Controlling Water Collection.

Water collector. Photograph by Linton Chilcott.

Tom Seeley’s Guest Blog.

Thank you, Tom, for another insightful peep into the world of honey bee behaviour and how science solves some more mysteries.

Water Needs.

Colonies of honey bees need water to survive.  A colony needs water to prevent dehydration of the adult bees, to prepare liquid food for the brood, and to cool the hive on hot days.  A colony’s water need is often met by the water that its foragers retrieve incidentally as they collect nectar, since nectar contains much water.  Sometimes, however, a portion of a colony’s work force must intentionally collect water from streams and other wet places.  These are the times when either the colony’s nectar collection is very low, due to a dearth of nectar-bearing flowers, or its water consumption is very high, due to high temperatures that create a strong need for evaporative cooling inside the hive.  It is not surprising, therefore, that a honey bee colony possesses mechanisms for controlling the rate at which its bees collect water.

Water is collected by elderly bees that work outdoors.  They fly out to whatever puddle or other water supply is near the hive, fill their honey stomachs with water, and return home and offload their water to middle-age bees that work inside the hive. 

A bee collecting water from wet moss. Photograph by Linton Chilcott.

Thus, there exists a division of labor between the bees that collect water and the bees that consume water.  This means that a colony must solve the problem of keeping a collection process and a consumption process in balance.  A prolonged imbalance between water collection and water consumption can be disastrous.  If consumption exceeds collection on a very hot day, a colony can overheat, causing abnormal development of the brood and sagging of the combs. 

Regulated water collection by a honey bee colony.

Fig. 2 shows how a colony can adaptively adjust its rate of water collection.  In the early 1950s, Martin Lindauer, a German bee researcher, moved a colony living in an observation hive into a greenhouse so he could closely monitor the colony’s rate of water collection from a water source (a soggy clump of moss).  When Lindauer turned on a lamp beside one of the hive’s glass walls, threatening the colony with lethal overheating, the colony boosted its water intake, to begin strong evaporative cooling of its home.  This stabilized the hive’s interior temperature.  When the heat stress was removed, the colony promptly lowered its water intake.

Solving Mysteries.

This shows that a colony can turn up and turn down its water intake, as needed.  To understand how it does these things, we must address two puzzles:  1) How do a colony’s workers know when to begin collecting water? and 2) How do a colony’s workers that have begun collecting water know whether to continue or to stop collecting water? 

What stimulates workers to begin collecting water?  Some forager-age bees are stimulated to start fetching water by heeding the waggle dances of hive mates that have already begun to collect water.  But what stimulates the first water collectors and hence starts the whole water-collection process? 

In the case of water collection for cooling purposes, one might suppose that it is sensing high temperatures inside the hive that initially tells workers that their colony needs water.  This is often, but not always, the case.  We know that sometimes a colony’s forager-age bees can become desperate to gather water when they are confined inside their hive by cold or rainy weather.  In my office, I keep a colony in an observation hive all year round.  The longest-lasting waggle dances that I have ever seen were performed on a winter day when the outdoor air temperature rose enough for water collectors in my “obs hive” colony to go outside and gather water from puddles of melted snow.  These bees must have responded to some indicator of water need other than high temperatures inside the hive.  What was it this indicator?

I suspect that what stimulated the first water collectors to action was the presence of highly concentrated sugar solution in their honey stomachs.  A bee might sense this either directly, as the fluid passes over her taste organs during food exchanges, or (more likely) indirectly, as a feeling of thirstiness. 


This mystery was investigated in the 1960s by Hans Kiechle, a PhD student of the renowned German bee biologist Martin Lindauer.  Kiechle worked with a colony living in an observation hive.  He assayed the thirstiness of this colony’s worker bees by placing a water-soaked cloth at the observation hive’s entrance and recording what fraction of the bees that contacted the wet cloth also drank from it.  He found that on rainy days, when the bees could not forage, most the bees at the wet cloth had an elevated sugar concentration in their honey stomachs and most drank from the cloth.  He also found that when he fed the colony a dilute (15%) sugar solution, this lowered both the sugar concentration of the crop contents of the colony’s worker bees and their desire to collect water.

Once a bee has begun collecting water, she must stay informed about her colony’s need for more water and respond accordingly.  If the need persists, then she should continue collecting and perhaps even perform waggle dances to recruit others to the task.  But if the need subsides, then she should cease collecting.  So, what tells a bee engaged in water collection that her colony does or does not need more water?

 We now know that a water collector acquires information about her colony’s need for more water each time she returns to the hive, and we know that she does so by sensing how easily she unloads her water to her hive mates.  When there is an acute shortage of water, a water collector’s load will be taken quickly and eagerly, and she will continue fetching water.  But when the colony’s water needs have been met, a water collector’s load will be taken slowly and reluctantly, and she will cease bringing in water. 

This raises the question:  How does a water collector sense the ease of unloading?  Several years ago, I studied this question with Susanne Kühnholz, a student from the University of Würzburg in Germany.  We took a colony of bees—living in an observation hive—to a location where there were no other colonies.  This was the Cranberry Lake Biological Station, which is in heavily forested Adirondack Park in northern New York State.  First, we heated the observation hive and labeled bees that we found collecting water on the shore of the lake not far from our hive.  Then varied our colony’s need for water by heating its combs with a lamp, and we painstakingly monitored the behaviors of this colony’s (now labeled) water collectors throughout a day, for several days in a row.  We wanted to closely see how a water collector’s unloading experience changes when her colony’s water needs drops from high (heating lamp on) to low (heating lamp off). 

Changes in the multiple variables of a water collector’s unloading experience as her colony’s water need goes from high to low.

Fig. 3 shows the results for one typical water-collector bee.  As long as the hive was heated (lamp was turned on at 11:20) and the colony’s brood nest was threatened with overheating, the water-collector bee needed just a few seconds to find a hive bee who accepted her load, each time the water-collector bee returned to the hive.  Also, she contacted only a few bees who refused to take her water, and she needed at most 1 minute from when she entered the hive to when she finished her water delivery.  But once the danger of overheating was over (lamp was turned off at 2:38), the water-collector bee’s search times, unloading rejections, and delivery times all increased.  Eventually, she needed nearly four minutes to find a bee who accepted some of her water, she was rejected by hive bees more than 40 times, and she hadn’t fully disposed her load after more than 10 minutes.  At this point, she stopped collecting water.

You may be wondering if water collectors pay attention to the whole constellation of variables of the unloading experience, some subset of these variables, or perhaps just one special variable.  To address this mystery, we calculated how much each variable of the unloading experience increased when we turned the heating lamp off.  Averaging our results for eight bees, we found that the mean search time increased by a factor of 6, the mean delivery time by a factor of 4, and the number of unloading rejections by a factor of 10!  These numbers suggest that the number of unloading rejections is the most salient variable of unloading ease, and so may be the primary sign of whether to continue or stop collecting water.

Why does a water collector’s ease of unloading change when her colony’s need for water changes?  We learned that water receivers are the middle-aged bees, that is, the bees that are no longer nurse bees but not yet forager bees.  Thus, when a colony starts to overheat, it is some of the middle-age bees in a colony that increasingly seek loads of water from water collectors and then walk about the hive, distributing it to other bees or smearing it over the ceilings and sides of cells.  We also learned that the additional water receivers that appeared when we increased a colony’s water need were NOT bees that were functioning as nectar receivers earlier in the day.  This means that a colony can increase the number of water receivers without strongly decreasing the number of nectar receivers.  This seems highly adaptive, since a colony will be more successful in foraging if its nectar collection is not lowered every time its water collection must be boosted.

The story of the control of water collection in honey bee colonies illustrates how, through patiently performed studies, we can understand our bees both at the level of a whole colony functioning as a unit and at the level of individual bees cooperating closely to build smoothly running society.         

4 thoughts on “Controlling Water Collection.”

Leave a Reply

This site uses Akismet to reduce spam. Learn how your comment data is processed.