and impacting society. During winter, V. velutina nests (n = 3) were collected in Galicia and data on internal and external aspects of the nests and the colony as a whole were collected. The whole colony population (WCPN; adult insects, larvae, and pupae in percentages) was as follows: nest A: n = 176 (49%, 3%, and 48%); nest B: n = 1979 (52%, 36%, and 12%); and nest C: n = 662 (5%, 27%, and 8%). The adult insect population (IAPN; males, workers, and gynes in percentages) was as follows: nest A: n = 87 (11%, 66%, and 23%); nest B: n = 1021 (3%, 62%, and 35%); and nest C: n = 430 (20%, 73%, and 7%). As a small number of queens is sufficient for a population to develop, it is necessary to avoid continued spread by deactivating and removing all nests, even those detected in winter. This practice can be of greatest importance in border areas where V. velutina is expanding into new territory.


Introduction
Galicia, located on the western end of the Iberian Peninsula (Europe), is a considerable distance for an insect to travel from Asia, the Americas, or Africa. Despite this, a number of species have migrated in the past 15 years and successfully colonized and spread, resulting in a broad range of consequences to recipient ecosystems and, thereby, human society [1]:

•
The Asian chestnut wasp (Dryocosmus kuriphilus Yasumatsu 1951 (Hymenoptera: Cynipidae)) was detected in 2004. It has caused significant losses in chestnut production and weakened vitally important chestnut agroforestry cultivars.

•
The potato flea (Epitrix similaris Gentner 1944 (Coleoptera: Chrysomelidae)) and the Guatemalan moth (Tecia solanivora Povolny 1973 (Lepidoptera: Gelechiidae)) were detected in 2009 and 2015, respectively. Their presence has important implications, and the growth of the potato is now forbidden in some areas-a plan implemented by an invasive species task force.

•
The red palm weevil (Rhynchophorus ferrugineus Olivier, 1790 (Coleoptera: Curculionidae)) was detected in 2013. The species actively destroys the interior of the palm tree.
the number of Hymenoptera stings [21]. In China, however, the Asian hornet is considered to be a dangerous predatory species. During the summer of 2013, hornet attacks were the cause of 42 human deaths and 1675 injuries in three cities in Shanxi Province. In-hospital mortality of 1091 sting victims was 5.1% from 2009 to 2011 in Hubei Province, a surprisingly high mortality rate compared to other poisonous animals in a single region [22]. In South Korea from 2010 to 2014, there were 483,233 calls requesting removal of wasp nests and Hymenoptera stings caused 78,860 injuries and 49 deaths, with Polistes rothneyi koreanus Vecht and V. velutina the most prevalent sources. Total medical costs associated with wasp stings over a five-year period were approximately US$3.2 million [23]. All this information could signal that V. velutina will be a major problem in cities in the future. The patterns, processes, and impacts of urban invasions differ significantly from invasions in other environments, posing unique and increasingly complex challenges [24]. Preliminary data on V. velutina management demonstrate that it has settled in Galician cities. In 2015, three V. velutina nests were destroyed in Santiago de Compostela, the capital of Galicia, which has an area of 220 km 2 and a population of 95,966. By the end of 2017, a total of 971 nests had been destroyed. Handling and destruction of V. velutina nests accounted for 52.47% of total calls to the city Corps of Firefighters, more than for any other issue [20]. Major cities would benefit from preparation of an action plan. Several strategies to manage V. velutina vary in their effectiveness and there has been little success to date to restrict its invasion and economic, ecological, and social impacts [25]. Methods for detecting and destroying V. velutina nests have to be prioritized in terms of performance [25,26].
Many biological factors critical in assessing management strategies for V. velutina are still unknown [27], specifically, more knowledge about the ecology, phenology, behavior, and life history. The goal of this study was to describe the nest architecture and colony composition of V. velutina nests in winter, the first such study conducted during January within its introduced or native range.

Sampling Area
Three nests were removed on 2, 5, and 11 January 2019. All nests were located in the municipality of Nigrán (Galicia, Spain) and were labeled and identified as A (42 • 06 54.8" N, 47 02.6" W), B (42 • 06 47.0" N, 8 • 47 49.7" W), and C (42 • 08 48.2" N, 8 • 48 37.7" W), respectively ( Figure 1). Nigrán is a 35 km 2 municipality and contains the metropolitan area of Vigo, the most populous city in Galicia with a population of approximately 18,000. It is surrounded to the west by the Vigo estuary and the Atlantic Ocean and has a temperate climate and a greatly varying orography, including wide valleys. The highest altitude is approximately 500 m. The coast is composed of beaches and mountainous forests. Two rivers flow into the coast of Nigrán: the Miñor and the Muiños, creating an estuary that turns into marsh. Accidents are common and reported in the local press. Although they caused three deaths in Galicia in 2018 and one death in Nigrán in 2019, there is a general lack of data regarding the health risks of V. velutina stings. Poison Control Centers in France have noted that the increase in the Asian hornet population in the southwestern region of the country has not thus far resulted in an increase in the number of Hymenoptera stings [21]. In China, however, the Asian hornet is considered to be a dangerous predatory species. During the summer of 2013, hornet attacks were the cause of 42 human deaths and 1675 injuries in three cities in Shanxi Province. In-hospital mortality of 1091 sting victims was 5.1% from 2009 to 2011 in Hubei Province, a surprisingly high mortality rate compared to other poisonous animals in a single region [22]. In South Korea from 2010 to 2014, there were 483,233 calls requesting removal of wasp nests and Hymenoptera stings caused 78,860 injuries and 49 deaths, with Polistes rothneyi koreanus Vecht and V. velutina the most prevalent sources. Total medical costs associated with wasp stings over a five-year period were approximately US$3.2 million [23]. All this information could signal that V. velutina will be a major problem in cities in the future.
The patterns, processes, and impacts of urban invasions differ significantly from invasions in other environments, posing unique and increasingly complex challenges [24]. Preliminary data on V. velutina management demonstrate that it has settled in Galician cities. In 2015, three V. velutina nests were destroyed in Santiago de Compostela, the capital of Galicia, which has an area of 220 km 2 and a population of 95,966. By the end of 2017, a total of 971 nests had been destroyed. Handling and destruction of V. velutina nests accounted for 52.47% of total calls to the city Corps of Firefighters, more than for any other issue [20]. Major cities would benefit from preparation of an action plan. Several strategies to manage V. velutina vary in their effectiveness and there has been little success to date to restrict its invasion and economic, ecological, and social impacts [25]. Methods for detecting and destroying V. velutina nests have to be prioritized in terms of performance [25,26].
Many biological factors critical in assessing management strategies for V. velutina are still unknown [27], specifically, more knowledge about the ecology, phenology, behavior, and life history. The goal of this study was to describe the nest architecture and colony composition of V. velutina nests in winter, the first such study conducted during January within its introduced or native range.

Sampling Area
Three nests were removed on 2, 5, and 11 January 2019. All nests were located in the municipality of Nigrán (Galicia, Spain) and were labeled and identified as A (42°06′54.8″N, 47′02.6″W), B (42°06′47.0″N, 8°47′49.7″W), and C (42°08′48.2″N, 8°48′37.7″W), respectively ( Figure 1). Nigrán is a 35 km 2 municipality and contains the metropolitan area of Vigo, the most populous city in Galicia with a population of approximately 18,000. It is surrounded to the west by the Vigo estuary and the Atlantic Ocean and has a temperate climate and a greatly varying orography, including wide valleys. The highest altitude is approximately 500 m. The coast is composed of beaches and mountainous forests. Two rivers flow into the coast of Nigrán: the Miñor and the Muiños, creating an estuary that turns into marsh.

Nest Collection
All of the nests ( Figure 2) were detached and removed during the night and a quick-drying foam was used to seal the nest entrance. Smaller branches of the tree that had become part of the nest were removed and height of the nest (H N ) from the ground was recorded. The nests were sealed using two bags, 105 × 85 cm in size, and the main branch of the tree to which the nest was secured was manually cut using a hand saw. The nests were placed in a freezer for 12 h at a temperature of −18 • C.  (2) Burela. The right image shows the sampling region (Nigrán).

Nest Collection
All of the nests ( Figure 2) were detached and removed during the night and a quick-drying foam was used to seal the nest entrance. Smaller branches of the tree that had become part of the nest were removed and height of the nest (HN) from the ground was recorded. The nests were sealed using two bags, 105 × 85 cm in size, and the main branch of the tree to which the nest was secured was manually cut using a hand saw. The nests were placed in a freezer for 12 h at a temperature of −18 °C.

Nest Structural Features
Various external and internal measurements were taken of nests A-C to describe their form and structure. The measurements were carried out using a Vernier caliper, a ruler, and measuring tape.

•
The overall length (LN), width (WN), and circumference (CN) of the nests were recorded. The volume of the nest (VN) was calculated using the formula of a prolate spheroid: Tree branches inside the nest were denoted as main (MB) and secondary (SB) branches. The number of secondary branches (NSB), length of the main branch covered by the nest (LMBC), and the diameter of both main and secondary branches (DMB and DSB) were recorded.

•
After external measurements and inspection, the nests were dissected. The number of combs (NC), with their maximum (LC-max) and minor (LC-min) lengths, were recorded and the number of pillars (NP) that connected the combs was recorded. The presence of upholstered cells (UHCs) was recorded in each comb from all nests. Surface of the combs of the nest (TSN) and the surface of the combs (SC) were calculated using the formula for the area of an ellipsoid: The total cells (TCC) of a given comb and the total cells of a nest (TCN) as ∑TCC were calculated using the following formula: • where "n" is the number of cells counted along its longest diameter [25].

Nest Structural Features
Various external and internal measurements were taken of nests A-C to describe their form and structure. The measurements were carried out using a Vernier caliper, a ruler, and measuring tape.

External
The overall length (L N ), width (W N ), and circumference (C N ) of the nests were recorded. The volume of the nest (V N ) was calculated using the formula of a prolate spheroid: Tree branches inside the nest were denoted as main (M B ) and secondary (S B ) branches. The number of secondary branches (N SB ), length of the main branch covered by the nest (L MBC ), and the diameter of both main and secondary branches (D MB and D SB ) were recorded.

Internal
After external measurements and inspection, the nests were dissected. The number of combs (N C ), with their maximum (L C-max ) and minor (L C-min ) lengths, were recorded and the number of pillars (N P ) that connected the combs was recorded. The presence of upholstered cells (UHCs) was recorded in each comb from all nests. Surface of the combs of the nest (TS N ) and the surface of the combs (S C ) were calculated using the formula for the area of an ellipsoid: The total cells (TC C ) of a given comb and the total cells of a nest (TC N ) as TC C were calculated using the following formula: where "n" is the number of cells counted along its longest diameter [25].

Nest Colony Composition
The number of adult individuals was counted manually, obtaining the total insect adult population of each nest (IAP N ). Females were differentiated from males by the presence of 12 segmented antennae (males have 13). Males have an apex of the last sternite bilobate which is sharp in females. The difference between workers and gynes was established at a mesoscutum width of 4.5 mm [28].
The whole colony population of each nest (WCP N ), that is, adult insects and nonmature (larvae and pupae), present in different combs of the nests was recorded. Later, IAP N were separated into males, workers, and gynes.
The theoretical adult production (TAP N ) of individuals of each nest was inferred using the following function or method function [26]: where TAP N is the number of individuals produced by the colony and L C-max is the diameter of the largest comb.

External Nest Structure
All nests were suspended from trees at heights of 11, 12, and 10 m for nests A-C, respectively. Nests A-C were found in tree species Alnus glutinosa, Salix alba, and Salix babylonica, respectively. In all nests, a branch of the tree was identified as being inside the nest, covered by it to varying lengths and serving as the main nest support. Nests B and C had additional secondary branches, six and four, within the top structure. The length (L N ), circumference (C N ), width (W N ), volume (V N ), and height (H N ) from the ground of all nests are shown in Table 1. Nest A was the smallest and had a spheroidal shape with a volume of 9.7. Nests B and C had more of a pear shape, with calculated volumes of 101.8 and 57.0, respectively.
A detailed inspection of nest A revealed it was completely covered with various fragile leaves. Nest A was generally brown in color with strips of alternating lighter and darker shades, indicating that various sources of material were used in its construction. Nests B and C were externally covered mostly by spit pulp and only sheets of paper covering the bottom of these nests were found. The outsides of nests B and C were more durable than nest A, as they did not crumble with pressure upon handling. Nests B and C were externally a bright and homogenous tone. Nest C externally had a slight "varnish" or shell-like effect and possessed a distinct beehive aroma. It appeared nest A was formed using large sheets of paper, yet nests B and C only had this laminar structure in the lower part of the nest, the remaining parts externally covered with sculpted or stuck pulp on the surface.

Internal Nest Structure
Nest A was soft with a friable envelope consisting of different numbers of sheets (between 6 and 10) overlapping, covering the inner chamber formed by the combs. Nests B and C possessed a rigid Insects 2019, 10, 237 6 of 17 structure with hard and fibrous envelopes. Only in the lower portion was it possible to manually remove the sheets (between four and six) wrapping this part of nests B and C.
Once the outer cover was removed from all nests, mostly found were (a) a conical top or roof, with lengths of 8, 25, and 21 cm for nests A-C, respectively, and (b) the sets of the combs that made up the internal chamber, with four, eight, and nine combs found in nests A (Figure 3), B (Figure 4), and C ( Figure 5), respectively. Once the outer cover was removed from all nests, mostly found were (a) a conical top or roof, with lengths of 8, 25, and 21 cm for nests A-C, respectively, and (b) the sets of the combs that made up the internal chamber, with four, eight, and nine combs found in nests A (Figure 3), B (Figure 4), and C ( Figure 5), respectively.  (1)(2)(3)(4) in the same order that they were found within the nest.
In nest A (Figure 3, 1s), it was noted that the first comb was directly joined to the support branch of the tree, having a smaller roof cone surrounding it. In the other nests (B and C), the first combs did not attach directly to the branch (Figure 4, 1s and Figure 5, 1s).
There was an absence of UHCs in all combs of nest A, but UHCs were present in nest B (combs 1-6) and nest C (combs 1-5 and 8).
There was a free internal space that laterally separated the internal part of the outer cover of the nest and the internal chamber of the combs, between 3 and 5 cm. However, in nest C, this space was not always present, establishing several connections between the inner wall of the outer cover and the outer edges of the combs ( Figure 5, combs 2-4).  (1)(2)(3)(4) in the same order that they were found within the nest.
In nest A (Figure 3, 1s), it was noted that the first comb was directly joined to the support branch of the tree, having a smaller roof cone surrounding it. In the other nests (B and C), the first combs did not attach directly to the branch (Figure 4, 1s and Figure 5, 1s).
There was an absence of UHCs in all combs of nest A, but UHCs were present in nest B (combs 1-6) and nest C (combs 1-5 and 8).
There was a free internal space that laterally separated the internal part of the outer cover of the nest and the internal chamber of the combs, between 3 and 5 cm. However, in nest C, this space was not always present, establishing several connections between the inner wall of the outer cover and the outer edges of the combs ( Figure 5, combs 2-4).
Inside nest C (Figure 6), we found pollen loads in a pocket of the nest envelope. Unfortunately, we could not conclude if this was accidentally included in the colony due to tree growth and packing by insects or if it was intentionally recovered and stored by the hornets.
Total surface of the combs for each nest, in decreasing order, was nest B > nest C > nest A. Nest B contained the largest number of combs (nine combs), followed by nest C (eight combs), and A (four combs). In nest C, comb 5 had the greatest length (L C-max = 36.1 cm) of all combs of the three nests studied ( Table 2).
The theoretical adult production (TAP N ) was 1129 for nest A, 7492 for nest B, and 7926 for nest C. The total cells of the nest (TC N ) was 1074, 9962, and 7784 for nests A-C, respectively.

Discussion
Approaching active colonies of Vespidae in general, and V. velutina in particular, is complicated and potentially risky. Vibrations, impacts, or panicked movements near a colony may be perceived as threats, provoking a defensive attack. However, V. velutina colonies reportedly show unsolicited assault behavior and the recommended minimum distance to observe a colony is several meters [29,30]. The observation of Vespidae nest structure is quite difficult, as one cannot generally study a given nest thoroughly using nondestructive methods. Details regarding seasonal modifications need to be obtained by breaking open nests [26].
Study nests were collected in Nigrán, 25 km away from O Rosal, one of the first two original entry points of V. velutina detected in Galicia in 2012. The nests represented the seventh annual known generation of V. velutina in Galicia. The collection protocol ensured complete sampling of the whole colony without using pesticides or otherwise damaging the nest. In previous studies, nests were obtained during the day; however, as V. velutina does not possess nocturnal habits, such as the European Vespa crabro, this enabled nighttime collection to include all adult individuals present.
Study nests were located in trees at heights greater than 10 m. V. velutina is a well-known aerial-nesting species within and outside its native range but also displays a broad and flexible nest site preference, introducing difficulty in its detection and destruction. They can build nests in trees at high altitudes (between 10 and 40 m) in trees in urban areas and also at lower altitudes in fruit trees, shrubs, and bushes in cities, rural areas, and gardens. It takes advantage of human construction to settle at high altitudes in buildings, on electricity pylons, houses, and lamp posts or at lower heights (less than 5 m) in construction interiors (attics and ceilings of barns and warehouses) [29].
A geo-referenced map of V. velutina nests in France revealed that of 6073 nests, 48.5% were observed in artificial and 42.25% in agricultural surroundings. Less than 10% occurred in natural areas (8.1%) and wetlands (1.1%). The majority (70.0%) were located more than 10 m above the ground, 26.3% were between 2 and 10 m above the ground, and only 3.7% were situated <2 m above the ground. In addition, 87.0% of the 3296 nests were built in vegetation (trees, shrubs, bushes, etc.), 12.8% on construction (buildings and/or houses), and only 0.2% underground [27]. This demonstrates that V. velutina, like Vespa simillina, can nest in a diverse range of places with considerable adaptability, and both species are able to change nest sites in response to their environment [31]. Nests at high altitudes in trees can be dangerous to forestry management (wood cutting) personnel working in cities or forests. In 2018, many accidents were incurred by forestry workers due to the presence of V. velutina undetected in tree canopies, which required the mobilization of aerial transport to transfer the injured to the nearest health center.
The nest envelope of V. velutina has been characterized as a shell-like structure with short air pockets created using a maximum of five layers [10]. A nest may be defined as any modification of the environment by adult insects that provide shelter for the rearing of their offspring [32]. V. velutina make their nest out of paper-like fiber. Wood pulp collected from different sources is chewed and mixed with saliva by the hornets. The size of a hornet nest grows in proportion to the size of the inhabitants and the development rate of the colony. In our study, the insect adult populations (males, workers, and gynes) were as follows: nest A, n = 87 (11%, 66%, and 23%); nest B, n = 1021, (3%, 62%, and 35%); and nest C, n = 430, (20%, 73%, and 7%). Previous work showed that the number of combs, cells, and individuals produced from nests (n = 3) in December in France varied between 7 and 10 (mean of 8), 3685 and 5607 (mean of 4756.7), and 4039 and 4583 (mean 4382.4), respectively [26]. A nest removed in Germany in November consisted of a fully developed, medium-sized nest of six honeycombs with approximately 2637 used brood cells.
Availability of food and nesting resources is important to insects, as well as weather and climatic variables, as the thermoregulation capability of the nest is determined by ambient temperature. Nest architecture is an extended phenotype of a colony and thus under natural selection. Within a species, the rule may be context dependent, varying adaptively according to colony age, size, nutritional status, or reproductive state [33]. Previous research suggests that seafood might provide a valuable alternative food source favoring V. velutina colony development [34]. In the area from which nests were collected in the present study, V. velutina has a long colony cycle; the queens are normally spotted awakening from hibernation as early as late February. The present work shows colonies are active in January.
Different organizations, mainly public administration and academia, share the opinion that "the destruction of nests in winter is not seen to be an important method of control since the queen is dead and the future founders have already left the nest to winter" [35,36]. However, the results obtained in the present work show us that three nests captured in winter were active with a variable number of individuals, among which were a non-negligible number of founding queens. Since only one or a small number of queens is sufficient for the development of the population, it becomes necessary to change previous approaches to avoid future spread. Any detected nest must be immediately deactivated and removed. This may be especially important in areas where V. velutina is expanding into new territories.
An in-depth understanding of the behavioral ecology of invasive species in unique surroundings is needed to determine the behavior that maximizes health and thereby enlarges the alien species population [37,38]. A V. velutina colony life history can be described by understanding age and size at maturity, pattern of reproductive periods, size and number of offspring produced, and demise of the colony [39]. Better knowledge of invasive species is important to address the most suitable actions to prevent their spread. Several important questions we could not address include whether the queens found in the nests belonged to the nests originally, if they were fertilized or not, their age, or if they were using the nest as protected space while waiting for better conditions to start new nests and colonies or as a stop prior to hibernation.

Conclusions
To the best of our knowledge, the work presented here represents the first reported study of nest architecture and colony composition over winter of the yellow-legged Asian hornet. Since all overwinter nests contained living individuals of all castes, as well as brood and eggs, it is extremely important to remove V. velutina nests during the winter as well as the rest of the year. It is essential to work together to try to reduce or stop the spread of this invasive species by way of educating the public about V. velutina and supporting further research with a greater number of nest studies in different areas and at different times of their development.
Funding: This research received no external funding.