Florivory by lizards on Tacinga inamoena (K.Schum.) N.P.Taylor & Stuppy (Cactaceae) in the Brazilian Caatinga

Florivory can alter plant reproductive success by damaging sexual reproductive structures and disrupting plant-pollinator interactions through decreased flower attractiveness. Here, we report new records of the cactus Tacinga inamoena flower consumption by the Tropidurus hispidus lizard in the Brazilian Caatinga. We monitored 53 flowers from 11 T. inamoena individuals with camera traps over a 1-year period, totalling 450 camera-days of sample effort. We detected four florivory events. In three of these events, flowers were entirely consumed or had their reproductive structures severely damaged, leading to no fruit formation. Florivory events occurred in the morning, right after anthesis, in flowers near the ground, and lizards did not climb the cactus. Our results suggest that T. hispidus florivory on T. inamoena could have a negative impact on fruit set, since the consumed flowers were entirely destroyed. However, the long-term effects of florivory by lizards on T. inamoena reproductive success in the Caatinga still needs to be elucidated.


Introduction
Florivory refers to the consumption of floral structures by a wide variety of animal taxa (McCall & Irwin 2006). It can decrease plant reproductive success by damaging floral sexual structures or altering flower availability and attractiveness to pollinators (Cascante-Marín et al. 2009;Carper et al. 2016). Vertebrate florivory distinguishes from invertebrate florivory since entire flowers are frequently eaten in the former (Riba-Hernandez & Stoner 2005). Such is the case of florivory by lizards, as their feeding on flowers is often destructive (Olsson et al. 2000). Many lizard species feed on nectar and other flower parts because of their high nutritional and energetic values, potentially playing a significant positive role in pollen transfer (Pérez-Mellado & Casas 1997;Gomes et al. 2013;Hervías-Parejo et al. 2020).
Cactaceae Juss. is one of the most diverse plant families in the Neotropics, representing an important floristic element in arid and semiarid environments with extreme conditions (Mutke 2015). For this family, florivory has been found in drier areas, where cacti flowers stand out for their water content and nutritional value (Nobel 2002). In such environments, cacti flowers are partially or entirely consumed by invertebrates, e.g., Opuntia Mill. flowers eaten by Lepidoptera larvae in Chihuahuan Desert (Pinã et al. 2007(Pinã et al. , 2010 (Rocha et al. 2020) by flower-damaging bees Trigona spinipes (Fabricius 1973), and vertebrates, e.g., Opuntia flowers eaten by lizards in Galápagos Island (Nobel 2002), Melocactus flowers eaten by Tropiduridae lizards in the Brazilian Caatinga (Gomes et al. 2013), and the short columnar cactus Echinopsis rhodotricha K.Schum. florivory by deer and peccaries in the Brazilian Chaco (Gomes et al. 2016).
Tacinga is a genus of the subfamily Opuntioideae, composed of eight species endemic from Brazil, restricted to the semiarid lands in the Eastern corner of the country (Zappi & Taylor 2020). Tacinga inamoena (K.Schum.) N.P.Taylor & Stuppy, commonly known as "cumbeba" or "quipá", is a subshrub cactus, with round to oblong pads and solitary subapical flowers at the top of the branches, 4-6 × 3.5-4 cm, with a green globose pericarpel, 10 mm and orange-red, acute, fleshy bract-scales subtending bristles (Taylor & Zappi 2004;Peixoto et al. 2016) (Figure 1a). Their flowers, which are the showiest of the genus, are large, shiny, have erect stamens forming a column surrounding the stigma, provide pollen and nectar as floral rewards, and are pollinated by hummingbirds (Lambert 2009). The flowering period starts at the end of the rainy season, continues throughout the dry season, peaking in September/October (Quirino 2006). Flower anthesis is diurnal, usually lasting one day, from eight to 11 hours in prickly pears, and sometimes opening on a second day (Pimienta- Barrios & del Castillo 2002).
Tropidurus hispidus (Spix, 1825) is a heliophilous lizard that can reach up to 35 cm in total length, the largest of the genus (Kolodiuk et al. 2010). It is widely distributed in South America, occurring in open landscapes such as Caatinga, and living in forest borders, rocky outcrops, and human-altered areas (Fialho et al. 2000). Its generalist diet comprises plant material, ants, insect larvae, and vertebrates. Plant material (e.g., leaves, non-identified flowers, and fruits) have been reported as one of the three most important food items for T. hispidus populations in Caatinga and in rocky outcrops areas, especially during the dry seasons (Van Sluys et al. 2004, Kolodiuk et al. 2010Albuquerque et al. 2018), yet little is known about the identity of the plant species consumed (Koski et al. 2018).
During a study about Cactaceae frugivory by vertebrates in the Caatinga, a seasonally dry tropical forest in northeastern Brazil, we recorded florivory events by lizard Tropidurus hispidus on the cactus Tacinga inamoena. Herein, we quantified and described this flower herbivory interaction, which is new information for both species.

Material and Methods
Our study site was the Santo Antônio Ranch private property located in Lajes municipality (05°47'48" S, 36°14'23" W), Rio Grande do Norte state, northeastern Brazil (Brazilian Semiarid region). The region is characterized by low irregular rainfall (>300 mm/year), long periods of drought, and high average annual temperatures (30°C). We monitored 53 flowers from 11 T. inamoena individuals using camera traps (Bushnell Trophy CAM 8 MP HD) from September 2018 to July 2019. We had a total of 450 camera-days of sample effort distributed throughout the months (on average 45 cameradays per month). We placed the cameras one to two meters away from the flowering individuals and configured it to capture two photos followed by 10 second videos, in intervals of one minute. We then left the cameras in the field working 24 hours/day every month, recording the whole anthesis period of blooming flowers. We chose plants that were located at least 70 m from each other to ensure spatial heterogeneity.
can indirectly interfere in plant reproductive success by damaging the petals, making the flower unrecognizable or less attractive to pollinators Mothershead & Marquis 2000). In turn, this can affect the sexual display of the plant and disrupt plant-pollinator mutualistic interactions Leavitt & Robertson 2006). In general, petal removal can indirectly decrease reproductive output (seed set) via decreased flower visits by pollinators (Leavitt & Robertson 2006;Carper et al. 2016;Tsuji & Ohgushi 2018).
In one of the florivory events recorded on T. inamoena, the petals were not entirely consumed by the lizard T. hispidus, and, although through the camera traps we could not determine whether the reproductive structures were damaged or not, we did not observe any flower visitors after consumption by the lizard. In the prickly pear Opuntia microdasys (Lehm.) Pfeiff., florivory by Lepidoptera larvae caused a substantial decrease in their reproductive success by destroying the plant reproductive structures (Pinã et al. 2010). Also, florivory by Lepidoptera larvae on O. microdasys floral buds caused a high percentage of abortion and, consequently, no fruits (Pinã et al. 2007). In E. rhodotricha, a short columnar cactus species from the Brazilian Chaco, flowers that were partially damaged by deer and peccaries did not set fruits (Gomes et al. 2016). In contrast, flowers of the globular cactus Melocactus ernestii Vaupel also had their petals consumed by Tropiduridae lizards. However, such florivory did not impact the plant reproductive success, probably because the flower reproductive structures of Melocactus species are protected inside a modified stem (cephalium) and M. ernestii is capable of self-pollination (Gomes et al. 2013).
It is expected that bigger animals such as vertebrates would remove the entire flower when feeding on this resource (Riba-Hernandez & Stoner 2005; Gomes et al. 2016). In this case, florivory can have a direct effect on the plant reproductive output because completely removing the flower takes away its chances of reproducing developing fruit (McCall & Irwin 2006;Cascante-Marín et al. 2009). This pattern has been observed in E. rhodotricha florivory by deer and peccaries (Gomes et al. 2016) and in this present study. In our records, in three of the four florivory events, T. inamoena flowers were severely damaged.
Lizards and cactus flowers appear to have an old relationship, especially Iguania lizards and prickly pear cacti on islands (Christian & Tracy 1982;Christian et al. 1984;Schluter 1984). A study about the diet of the Tropiduridae lizard Microlophus pacificus (Steindachner 1876) on Isla Pinta Galápagos found Opuntia flowers in its stomach and a positive correlation between lizard body size and herbivory (Schluter 1984), followed by a recent study that has observed M. pacificus consuming Opuntia galapageia Hensl. flowers in the field, possibly transporting its pollen and acting as a pollinator (Hervías-Parejo et al. 2020). Opuntia florivory by a non-insular lizard species of Cnemidophorus was first We categorized the flower consumption records into four categories (from four to one) according to the Damage Intensity (D.I.) caused by the lizard on the flower. In the fourth category, the flower was entirely consumed (flower and pericarpel, which could be entirely consumed or detached from the plant); in the third category, the flower was entirely consumed and pericarpel was damaged; in the second category, the flower was entirely consumed without damaging the pericarpel, and in the first category, the flower was partially consumed, without damage to the pericarpel. Each record was classified into the four categories by analyzing the photos and videos. Day and time of consumption as well as the height of flowers consumed were also recorded.

Results
Tropidurus hispidus ate four T. inamoena flowers, which corresponded to 7.5% of the flowers monitored in the study period (N = 53). Three of the florivory events happened in the same cactus individual at the end of the 2018 dry season, and the other one during the 2019 rainy season ( Table  1). All the florivory events occurred in the morning, right after anthesis, during single visits that, in some cases, lasted several minutes (Table 1). In the last florivory event we recorded, the flower started anthesis on one day and opened again on a second day, when it was then fully eaten by the lizard. However, we did not record flower visitors on this flower during its first day of anthesis.
Lizards approached the plant and started gradually eating the flowers located near the ground, eventually leaning on wooden sticks on the ground to avoid climbing the cactus, and began by tearing the petals until reaching the globose pericarpel (Table 1, Figure 1). In one of the events, the lizard jumped to finish consuming the globose pericarpel remains (Video).
In three florivory events, the flowers were either entirely consumed (N = 2, D.I. = 4), including petals, stamens, and stigma along with the globose pericarpel, or severely damaged (N = 1, D.I. = 3) (Table 1). No fruits developed in these damaged flowers. In only one of the events, the petals were not completely consumed and the pericarpel was not damaged (D.I. = 1), remaining attached to the plant for more than 34 days until we collected the camera. In this case, we cannot affirm whether this flower developed into a fruit, considering that T. inamoena fruits take an average of six months to ripen after fertilization (Quirino 2006). As we observed in the footage, T. inamoena buds took about 25 days to develop and start anthesis, remaining open for a day or sometimes opening a second day. Throughout the monitored months, other available structures (e.g. flower buds and ripe and/or unripe fruits) were not consumed.

Discussion
Florivory can greatly impact plant species, similar to or sometimes more than tissue herbivory (McCall & Irwin 2006 (Mijares-Urrutia et al. 1997). In this last case, flowers were located near the ground and the lizard did not have to climb the cactus.
Opuntioideae cacti offer a great amount of resources including cladodes, fruits, and flowers (Mellink & Riojas-López 2002). A common reproductive feature of cacti is asynchronous flower development, which is, many floral buds, flower at anthesis, and unripe fruits available at the same time (Pimienta-Barrios & Nobel 1995;Pimienta-Barrios & del Castillo 2002;Quirino 2006). Despite that, two-thirds of reptiles that consume platyopuntias on the Galápagos Islands specifically eat their flowers, probably because of the pollen and nectar of cacti, which are often rich in sucrose (Pimienta-Barrios & del Castillo 2002;Fleming et al. 1994). Similarly, when feeding on T. inamoena flowers in anthesis, T. hispidus lizards ignored flower buds and ripe or unripe fruits.
Tacinga flowers have erect stamens and showy colours, such as the orange-red T. inamoena flowers, which have evolved to be pollinated by hummingbirds (Lambert 2009). Lizards use their vision to search for food near the ground and some diurnal species may also have visual pigments and photoreceptors, similar to those in birds, which could help them detect food (Ellingson et al. 1995;Janzen & Brodie 1995). Such features may have favoured T. inamoena flower detection by Tropiduridae lizards, as these flowers are located near the ground ( = 0.35 m in height, according to Table 1). Also, T. inamoena flower anthesis starts in the morning, overlapping with T. hispidus activity patterns, which is also when flowers are most conspicuous and provide copious amounts of nectar 22 ± 1.8 µl [%] = 26 ± 1.0 (N = 5 flowers), corresponding to 6.3 ± 0.5 mg sugar per flower (Galetto & Bernadello 2005).
According to Albuquerque et al. (2018) flower eating by T. hispidus was observed in populations of Caatinga and Cerrado, being higher during water scarcity periods in dryer areas. Tropidurus hispidus searches for floral food resources due to its need for water, as well as the high nutrient contents of flowers. The peak flowering of T. inamoena occurs in the dry season (Quirino 2006). Thereby, its flowers are an important resource for lizard energy requirements, especially in an environment as the Caatinga, where water and food resources are scarce in the dry months, which can last for a long part of the year. Despite this plant-animal interaction could bring negative impacts to the plant reproductive output, additional investigation is needed to evaluate the long-term effects of florivory by lizards on T. inamoena populations and whether it can influence its reproductive success by reducing the fruit and seed set.