This post is based on some recent categorization questions raised by Phil Lussier. Since many of you are likely contemplating these and related issues, I am providing my responses here. – Scott
Question 1
In seeing a group of crows on the compost pile I could just barely make out the very top of the head of one obscured by the pile. I wouldn’t have known it was there if not for consecutive images that showed it moving. However it never fully came into view. What I saw was glossy black but, if a crow, it was only the very top of its head. It was on the opposite side of the pile from the other crows. All of which led me to assume it was a crow as well. But I got to mulling it over and realized I could not make an i.d. from what little was visible. If I hadn’t thought about it I would have counted it as another crow but in trying think objectively I realized I shouldn’t label it so specifically without more evidence so I listed it as “unidentified”, “1”. It would be the logical thing to do if there had been no other crows in the image upon which to make the inference. Am I correct in my reasoning?
Response 1
Sounds like your encountering some of the complexities of this type of ecological research. In this particular case, where you are dealing with a group of crows and see an incomplete view with the top of a glossy black head of a head just visible behind the pile, we would routinely count that obscured animal as a crow. Here is how we justify that categorization: 1.) There are no other large glossy black avian species that we encounter (things would be more complicated if fish crows or ravens were in the mix of species). The only other rather black bird in the study is the common grackle, which is considerably smaller and very rare (only a single visit recorded in the study). 2.) American crows are very social, and foraging group members are commonly observed disappearing and reappearing from behind the pile. 3.) Although crows do associate with other species (e.g., red-shouldered hawks, turkey vultures, wild turkeys) at the piles, none of these other species have that characteristic black head. So as you can see, this determination is somewhat based on having had extensive experience viewing these images throughout the course of the entire experiment. I can see why you leaned toward the cautious approach of counting it as “unidentified 1.” Under a different scenario, say a very limited view of a solitary, gray/brown animal behind the pile, I would be apt to go with “unidentified,” given the higher degree of ambiguity.
If in doubt, going with “unidentified” is likely the safest approach since that will trigger us in the lab to view it, provided more than one person enters that category for a particular image.
Question 2
Here’s another puzzler. I was viewing image 2010-10-18 8:46:57 am M 8/10 as my focal image. I clicked through the consecutive images quite a bit because of my previous question that you answered below. It was similar to the image in that question in that there was a fourth crow possibility behind the pile and even though it never raised its head throughout the sequence I was more inclined to call it a crow just because more of the back was visible and its size was not as big a question mark as in my previous question. However in the attempt to apply it as an additional example to that particular type of case I happened to notice a squirrel jumping across the face of a tree in the upper left. It’s very hard to notice and I would have except for its movement in the sequence.
Should I have counted it or not? I opted not to count it even though it is clearly in the focal image (8/10 as stated above) because it is so far out of the range of normal view. I expected I’d seen it only because of the special circumstance.
P.S. For what it’s worth the sequence I clicked through for consecutive images seems to jump from 10/10 to 3/10. There is no 1/10 or 2/10.
[To see the focal image go to http://reindeer.cc.trincoll.edu/CompostResearch/Actual/R9/MIX_R9S_101018_08-46-57_V_08_10.JPG ; squirrel is on dark tree trunk in far left, hard to see without consecutive images.]
Response 2
Don’t be concerned about asking many questions. Since this is an actual research project, we want participants to approach it very carefully, being knowledgeable about what they are doing and why they are doing it. Sound like you’re set as to how to handle the crows, counting all, even those only partially visible, in your focal image, so on to the squirrel.
Our standard operating procedure is to count any and all animals that are visible in the entire focal image (directly on pile, in foreground, in background). Thus, if the squirrel were visible in the background of the focal image, it should be counted. This approach provides an objective assessment all potential visitation. Certainly, many animals in the background are not necessarily appearing due to the kitchen scraps in a treatment pile, but that will come out in the wash. Animals are recorded in the background of images of the control pile as well. When we look at effects of the treatments, we are always comparing the frequency of encounters at the treatment piles to that at the control pile, thus this “background” activity is factored out in an objective manner.
Your experience with the consecutive image numbers, mentioned at the close of your message, is not that uncommon. The database when serving up the consecutive images for a given focal image, may at times present that image outside of the context of its complete set of 10 images. Sounds like your focal image 8/10 was in the midst of a series of eight consecutive images (3/10 to 10/10). After viewing 10/10 it then loops back to 3/10.