P. Wells,* H. Wells,** V. Vu, N. Vadehra, C. Lee, R. Han, K. Han, and L. Chang

Department of Biology, Occidental College, Los Angeles, California 90041
**Department of Biological Sciences, The University of Tulsa, Tulsa, Oklahoma 74104

Abstract. - Conventional wisdom accords to the honey bee Nasanov gland pheromone a forager attractant function. Our experiments failed to support three predictions of that hypothesis. 1) Fragrant components of Nasanov gland secretion were not more effective than other scents as recruitment incentives for new (naive) foragers to a food source. 2) Honey bees harvesting scented rewards did not choose "Nasanov mixture" scent in preference to other odors. 3) After training to "mixed scent," containing Nasanov and nonNasanov components, bees did not prefer Nasanov pheromonal components. Our experiments did not support the conclusion that Nasanov gland secretions function as a "forager attractant" pheromone.

Table 1. Recruitment of naive bees by foragers harvesting 1.25 M sucrose rewards that were unscented, or perfumed with 100 ul/l cinnamon oil, citral, cajeput oil, geraniol or nerol.
Sucrose reward	Expt. days	Forager visits	Naive recruits	Recruits/ 100 visits
Unscented*	2	258	3	1.2
Cinnamon	3	822	110	13.4
Citral**	3	888	99	11.1
Cajeput	2***	611	70	11.5
Geraniol**	3	648	64	9.9
Nerol**	3	914	100	10.9
* Presented on the first and last days of the experiment. ** Nasanov scents. *** One day rained out.

Several attempts to verify the "forager attractant" hypothesis have used dish preference tests. For example, more bees "hovered or landed" at geraniol-scented dishes than at control scents or unscented control dishes; the sum of activity at control dishes, however, exceeded that at the expenmentals (Free 1962). A very similar result was obtained when excised Nasanov glands, rather than geraniol, provided the experimental scent (Free 1968).

In other experimental designs, more honey bees were captured in insect traps baited with a mixture of Nasanov compounds, a "Nasanov lure," than in unbaited (odorless) control traps (Free et al. 1984). Similarly, Mayer et al. (1989) reported that a commercially available pheromone mixture, "Bee-Scent" (Scentry Corp.), when sprayed on blooming fruit trees, increased numbers of foraging bees over those observed in untreated orchard plots. Neither of these studies included alternative (nonbee) scents as additional control treatments.

Despite von Frisch's ambiguous results, alternative interpretations of dish preference study results and a paucity of controls in trap and field tests, the body of evidence cited above has been viewed as generally supportive of the Nasanov "forager attractant" hypothesis. Many contemporary accounts of honey natural history accept that function for Nasanov pheromone (e.g., Seeley 1985; Free 1987; Winston 1987; Gould and Gould 1988).

On the other hand, citral and geraniol failed to attract bees in an olfactometer (Woodrow et al. 1965). Neither did these fragrances regularly increase bee populations when sprayed on test plots of alfalfa, unless coupled with sucrose rewards (Waller 1970). Indeed, in Waller's (1973, Table 1C) "dish preference" experiments, results seem to refute the "forager attractant" hypothesis.

Also, foragers seldom expose their Nasanov glands when visiting natural flowers (Free 1968), but they often do so at experimental feeders containing unscented sucrose solutions (Wenner et al. 1969; Wells and Wenner 1971). Paradoxically, when scent levels in food rewards were increased in those experiments, Nasanov exposure decreased and recruitment increased.

Thus, some published reports support the Nasanov "forager attractant" pheromone hypothesis, while others do not. Additional experimental studies may help to resolve the issue. If Nasanov secretion were in fact a forager attractant pheromone one might expect that: 1) one or more of its fragrant constituents would be more effective than other scents in recruitment of foragers to a food source; 2) honey bees harvesting scented rewards should choose a "Nasanov mixture" scent in preference to alternative odors provided after removal of the training scent; 3) honey bees harvesting a "mixed scent" reward, containing both Nasanov and control components, should choose Nasanov compounds when the training mixture is replaced by a set of "single scent" rewards that each contain only one of the training mixture components.

We have experimentally examined these three predicted properties of a presumptive Nasanov gland "forager attractant" pheromone by testing the null hypotheses of equality among scents under the challenge of these predictions.

Materials and Methods

All experiments were done on the Occidental College Campus, Los Angeles, California. Individually-marked foragers from a colony of approximately sixty thousand honey bees were fed 1.25 M sucrose solution from syracuse watch glasses at a distance of 50 m from the hive. The sucrose rewards were unscented or were scented with Eastman Organic Chemicals No. T 378 geraniol, Sigma Chemical Company No. N-7761 nerol, No. C-1645 citral, No. A-6769 anise oil, No. B-4258 bay oil, No. C-7517 cajeput oil, No. C-7267 cinnamon oil, No. C-8392 clove oil, or mixtures of the above at rates specified below. In control experiments excised Nasanov glands or Scentry Inc. "Bee-Scent" commercial pheromonal mixture was used.

Experiment 1. - Eleven individually paint-marked foragers, initially trained on clove-scented sucrose solution, were given two-hour feedings of unscented 1.25 M sucrose for three days prior to the experiment. Then, and throughout the experiment, all unmarked bees (newly recruited "naive" bees) that landed on the feeding dish were captured and killed. Hence, only the marked foragers made repeated trips from hive to feeder.

Food was provided only from 10:00 AM-12:00 M during each day of the experiment. All scented solutions were at the rate of 100 ul fragrant oil/l. Unscented sucrose reward was provided on control days 1 and 22 of a 22-day experiment. On each of days 2-21, a sucrose solution scented with cinnamon, citral, cajeput, geraniol, or nerol was provided, sequentially in that order, one scent per day, with four repetitions of the sequence. Days on which a given scent was provided were separated by four feedings of other odors. During one sequence (days 14-19), however, recruitment could not be monitored because of intermittent rain. Thus, we recorded recruitment on two days each for unscented and cajeput-scented rewards and on three days each for rewards containing cinnamon, citral, geraniol, or nerol scent.

Experiment 2. - Ten individually-marked foragers were initially trained to 1.25 M clove-scented sucrose and were fed daily from 10:00 AM to 12:00 M during each day of the experiment. On day 1 of the experiment, instead of the training scent, separate dishes of cajeput-, anise-, bay- and "Nasanov mixture"-scented sucrose were provided. The specific nonNasanov scents were at the rate of 100 ul/l; the "Nasanov mixture" included 100 ul geraniol, 100 ul citral, and 50 ul nerol/l.

The scent at which each marked forager first landed and drank, and its choices at subsequent visits were recorded. On day two the most popular scent was omitted; only the three less favored scents were provided. On day three the new most popular scent was omitted, leaving two choices. On day four, the remaining (least popular) scent was pitted against clove, the original training scent. On each day, first and subsequent visits of the marked foragers were recorded. This experiment was repeated thrice with new sets of marked bees each time, and with minor variations (once without "day three," once without "day four" and once with "day one" only).

Experiment 3. - Ten individually-marked foragers were trained to a mixture of scents that included 100 ul/l each of anise oil, bay oil, citral, geraniol and nerol.

On day one of the experiment, instead of the mixture, individual scents (100 ul/l, as in the mixture) were provided in separate dishes. The scents at which the experimental bees first landed and drank, and their choices on subsequent visits, were recorded. These bees were again offered the array of scented-sucrose rewards in separate dishes on the second day of the experiment, and their choices were recorded.

Control experiments. - Three additional experiments were run as controls. In the first, ten bees were trained on 100 ul/l clove-scented sucrose. Then, on the day of the experiment, they were offered separate dishes containing either unscented, 100 ul/l anise-scented, 100 ul/l cajeput-scented or, as a fourth choice, unscented 1.25 M sucrose to which eight fresh surgically excised Nasanov glands had been added. Data were recorded as in experiment 2.

In a second control, bees were trained on 100 ul/l anise-scented sucrose and then tested on separate dishes of 100 ul/l bay-scented, 100 ul/l cinnamon-scented rewards or, as a third option, 1.25 M sucrose solution to which 100 ul/l of Scentry "Bee-Scent" had been added. The scent(s) chosen by each bee were recorded.

As a variation of this experiment, these same bees were then fed bay-scented sucrose for two days. Then, on the third (test) day, they were offered a choice of 100 ul/l clove-, 100 ul/l cinnamon-, or 200 ul/l Scentry "Bee-Scent"-perfumed 1.25 M sucrose. Foraging visits by each bee to the test scents were recorded.

A third control experiment measured Nasanov gland exposures by foragers harvesting 1.25 M sucrose scented with 100 ul/l of "Bee-Scent," 100 ul/l of bay oil, or the unscented sucrose reward. Twelve bees were allowed to harvest one of the above solutions for an hour, then, during the next thirty minutes, numbers of visits and observable Nasanov gland exposures at the feeder were tallied. Two trials were run for each reward type; the same set of marked foragers was used throughout the experiment.

Results

Experiment 1. - Results recorded on separate days when the reward had a given scent did not differ systematically. Therefore, cumulative visits to the feeder by marked foragers, total numbers of naive recruits, and a ratio of recruits to visits are reported in Table 1. Marked foragers readily harvested unscented sucrose and each scented reward when it was offered. On days when unscented reward was provided, the first visits of some foragers were delayed, resulting in fewer cumulative visits. Once started, however, those foragers made repeated trips to the feeder.

Each of the scents we used was effective in the recruitment of naive foragers.

Table 2. Reward choices by honey bee foragers. After training to clove-scented sucrose, bees were allowed to choose bay-, anise-, cajeput- or "Nasanov mixture"-scented rewards. On day 2 only the latter three scents, and on day 3 the latter two scents were provided. On day 4, "Nasanov mixture" was pitted against clove-scented reward. N = number of trained bees.
Day	N	Bay	Anise	Cajeput	Clove	"Nasanov mixture"	Did not visit
1	40	29	4	0	-	0	7
2	30	-	20	1	-	0	9
3	20	-	-	13	-	0	7
4	20	-	-	-	11	0	9

Recruitment rates among citral-, geraniol-, nerol-, cajeput- and cinnamon-scented rewards did not differ significantly (X2 = 3.894, df = 4). Neither did recruitment to the three Nasanov scents, as a group, differ significantly from recruitment to the grouped control scents (X2 = 2.363, df = 1). Recruitment did differ significantly between scented sucrose rewards and the unscented control (X2 = 23.176, df = 1, P < .01).

Experiment 2. - Data from four stepwise repetitions of this experiment are summarized in Table 2. Prior to the first test, the individually-marked foragers harvested clove-scented sucrose from a feeder dish. On the first experimental day, when they were offered an array of separate dishes containing bay-scented, anise-scented, cajeput-scented and "Nasanov mixture"-scented sucrose rewards, respectively (but none with the clove-scented training solution), forager choices among the test solutions were not equal (X2 = 70.879, df = 3, P < .01). Most of the foragers harvested the bay-scented rewards.

When the experiment was repeated on day 2, without bay, significant inequality of visitation again occurred (X2 = 36.286, df = 2, P < .01). Most of the foragers harvested anise-scented sucrose. Similarly, on day three, numbers of foragers visiting the two remaining test dishes were not equal (X2 = 13.000, df = 1, P < .01). With neither bay nor anise present, foragers chose cajeput. On day 4, clove, the control scent, was preferred to "Nasanov mixture" (X2 = 11.00, df = 1, P < .01).

The null hypothesis of equal visitation to all dishes was not supported in any of the experiment 2 tests, and "Nasanov mixture" was never the scent of choice. Under these experimental conditions, bees were constant foragers on their chosen scents, with a cumulative error rate (imbibe from any other dish) of only one percent.

Experiment 3. - During the training period, marked foragers readily harvested sucrose reward perfumed with a blend of anise, bay, citral, geraniol and nerol scents. On test days 1 and 2 when sets of individual dishes, each containing only one of those scents, were provided, they drank from the dishes on which they landed.

Visitation patterns of seven experimental bees are summarized in Table 3, with an asterisk indicating the dish first visited by each bee. Bay and geraniol were favored scents, both for first visits and largest numbers of harvesting trips. Bees also drank from anise and nerol, but never from citral. Neither the "forager attractant" nor the null hypothesis predicted this result (X2 = 19.891, df = 4, P < .01 of equal visitation, bees 1-6).

Table 3. Individual scents visited by bees first trained to a mixture of all scents. A = anise, B = bay, C = Citral, G = geraniol, N = nerol. * = first visit.
Bee No.	Day 1					Day 2					Total
	A	B	C	G	N	A	B	C	G	N	A	B	C	G	N
1	-	-	-	-	-	0	0	0	19*	3	0	0	0	19	3
2	0	15*	0	0	0	0	12	0	16	0	0	27	0	16	0
3	0	13*	0	0	0	0	23	0	0	0	0	36	0	0	0
4	0	6*	0	6	1	0	0	0	8	1	0	6	0	14	2
5	0	4	0	6*	1	0	3	0	17	2	0	7	0	23	3
6	0	11*	0	2	2	4	4	0	10	4	4	15	0	12	6
7	-	-	-	-	-	0	11	0	1*	0	0	11	0	1	0

Only bee No. 3 was completely constant to one scent. The others drank from two or more differently scented dishes during the two-day experiment. The cumulative error rate (imbibe from a dish other than that most frequently visited) in experiment 3 was 29 percent, but it varied considerably among bees.

Control experiments. - When ten marked foragers that had been harvesting clove-scented sucrose were offered a choice among three test dishes, one containing freshly excised Nasanov glands and two with control scents, six bees chose the anise control and two chose cajeput. No bees landed at the dish of fresh Nasanov glands. Two of the marked foragers did not visit any dish on the test day.

When ten marked foragers, trained on anise-scented rewards, were offered a choice of Scentry "Bee-Scent"-, bay- or cinnamon-scented sucrose, six bees chose bay, two landed on cinnamon, and two did not visit. Eight of these same bees, after two days of harvesting bay-scented sucrose, chose clove, while the other two visited cinnamon. None of the bees regularly visited the dish perfumed with "Bee-Scent." Individual forager constancy to chosen scent was high, as in experiment 2, but when bees landed on the "Bee-Scent" dish or the "excised Nasanov gland" dish, they drank from it.

In the third control experiment (sums of two trials for each reward type), 106 visits to Scentry "Bee-Scent"-scented reward yielded ten Nasanov gland exposures (9%); 87 visits to bay-scented sucrose yielded nine gland exposures (10%); and 104 visits to unscented sucrose yielded 34 gland exposures (33%). When added to a sucrose reward, "Bee-Scent" suppressed Nasanov gland exposure as effectively as did bay oil.

Discussion

Naive bees, seeking a food source for the first time, rely heavily on odors to which they have been introduced in the hive by the experienced foragers that recruit them (Wells and Wenner 1971, 1973; Wenner 1974; Wenner and Wells 1990). Recruitment of new foragers to an unscented source is negligible (Wenner et al. 1969; Wells and Wenner 1971; Friesen 1973; our experiment 1), and foragers may be individually constant to specific scents in an array of food sources (plastic flowers) polymorphic for color and odor (Wells and Wells 1985).

Nasanov scent experiments which used unscented controls have confirmed the importance of odors to field bees (e.g., Free et al. 1984; Mayer et al. 1989; experiment 1). However, these experiments do not establish that Nasanov components are unique "forager attractant" pheromones. In our experiment 1, for instance, Nasanov scents did not differ significantly from control fragrances in attractant properties. Our results, and those of others (e.g., Mamood et al. 1992), do confirm conditioned responses by bees to odors associated with food rewards; Nasanov scents are effective conditional stimuli.

In our experiment 2, on redistribution of experienced foragers to an array of unfamiliar scents at a feeder location, "Nasanov mixture" was never the option of choice. Even in experiment 3, when foragers were allowed to choose among the individual components of an odor mixture to which they had been trained, all bees did not gravitate to Nasanov compounds. Some did prefer geraniol, but others chose bay, and forager constancy to individual scents was low.

The results of experiment 2 argue against the notion that components of the Nasanov secretion only acquire effective forager attractant properties when mixed together. Furthermore, neither excised Nasanov glands nor Scentry "Bee-Scent" gave results appreciably different from those we obtained with specific Nasanov components; suppression of Nasanov gland exposure shows that bees perceived Scentry "Bee-Scent" at our experimental concentrations.

Neither the existing body of evidence nor the results reported here justify the conclusion that Nasanov gland secretions function as a "forager attractant" pheromone.

Acknowledgments

We thank Mr. Bruce Steele for apicultural assistance; Scentry, Inc. kindly provided a sample of Scentry "Bee-Scent"; manuscript critiques by Drs. Frederick R. Prete, Gordon D. Waller and Adrian M. Wenner were very helpful.

Literature Cited

Boch, R., and D. A. Shearer. 1962. Identification of geraniol as the active component in the pheromone of the honey bee. Nature, 194:704-706.

Free, J. B. 1962. The attractiveness of geraniol to foraging honeybees. J. Apic. Res., 1:52-54.

________. 1968. The conditions under which foraging honeybees expose their Nasonov gland. J. Apic. Res., 7:139-145.

________. 1987. Pheromones of social bees. Cornell Univ. Press, Ithaca, N.Y.

________, A. W. Ferguson, and J. R. Simpkins. 1984. A synthetic pheromone lure useful for trapping stray honeybees. J. Apic. Res., 23:88-89.

Friesen, L. J. 1973. The search dynamics of recruited honeybees, Apis mellifera ligustica Spinola. Biol. Bull., 144:107-131.

Frisch, K. von. 1923. Uber die "Sprache" der Bienen, eine tierpsychologische Untersuchung. Zoologische Jahrbuecher. Abteilung fuer Allgemeine Zoologie und Physiologie der Tiere, 40:1-186.
1-186.

________. 1947. The dances of the honey bee. Bull. Anim. Behav., 5:1-32.

________. 1967. The dance language and orientation of bees. Harvard Univ. Press., Cambridge, Mass.

Gould, J. L., and C. G. Gould. 1988. The honey bee. Scientific American Library: Freeman and Co., New York.

Mamood, A. H., G. D. Waller, and G. M. Loper. 1992. Classical conditioned responses of honey bees to biological scents. Amer. Bee J., 132:403-405.

Mayer, D. F., R. L. Britt, and J. D. Lunden. 1989. Evaluation of BeeScent(r) as a honey bee attractant. Amer. Bee J., 130:41-42.

Morse, R. A., and R. Boch. 1971. Pheromone concert in swarming honey bees (Hymenoptera: Apidae). Annals Entom. Soc. Amer., 64:1414-1417.

Pickett, J. A., I. H. Williams, A. P. Martin, and M. C. Smith. 1980. Nasonov pheromone of the honey bee, Apis mellifera L. (Hymenoptera: Apidae). Part I. Chemical characterization. J. Chem. Ecol., 6:425-434.

_____, _____, M. C. Smith, and A. P. Martin. 1981. Nasonov pheromone of the honey bee, Apis mellifera L. (Hymenoptera: Apidae). Part IIId. Regulation of pheromone composition and production. J. Chem. Ecol., 7:543-554.

Renner, M. 1960. Das Duftorgan der Honigbiene und die physiologische Bedeutung ihres Lockstoffes. Z. vergleich. Physiol., 43:411-468.

Ribbands, C. R., and N. Speirs. 1953. The adaptability of the homecoming honey bee. Brit. J. Anim. Behav., 1:59-66.

Schmidt, J. 0., and S. C. Thoenes. 1987. Honey bee swarm capture with pheromone-containing trap boxes. Amer. Bee J., 127:435-438.

_____, _____, and R. Hurley. 1989. Swarm traps. Amer. Bee J., 129:468-471.

Seeley, T. D. 1985. Honeybee ecology: a study of adaptation in social life. Princeton Univ. Press, Princeton, N.J.

Shearer, D. A., and R. Boch. 1966. Citral in the Nassanoff pheromone of the honey bee. J. Insect Physiol., 12:1513-1521.

Sladen, F. W. L. 1901. A scent-producing organ in the abdomen of the bee. Glean. Bee Cult., 29:639.

Snodgrass, R. E. 1956. Anatomy of the honey bee. Comstock Publishing Associates (Cornell University Press), Ithaca.

Waller, G. D. 1970. Attracting honeybees to alfalfa with citral, geraniol and anise. J. Apic. Res., 9:9-12.

________. 1973. The effect of citral and geraniol conditioning on the searching activity of honeybee recruits. J. Apic. Res., 12:53-57.

Wells, P. H., and H. Wells. 1985. Ethological isolation of plants 2. Odour selection by honeybees. J. Apic. Res., 24:86-92.

________, and A. M. Wenner. 1971. The influence of food scent on behavior of foraging honeybees. Physiol. Zool., 44:191:209.

________, and ________. 1973. Do bees have a language? Nature, 241:171-174.

Wenner, A. M. 1974. Information transfer in honeybees: a population approach. Pp. 133-169 in Nonverbal communication: Vol. 1. Advances in the study of communication and effect. (L. Krames, T. Alloway, and P. Pliner, eds.), Plenum Press, New York.

________, and P. H. Wells. 1990. Anatomy of a controversy: the question of a "language" among bees. Columbia Univ. Press, New York.

_____, _____, and D. L. Johnson. 1969. Honeybee recruitment to food sources: olfaction or language? Science, 164:84-86.

Winston, M. L. 1987. The biology of the honey bee. Harvard Univ. Press, Cambridge, Mass.

Witherell, P. C. 1985. A review of the scientific literature relating to honey bee bait hives and swarm attractants. Amer. Bee. J., 125:823-829.

Woodrow, A. W., N. Green, H. Tucker, M. H. Schonhorst, and K. C. Hamilton. 1965. Evaluation of chemicals as honey bee attractants and as repellents. J. Econom. Entom., 58:1094-1102.

Accepted for publication 1 October 1992.