Nectar for two partners. Ants and bees around Ferocactus

Nectar within and outside the flower

At the apex of Ferocactus wislizeni, a bee may descend among the stamens while worker ants walk over neighbouring areoles. The plant supplies nectar to insects in two locations. Inside the flower, the secretion rewards visiting animals. Extrafloral nectaries formed from modified spines also attract ants when the flower is closed. The distance between these locations is small, so the paths of the two groups regularly cross. [McIntosh 2005] [Ness 2006]

The nectaries of F. wislizeni are concentrated in the upper areoles, near developing buds, flowers and fruits. Their secretion provides ants with a predictable food source. The presence of patrolling workers can limit herbivores attacking the reproductive organs. Protection and pollination nevertheless take place within the same small space, and one partner's behaviour changes the other's operating conditions. [Ness 2006]

Which bees pollinate the flowers

McIntosh studied pollination in two Arizona populations. F. wislizeni was observed at the Santa Rita Experimental Range and F. acanthodes at Desert Peak. The second species was described in the publication as F. cylindraceus, a name now assigned in POWO to the nominate subspecies of F. acanthodes. Research was conducted in the 1996–1998 seasons, with flower observations between approximately 9 and 13 hours. [McIntosh 2005] [2-WWW]

A flower was covered with mesh, then allowed a single visit and covered again. Later, fruit formation was checked and seeds were counted. Pollinator effectiveness had two components. The first was its number of visits. The second included the probability of fruit formation and the number of seeds after a single visit. Linking these measures allowed the contribution of a particular insect group to reproduction to be established. [McIntosh 2005]

The three most important species were solitary bees associated with cactus pollen, Diadasia rinconis, Svastra duplocincta and Ashmeadiella opuntiae. D. rinconis dominated at Desert Peak. In F. wislizeni, its single visits produced fruits more often than visits by the other two species, and seed numbers were greater than after visits by S. duplocincta. Rare visits by other insects could also lead to fruit formation. For some small bees, this outcome did not occur in the single-visit trial. [McIntosh 2005]

Flower structure guides a bee among the stamens

The broad perianth of both ferocacti surrounds numerous stamens and a stigma projecting above them. A bee collecting pollen moves over the anthers. To obtain nectar it must descend deeper, among the stamens, to the base of the floral tube. These two behaviours change the insect's route, but both can ensure contact with the organs that transfer and receive pollen. [McIntosh 2005]

The study found no significant difference between nectar and pollen visits in fruit-set frequency or seed number. Nectar visits were most frequent and, through their number, made a large contribution to the overall outcome. Effectiveness therefore depended on a combination of flower geometry, insect behaviour and return frequency. A single successful contact produced an average of 149 seeds in F. acanthodes and 148 in F. wislizeni. These means concern only flowers whose fruits matured. [McIntosh 2005]

Workers near buds and fruits

Another part of this network was studied at Tumamoc Hill in Tucson. The ants most often present around F. wislizeni were Crematogaster opuntiae, Forelius sp., Solenopsis aurea and S. xyloni. In the 2003 and 2004 flowering seasons, 259 plants were observed and inspected approximately once a week. A cactus was usually tended by one ant species at a given time, but the partner could change. Differences between them included worker numbers, aggression and their tendency to enter flowers. [Ness 2006]

In an accompanying protection study, Manduca sexta caterpillars were added as surrogate herbivores. All four ants attacked them and recruited further workers. Their effectiveness and patrol frequency allowed differences in fruit maturation to be predicted. The abstract indicates that the benefit of protection saturated at high intensity. Further increases in guard numbers therefore produced progressively smaller increases in the service. [Ness et al. 2006, abstract]

Aggression changes pollinator behaviour

The most aggressive species, S. xyloni, was also an ant frequently found inside flowers. In 2004, on plants tended by this species, workers were found in an average of 44,2% of available flowers and bees in 4,2%. With C. opuntiae, the corresponding values were 10,8% and 23,3%. These were means of percentages calculated for individual plants, with 84 and 62 plants in the respective groups. Each flower inspection lasted about 30 seconds. [Ness 2006, Table 1]

Separate ten-minute observations showed shorter bee foraging times in flowers occupied by ants. Mean time was 36 seconds in the presence of workers and 56 seconds in flowers without them. The shortest visits occurred with S. xyloni. Bees were sometimes driven away by workers, and several captured insects had ant heads clamped onto their legs. Contact could therefore pose a direct risk to the pollinator. [Ness 2006]

The danger estimate combined ant numbers in flowers with their previously measured effectiveness in attacking caterpillars. This was a comparative model. It did not directly measure the probability of a bee being killed. Insects might also respond to reduced amounts of nectar or pollen after ants had fed. The author therefore considered both attack risk and competition for food. [Ness 2006]

Petals as a boundary between partners

Ness tested workers' responses to signals originating from petals. Half of a dish was treated with a flower petal, while the other half served as a control. One worker was observed for five minutes. Only C. opuntiae significantly avoided the petal-treated half. The difference was not significant for other species. The tendency to avoid this surface corresponded to this ant's lower presence inside flowers in the field. [Ness 2006]

The result indicates the action of a petal-associated stimulus without identifying a particular chemical compound. This boundary affected individual ant species differently. Nectar outside the flower and a local repellent signal can together direct worker activity towards the area surrounding reproductive organs. The effectiveness of this separation depends on the partner's sensitivity to the stimulus. [Ness 2006]

Fruit retention and seed number are separate outcomes

Plants tended by S. xyloni had a lower total seed mass per fruit and lower individual seed mass than plants with the other partners. A pooled comparison of S. xyloni with the three other ants also found fewer seeds per fruit. When each of the four species was analysed separately, the overall test of seed number was not significant. The strength of the conclusion therefore depends on the precisely formulated comparison. [Ness 2006]

Ant species occupied plants naturally. They were not randomly assigned to cacti. The associations therefore link partner behaviour with reproductive outcome, with possible influences from growing location and other plant characteristics. Protection can increase the number of fruits that mature, while repelling bees can limit seed number in each fruit. A complete balance requires consideration of both stages and the subsequent fate of the offspring. [Ness 2006]

Nectaries drawing ants away from flowers

The 2025 study by Sandoval-Molina and colleagues extends this subject to Ferocactus latispinus subsp. spiralis (F. recurvus in the publication). The available abstract describes an ant-exclusion experiment. Ant presence in flowers was associated with lower seed production. A greater number of extrafloral nectaries was associated with a lower probability of workers entering flowers. The result supports the hypothesis of drawing ants towards a separate reward. [Sandoval-Molina et al. 2025] [2-WWW]

The abstract was checked for this study. It allows the direction of the observed relationships to be compared with earlier studies of F. wislizeni. In both cases, reproduction depends on nectar distribution and partner behaviour. Traits attracting protection and traits enabling pollination interact within one plant, and the result becomes apparent only when insect observations are combined with fruit and seed formation. [Sandoval-Molina et al. 2025] [Ness 2006]