By Thomas C. Grubb, Jr. , Department of Zoology, Ohio State University
The Boxwood Press; Pacific Grove; 1986
(Used with Permission)



Chapter Seven  When Do Great Blue Herons Give Up?
The hunting strategy of Great Blue Herons might best be termed the "mobile ambush." In almost any habitat containing sufficient frogs, fishes, or other prey, this largest of our herons may be seen poised along the water's edge or wading out to its featherbottoms, searching the water. Periodically, the neck is cocked into the familiar S shape, then the rapier bill is launched to grasp the victim. Careful attention will show you that a heron does not stand for very long in one place. After watching intently, but in vain, for awhile, the bird will take one or several steps before becoming stationary again. How does it decide how long to wait in one place before moving? Put another way, what determines its givingup time?
Suppose herons have to learn when to give up hunting in any one particular place. One very useful piece of information could be the average time interval between arriving at a hunting site and catching a fish or some other prey there. If herons can make such a calculation, or something like it, we are led to the hypothesis that givingup times are always slightly longer than the average interval between coming to a stop at a site and catching a prey item there.
Great Blue Herons hunt over a wide range of habitats and under a variety of climatic conditions. We find them in swamps and marshes, along the shores of lakes, and wading about in estuaries and shallow bays of the ocean. They are hardy birds and can be seen fishing in snowstorms as well as during high summer. This great spread of times and places gives us a way to test our hypothesis. We can predict that time to a catch will vary greatly over the species' range, but if we measure time to a catch in any particular location, we should always find it slightly less than givingup time. Stated another way, catch time and givingup time should be positively correlated; the longer the time to a catch, the longer the givingup time.
This project requires taking records over the course of the annual cycle and in as many habitats as possible. You will need a stopwatch, notebook and binoculars.
Two kinds of records are needed, the number of seconds that a heron remains at each stop when it fails to catch a prey, and the number of seconds between coming to a stop and catching a food item. For our study, we want to treat one habitat on one day as a single record. Do this by filling out a table in your notebook similar to Table 7.1. You will need to make a new table for each day and each habitat. Include in each table the location, date, and climatic conditions such as temperature, an estimate of wind velocity, and percentage of the sky obscured by cloud cover. After each field session, calculate the average time to a catch and the average givingup time. Before you can proceed further, you will need to build up, say, 20 to 30 of these pairs of averages.

(in seconds) 
(in seconds) 








































Time = 

To evaluate our hypothesis, we want to see whether there is any relation between catch time and givingup time. If, over a spread of habitats and seasons, the two numbers get larger or smaller together, this would support our hypothesis that givingup time is determined by catch time. As you accumulate pairs of average numbers add them to Fig. 7.1. Each point in 7.1 will correspond to one habitat on one day. It might be easier to make your own figure using graph paper, and you may have to if you find average catch times or givingup times longer than those in Fig. 7.1.
Fig. 7.1 is called a scattergram. You can inspect yours as a method for assessing the hypothesis. First, though, we will make up some results to see what they can tell us. In Fig. 7.2A, I have added to the scattergram 25 points taken from various imaginary habitats on 25 days. As we look over these results, we see there is a close correspondence between the two values. It is clear that as average catch time increases, so does average givingup time. These presumed results would certainly support our hypothesis that catch time determines givingup time. Another imaginary set of 25 records is plotted in Fig. 7.2B. In this case, we look in vain for any relationship. Givingup time appears to be independent of catch time. For example, in Fig. 7.2B, catch times from 30 to 110 seconds can all correspond to a givingup time of 60 seconds. From results such as these, we would reject our hypothesis; how long a heron hunts in one place is not determined by its average catch time in that habitat on that day. It is very likely that when you graph your own results, they will have a look intermediate between the appearances of Figs. 7.2A and 7.2B. Perhaps your curiosity will be sufficiently piqued that you will work through the statistical analysis described below. By doing so, you can make a firm decision about the validity of our hypothesis.
The methods of this project are directly applicable to a large number of species that hunt using the mobile ambush strategy. The hypothetical relationship between catch time and givingup time can be tested in other herons, flycatchers, kingfishers, hawks, bluebirds and shrikes. Maybe you can think of other possibilities.
Let's consider some numbers which illustrate the troublesome sort of intermediate results produced by many field studies. Examine the 25 points in Fig. 7.3. It looks like there might be relationship between catch time and givingup time, but it is not very clear cut, and the apparent relationship could simply be due to chance variation in the heron's behavior. We really cannot decide whether catch time determines givingup time without performing a statistical test. Wehen we want to know if two quantities vary together in some predictable way, we do what is called a correlation analysis.
Appendix 2 presents the Spearman Rank Correlation Analysis, and uses hunting behavior of Great Blue Herons as an illustrative example. Try this corelation analysis, using Table 7.2 to arrange your numbers.