In commenting on my first post about using shadows to date photographs Michael Smith asked a number of questions about the process. This post is a response to those questions (the questions are marked in italics below.)
When discussing our dating experiments at the University of New England and Regional Archives, I should first point out that I have the services of a most talented volunteer with a background in both history and information technology. This gentleman is Mr Michael Reed who is refining his expertise in the technique based on the book:
“Forensic Genealogy” by C Fitzpatrick ISBN: 9780976716006 (This book contains the CD to generate the math.)
This is a sample image that we are going to work on. The clothing and equipment gives us an approximate date range and then we have the exact location of the site. The question here revolves about whether we have enough information to judge the length of our shadows.
(1) Did you already know the date of the photos you tested? i.e. Did you test first on known dates to see how accurate the calculations are?
We had an estimated date range based in information gained from details of the award of the Victoria Cross to Jackson and his service records. This placed us to within 18 months of the event. We then looked for shadow and applied the algorithm that produced a table of possible dates. Then we were able to seek newspaper confirmation of the event.
(2) How many dates does the algorithm spit out? (I’m sure it varies but ball park)
The algorithm is basically a reverse sundial producing many dozens of possible dates over many years from the photo location and the length of shadow.
(3) Do you need to know where the photo was taken to get an accurate date. We use a range of factors to identify dates in our photographic collection. This can be very time consuming and not always terribly accurate. And of course there are some where I suspect we will just never know.
The exact location of the photo is critical to the process. Without knowing your location on the planet, you cannot work out the location of the sun to get the date and time.
Finally, shall we see this written up in more detail in, say, a distinguished journal, at some stage?
I think that Michael Reed would be the potential author of any serious journal article generated out of this work. His understanding of the issue is complimented by his available time to experiment. I am just concentrating on just locating images that have both sun and location to give ourselves more practice at the technique.
This is an example of a photo that would be more challenging to date using this technique although here the date is a given. This image ALA22 provides dating evidence in its titling. However, none of the buildings in this photo still exist. ALA22 does not have any real hint as to the location of north even though I know the city block where this was taken. I suspect that Michael would delve into the tree line in order to identify this. When using a sundial in reverse, you have to know exactly where it stands. I like this photo.