Since 1995 I collect data on old bicycles. By October 1999, the resulting database contained data of more than 3,000 bicycles. 87% of these are Dutch, 7% are German, 4% are British and 2% are of a different origin. The spread of the data by year of construction is shown in the next figure.

Registered bicycles spread by year

The values for the period up to 1959 mirrors -in my opinion- a valid picture of the actual today's 'on the street' situation.
As the figure shows, the emphasis lies on Dutch bicycles of the fifties and sixties. Of this category, approximately 1 of 5,000 bicycles ever produced is registered in the bicycle database.

Initially, the database can be used to determine the year of construction by means of comparison. Apart from making lists of bicycles that are still provided with an original rear hub with a stamping of the year of manufacture, the year can be estimated by comparing certain parts. An example: 'Koets A'-type rear lights were almost exclusively fitted to Gazelle, Juncker, Fongers, and Locomotief bicycles older than 1955.

For Gazelle, there is a list of frame numbers and the corresponding year of manufacture which ends in 1950. The chart on this page includes 100 Gazelle bicycles of the period after 1950. The formula can be used to derive the year corresponding to a frame number. The resulting year is usually correct within a margin of one year.
The values shown in the chart represent the year of construction of the hub. The frame will be assembled some months later. In the formula this is taken into account. A resulting year, e.g. 1966.0, should be interpreted as 'middle 1966', thus not as January 1966, but as June 1966.

Dating Gazelle bicycles between 1950 - 1973

If a headset label is missing and no brand name can be found, specific characteristics may be the way out. What kind of frame number, where is it stamped-in, type of bottom bracket, type of headset-lugs etc. Sorting and selecting these data with the aid of a powerful database application, it can be possible to deduct the brand.

The database offers information about the approximate production rate of specific brands, and sometimes about which factory produced what kind of bicycles.
During the merger of Locomotief and Simplex (between 1952 and 1965), these brands sold 40% gents’ models and 60% ladies’ models. With the Gazelle company this ratio was the same. While L & S produced 25% roadsters (rim size 28 x 1 1/2 inch), with Gazelle this type of bicycles covered almost 40%. Moreover, L & S sold strikingly many ‘budget-bikes’ equipped with freewheel and caliper brakes. Of course, these figures assume that the bicycles still 'on the road' today give a correct image of what was sold then.
But there is more general information to be derived from this database: aluminium rims for sports and roadster bicycles were introduced in 1948. (Racers already used these in the thirties.) From 1949 onwards, these rims -mostly produced by Lepper- can be found on many Dutch bicycles of almost every brand. After 1964, this hype was more or less over.
Another example: the 45 mm Thompson bottom bracket, which was a typical Dutch feature. This type of bottom bracket vanished around 1960. Phoenix and Germaan stopped production at about that time, Juncker did so already ten years before, whereas Gazelle, Union (43 mm) and Veeno used these brackets for another couple of years. Ten years later, the same happened to the roadsters' seat stays made from ovalized tubes.
Another interesting subject is the frame geometry of these bicycles. The head-tube of the loop-frame type ladies' bicycles was mostly 270 to 280 mm long, making them the 'real granny's bicycles' with very comfortable steering properties. Fongers used head-tubes of a meagre 243 mm long (al least up from the twenties of the past century). Juncker and Union 'shortened' the measure of theirs in about 1950, Phoenix did so some years earlier, while Gazelle followed suit in two steps in 1956 and 1959. Some other brands made bicycles with a head-tube of even 300 mm long!

Rare brands might be troublesome to date reliably with the aid of the bicycle database, because of the small number of bicycles still 'alive'. However, for the purpose of brand identification data of these bicycles can be very useful. But even the well-known brands can be problematic, especially when the system of frame numbering is complicated.
Juncker, Locomotief and Simplex, who finally merged, give a good example. Until the late sixties, all of them used more than one numbering sequence, additionally using a coded system. By collecting as much frame numbers of these brands as possible and comparing the bicycles' characteristics, even that could be deciphered.