GROUP SIZE, VIGILANCE AND FORAGING

Animals vary greatly in sociality: some live solitary lives while others live in social groups. Understanding the costs and benefits of group living has been a central problem in behavioral ecology for half a century. In this lab we used time budget analysis to explore the costs and benefits of group foraging sparrows in the UCSC arboretum (mainly golden-crowned sparrows but some white-crowns too). These birds are ideal for such a study because they can be alone as well as in groups of varying size.

In theory, the costs and benefits that might change with group size include: (i) increased safety from predators in bigger groups due to the benefits of many eyes watching for predators which can translate into (ii) increased foraging time for individuals in large groups because they are able to spend less time scanning for predators. However, larger group sizes can also (iii) increase in the rate of aggressive interactions, which may or may not overwhelm the other factors. We collected time budget data on the frequency of these three different activities for individuals in different group sizes: feeding, scanning and aggression. We assumed that birds with their heads up chewing on seeds and scanning at the same time were scanning, not feeding.

Style of report:

This is a medium report (3-4 pages plus graphs and tables). Start with 1-2 introductory paragraphs that outline the conceptual basis of the study, what questions you sought to answer and a general (brief) overview of the approach of the study. Include brief Methods and Results sections and focus much of your work on the Discussion section where you present and discuss results. This study has many interesting dimensions to it, so take care structuring the logical flow of your arguments. The style and length are the same as your coot foraging paper, so keep in mind feedback given on those papers and double check your final product with the hints given in the ‘how to write a paper’ handout we gave you at the start of the class!

Use WEB of Science to find references:

The Web of Science (a database search engine) is a great way to track down useful references. If you go to the library home page and find Web of Science or use the following URL (http://isi2.isiknowledge.com/portal.cgi/wos), you can use the Web of Science to find papers that cite a particular paper of interest. In this case you can find papers that cite the Caraco paper on group size and time budgets. Once you are in the Web of Science web page, choose "Full search", then in the window that appears (i) tick the box "Science Citation" and then click below on the "Cited Ref Search". Caraco's paper was extremely influential -- you will find a list of 250 papers that cite it. Most of these will not be useful for your report; try using the titles to guess which ones are most likely to be useful.

What to analyze:

Alexis compiled your data in a Statview file (Sparrows Compiled.svd) and placed in the Sparrow folder on each of the computers. The count data for each of the three activities have already been converted to proportions (ppn). This file also includes the data collected in 2001 by the previous class, so you have two year's worth of data to analyze. Because most of the data are proportions we will use NON-PARAMETRIC tests in this exercise. The two tests you need are Spearman correlation (to look for a correlation between two continuous variables, be sure to pay attention to the sign of the correlation (+ or -)) and a Mann-Whitney U test (to compare two distinct groups, as in any comparisons between the two years). Both of these tests are in the "non-parametric" option in the "Analyze" option in the Statview menu bar. Note: Mann-Whitney will give you a statistic (report the Z value) and a P value, but not means or standard deviations for the two groups being compared. To obtain mean values for the proportions, repeat the exact same analysis you do with the Mann Whitney (i.e. same group and measured variables) but using an unpaired t test instead of the Mann Whitney or display “descriptive statistics” splitting the variable by the year. To use descriptive statistics, go to the menu bar, select ‘descriptive statistics’ (you can refine the display in the ‘more choices’ menu), and add the data you wish to view. To split a category, make sure the analysis is still selected (black squares at the corner of the display) and select the ‘split by’ button at the top of the new view window – you will then be asked to select the nominal variable that divides the data. Make sure that your split by category is the same as the criteria for your analysis, or your means will not be correct!

First, analyze the combined data (all species, all years) by choosing "no criteria" in the criteria box at the top of the datasheet. Look at the relation between group size and the three activities (with three Spearman tests). Also, pairwise analysis of the relation between time spent in the different activities (e.g. Spearman correlations between proportion of time spent (i) scanning and feeding, (ii) scanning and aggression and (iii) feeding and aggression) will indicate which activities have negative trade-offs (trade-offs indicated by significant negative correlations).

Second, repeat this for each year separately (done by choosing the appropriate year in the criteria box, this will only include the data from the selected year; Alexis has already set up the criteria).

Finally, if the patterns for each within year analyses differ, it would then be worth looking for differences between the years in some of the key variables with Mann Whitney tests (enter date as the "two group variable" and then the variable of interest in the "measured variable"). Depending on what your analyses of each separate year reveal, the sorts of things you could compare between years might include: group size and the proportions of time spent in the different activities.

IMPORTANT DATA: The observations in 2001 were done on a very cold and rainy day; the observations in 2003 were done on a sunny and warm day. This could have obvious implications for food demands and the critical assumption that time and food are limiting.

Other Important Hints:

We have data for two years and you can really think about what the different patterns mean. You should address the following questions:

(1) Is there evidence to support Caraco's predicted relation between group size, scanning and feeding for all years, or for either of the years separately? If you do not find evidence for the predicted pattern, provide a possible explanation.

(2) If there are differences between the years, why might this be? Do group sizes differ between the years? If so, why? If so, how might this explain the differences between years in behavior?

IN OTHER WORDS really think about the biology of what is going on here.