This history page contains a photograph which is one of several belonging to the photo gallery pages which are part of several pages relating to the invention of the world's first automatic totalizator in 1913 and Automatic Totalisators Limited, the Australian company founded in 1917 to develop, manufacture and export these systems. This is a small part of a mechanical computer long before the invention of electronic computers. It is an electromechanical shaft Adder which was probably part of the AJC system at Randwick Racecourse.

Early large system adder - part of a Mechanical Computer

This is an image of an early large system adder. I deduce that it is early, as, although it is clear that it has electrical elements, it still has a foothold in the purely mechanical era, as parts of it are driven by weights. The electrical elements are not on display in this image however twisted pair wires can be seen dangling from underneath the adder in the lower part of the image. I think it is fairly safe to assume the adding shafts at the rear of this adder will be activated by electrical impulses from the ticket issuing machines probably delivered by the twisted pair wire seen dangling under this adder. As the Julius totes became electromechanical in 1917, I suspect this adder was not manufactured prior to 1917. I know that the large adders for Longchamps, which were installed in 1928, did not utilise weights, so it probably was not manufactured after 1927. The earliest of the large Julius totes was for the AJC (Australian Jockey Club) at Randwick in 1917, which was a stand-out system supporting 150 TIMs (Ticket Issuing Machines). I first thought that this adder may have belonged to that system but that is not the case, which I will explain in the postscript below. The complete Julius tote system is large, requiring a sizeable machine room, as it consisted of many of these adders along with other equipment and each adder has a table arrangement to support it as seen in the image. The machine room can be thought of in modern parlance as a mechanical computer room.

More after the image...

Click here to go Back
Click on the image to go back to the photo gallery

There are no photographer's markings on this print

On the subject of the machine room, the image below, of the Longchamps machine room titled Similar to Figure 16 gives an idea of the amount of equipment, when looking at the runner number posts on top of each machine disappearing into the distance, with more adders behind the photographer.

As mentioned above the Randwick system of 1917 was a stand-out regarding size. The Midlands Advertiser paper in West Australia records the Bombay Julius Tote as being the largest automatic totalisator in the world in 1927. The Bombay system was installed in 1925 and there is no installation in that league in between Randwick and Bombay, certainly nothing that would require the likes of the adder in the image above. This leaves the question whether this adder dates back as far as the Randwick 1917 system, as this is the year Julius Totes became electromechanical and consequently Randwick could have been purely mechanical, although I find it highly unlikely that the purely mechanical technology could have supported anything like 150 machines. My PDP11 based computer totes, which replaced Julius Totes in the Brisbane region 62 years later, only had 160 machines, although this was not a limit!

An Automatic Totalisators Limited promotion document titled The PREMIER (JULIUS) AUTOMATIC TOTALISATOR, written in 1930, has an image of The Paddock at Randwick Racecourse. The annotation for this image reads: FIG. 3. "Paddock" Installation on the Randwick Racecourse, Sydney, N.S.W. This is the first all-electric "Premier" Totalisator built and installed in 1917, and still in operation. This indicates that in today's parlance the 1917 Randwick system was the world's first electromechanical totalisator system and that electricity has become a feature in the Randwick 1917 system making it possible that the shaft adder shown above is of the type used at Randwick.


The Adder above, although being electromechanical, was NOT used in the Randwick System and postdated it:

I have just discovered that I have an original document titled The PREMIER (JULIUS) AUTOMATIC TOTALISATOR as mentioned above. I also have what I thought is a photocopy of this document which has the same front cover except it has 1930 added beneath the title. I thought these documents were the same but on having a closer look, although they cover similar ground they are very different. In this photocopy version of the document with 1930 on it, I have found new information and discovered a very low resolution image of the same type of adder as the one in the above image. Although the photo was taken from the opposite side of the adder, it is quite unmistakable and shows more of the electrical wiring and it is clear that this adder is electromechanical, utilising solenoids to trip the escapement mechanisms. The constituent 5 shaft adder assemblies and their associated storage screws visible on top of this adder above, look very like the ones in the Longchamps adders, although it is clear that this adder pre-dates the Longchamps adders which are considerably larger.

The image in the photocopy version of the document is annotated "Horse Unit of largest type machine." Horse Unit refers to the adder. The associated note reads "The horse unit of the largest type has recorded accurately over 250,000 bets per minute (see page 11). The medium size records at the rate of 3000 bets per minute, and the small machine at the rate of 1000 per minute." Page 11 goes into technical details of the working of the Adder, however one sentence states "A machine has been manufactured and shown in Sydney that records bets up to 250,000 per minute." I recognise this as the machine that George Julius, the inventor of the Julius Tote, wrote a paper about, which he presented to the Institution of Engineers Australia on Thursday May 13th 1920. There are extracts from that paper on this website in the Mechanical Aids to Calculation chapter.

In George's paper he makes two references to the Randwick system, which as it is written in 1920 has to refer to the 1917 installation. On rereading George's paper I now know that the TDMs (Time Division Multiplexers) which are an essential element of the later electromechanical Julius Totes date back to at least 1920. George calls these TDMs distributors as it was not until the electronics era that the name TDM was created. George relates that expanding the Randwick system to greater demands would be very costly to maintain and install. He also relates that the adding gear with this new system is greatly reduced, as four or six escapements can now do the same as forty escapements in the Randwick machine. So in conclusion, the type of adder shown above was not used in the Randwick system. It does appear however that it was used as part of the demonstration to the Institution of Engineers Australia and that could explain why it looks so ornate. I suspect this adder along with possible others were specifically built for demonstration and promotion purposes focused on the 1920 white paper presentation to the Institution of Engineers Australia and is The horse unit of the largest type mentioned in the previous paragraph. It is also part of the system that was manufactured and shown in Sydney that records bets up to 250,000 per minute also mentioned in the previous paragraph. This system was also capable of supporting 1000 terminals.

From George's comments in his 1920 paper I deduce this demonstration adder, which followed the Randwick system, implemented improvements gleaned from the Randwick system. The greatest enhancement which made such a high performance system possible, was the introduction of what George calls distributors, mentioned above, which led to the Julius totes having what in computer terminology would be called a front end system based on TDMs (Time Division Multiplexers.) George acknowledges in his paper that it would have been cumbersome and costly to expand the Randwick system without these newly invented electromechanical TDMs.

George's comments I mentioned above appeared in his white paper titled Mechanical Aids To Calculation and extracts of his two comments follow:

In considering the application of the equipment to meet very much greater demands, it was apparent that the very large amount of gearing required under what may be called the Randwick system would be very costly both to install and to maintain. A modification of the system has therefore been developed, and may be briefly described as follows :-

This modification has very greatly reduced the amount of adding gear required in the machine, as in the new type four, or at most six, escapements perform the same duty as was previously performed by the forty escapements in the Randwick type of machine.

Finally, George's paper refers to a unit built to meet the condition of betting on the largest French racecourses which refers to a system utlilising the above type of adder. After seeing the adder above from the opposite end it is obvious that this adder could be the basis for the Longchamps adders.

I find this adder particularly interesting as it is so ornate particularly the legs of the table framework that supports it. I have often felt that engineering and technology products have an artistic element to them. They often are visually attractive imparting to them an artistic quality. It is an additional bonus to think that these attractive artworks were technology devices performing a function. In the case of much of this Julius Tote equipment when looking at these artistic pieces, it is additionally interesting to contemplate that they had been functional for so long, decades and in some instances almost half a century and in one known case, the Caracas system, the full half a century. To my mind a lot of Engineering Drawing has an artistic quality as well.

The five tubular shaped devices on the top of the adder extending from the rear to the front are called storage screws, a mechanical form of memory. They remember transactions recorded by the high acceleration adding shafts at the rear of the adder and transfer them to the inertia limited parts of the system at the front as they are able to catch up. The electronic counterpart to the mechanical device, consisting of the storage screw and the feedback path resulting from sensing of the position of the screw itself when in the vicinity of its rest position, so that the inertia limited parts can be slowed down and brought to a gentle stop, is a closed loop servo system. The rods extending from the centre of the near side of the tubular bodies of the storage screws are part of the screw position sensing. The storage screws are dealt with in more detail in other images in the photo gallery.

Following is an extract from a company document mentioned above, titled The PREMIER (JULIUS) AUTOMATIC TOTALISATOR, which is the ninth extract, immediately following the extract in the image page which appears in the second page of the photo gallery under the heading Ticket Issuing Machines (TIMs), with associated text starting The type of TIM in use at Randwick Racecourse in 1927. It relates to the Julius Tote Shaft Adders, called Adding Machines or Adding Units in this extract.

Adding Machines

The adding units total all the bets made, almost as quickly as tickets are issued by the ticket issuers, the maximum time that may elapse between the instant of printing a ticket and registering its sale being less than two seconds, even during very heavy betting periods. THESE UNITS ARE CAPABLE THEREFORE, OF RECORDING, PRACTICALLY INSTANTANEOUSLY, ALL THE TICKETS SOLD BY ALL THE ISSUERS, EVEN SHOULD SUCH ISSUERS ALL ISSUE A TICKET ON THE SAME HORSE IN A RACE AT THE SAME INSTANT. One "adding unit" is provided in any totalisator installation to total the investments on each horse, so that there are as many adding units at work as there are starters in the race. In addition, a further, and usually a somewhat larger, unit is installed which automatically totals the whole of the investments on the race--that is to say, its record is the sum of all the records on the various horse adding units.

In small installations, as before mentioned, the adding unit and the indicator are as a rule combined, particularly where independent auxiliary indicators are not required, but on large installations the adding unit is entirely distinct from the indicator and may be installed in any suitable machine room, far removed from the indicators, should such be found desirable. In Fig. 15 a view is given of the largest individual adding unit of the Premier type yet built, and this unit, which is installed in Longchamps Racecourse, Paris, is of very much greater capacity than any other unit in use, whether of the "Premier" or of any other type, and in Fig. 16 a view of a group of such adding units is shown.

Figure 15 An image of a large shaft adder at Longchamps Racecourse

The above image is the same as Fig. 15 in the company document which has the associated text: "Premier" Totalisator Adding Unit installed at Longchamps. Paris. Note that the adder in the image above is much more modern than the one shown at the top of the page. This adder has no driving weights. This is a reduced image with the full sized version in the photo gallery, several images below the entry for the image at the top of this page.

Similar to Figure 16 An image of the Longchamps Machine Room

The above image is similar to Fig. 16 in the company document, the main difference being that the one in the document has no people in it. The document has the following associated text: Group of "Premier" Adding Units at Longcamps. Paris. This is a reduced image with the full sized version in the photo gallery also several images below the entry for the image at the top of this page. Click on the image at the top of this page to go to the photo gallery directory. This is a good image showing the mechanical computer referred to in the title of this page. You can see the shaft adders extending into the distance and there will be adders behind the photographer as the nearest runner adder number visible is 29 and the system supported 42 runners. This is an interesting historic point in that all the totalisators I worked on, mainly in Australia, which were all computer based had a maximum field size of 24. As the functionality was implemented in software, this limit was a configurable variable set at build time, but the systems I worked on were built with this set to 24. Back to the extract:

Where units of this capacity are used, independent indicators are required to display to the public the progress of the betting, and Automatic Totalisators Limited can supply indicators of various types as more fully described on Pages 16 and 17. In smaller installations a combined adding and indicating unit is used, and this type has been installed on many Racecourses.

Such a unit is shown in Fig. 17, the machine being so arranged that the numbers displayed on the front of the equipment are made visible to the public through suitable windows in the totalisator building. A group of such units, is installed on the West India Turf Club's Course, in Bombay, is shown in Fig. 18.

Similar to Figure 17 An image of a combined adder and counter display

This image is very similar to the one in the company document and as Fig. 18 in the document shows the West India Turf Club's Course, Bombay system utilising these units, it is appropriate that the unit in the above image also belongs to that system. The annotation in the company document for this unit is: Combined "Premier" Adding and Indicating Unit for displaying numbers of Tickets sold as installed on various Racecourses.This too is a reduced image with the full sized version in the photo gallery a few images below the entry for the image at the top of this page. Back to the extract:

A simpler type of "Premier" equipment has recently been perfected, particularly suitable for smaller racecourses, and in these machines results are displayed by means either of dials and moving pointers or by a "barometer" type of Indicator, either of which system will inform the public quite sufficiently closely regarding the progress of the betting, whilst at the same time an exact count of the number of bets on each horse is recorded on a small counter on the unit itself, from which dividends can be calculated. Such a unit is shown in Figs. 19 and 20, and Fig. 19 shows in particular the lightness and compactness of this equipment, which is capable of recording all sales made on as many as 48 selling machines.

All Premier adding units are equipped with safety devices, which prevent errors in the recording of the betting should any unexpected failure of the mechanism occur, either mechanical or electrical. Such protective devices operate in such a way as immediately to stop the betting on any ticket selling machine where the record of such betting is not being correctly transmitted to and recorded by the adding unit. WITH THE PREMIER EQUIPMENT, THEREFORE, NO TICKETS CAN BE ISSUED WITHOUT ACCURATE AND PROMPT RECORDING BY THE ADDING MACHINE.

As previously described, Premier equipment is now available, and is, in fact, in use by means of which the result of the betting can be displayed to the public in terms of the "expected dividend" or "odds" on the horse, in place of the more usual and older method of recording only the "number" of tickets sold.

The next extract from this company document, which immediately follows on from this extract is below the entry for the image at the top of this page in the photo gallery under the heading Miscellaneous Images. To read it, click on the image at the top of this page, scroll down in the photo gallery index to the previously mentioned heading and select the image thumbnail with associated text starting The world's first Odds Computer....