Roberts Variable Star Project

It is now about 9 months since Tim Cooper and I started on this project and I thought it would be a good idea to inform those interested, and especially the members who have helped us in the data capturing exercise, of what has been achieved to date.

Alexander Roberts was a teacher and an amateur astronomer who spent some thirty years, from 1891 to 1921, observing variable stars from the Lovedale college near Alice in the E Cape.  He became a member of the Senate in the SA parliament, at which point it appears that he stopped his variable star observing.

In an article written by Sir David Gill, Astronomer Royal at the Cape, it is mentioned that Roberts did 250,000 variable star observations.  This figure has been repeated in a number of articles written about Roberts.  Amongst Roberts’ papers is a draft of a letter that he wrote in search of sponsorship to publish his observations in which he mentions having done a million observations.  Quite a feat!!  Well, if it were true it would be quite a feat.  We shall have a good look at how these figures arose in due course.

Roberts observed around 100 variable stars - Algol variables, Cepheids, Miras and some semi-regular variables.  He also observed some of the novae that were around and did a huge amount of mathematics involving double stars, including a determination of the period of Alpha Centauri, getting a period of about 81 years (compared to a modern value around 79 years).  He published many papers on all aspects of astronomy.

In the Roberts papers, which are at Boyden observatory (some 40,000 pages), we have found the reduced observations for about 98 stars.  Of these we have now captured the data for 82 stars, giving us a total number of observations around 57,000.  Most of this hard work was done by Tim’s dad with some of the local members also chipping in and here I would like to say thank you to Atze Herder, Chris Reitz,  Melvyn Hannibal,  Herman Wiechers and  Richard van Coppenhagen for their assistance.

Validating this data and checking that it is all correct has been a time-consuming task.  We found that Roberts himself made quite a number of errors in calculating Julian dates as well as in transcribing his observations onto the summary sheets.  And with the odd typing errors in the capturing exercise and difficulty reading some figures we have had to be very careful in checking the data.  But we think we are getting there.  In 2 or 3 months we may have completed the data input exercise.

So where does the discrepancy in the numbers come from?  Was Roberts some kind of a cheat?  I don’t think so.  The explanation lies in the rather strange method that Roberts used to do his observations - very different to that used by modern day variable star observers.  But this is an aspect of the story that demands a full explanation and an article of its own.

Brian Fraser