210
10 Statistics and Other Concerns
the kind of distributions concerned. The instrument has been lost. However,
there is a good picture of it in a publication by Kapteyn & Van Uven on skew
distributions [92], which has been reproduced in Fig. 10.3 (right). Recently
Florian L.R Lucas and Simon A.M. van der Salm, chemist and mathematician,
have built a replica of the machine with financial support of the Royal Natural
Sciences Society in which Kapteyn had been so active and donated it to the
University Museum Groningen. Actually, the Museum contributed equally and
Lucas and van de Salm put in money out of their own pocket. Their publication
on this contains a thorough treatment of its workings and properties of the
log-normal distribution [93].
In 1903 Kapteyn gave a lecture for the biologists at the University of Groningen with the title Skewed frequency curves in biology and statistics. The lecture
was actually given in English, at least the text of the published version was
in English. He began to say that he had been ‘asked by several students and
some other persons to discuss the statistical methods of Quetelet, Galton and
Pearson’. The Belgian Lambert Adolphe Jacques Quetelet (1796–1874) and
the British Francis Galton, the designer of the quincunx, and especially Karl
Pearson (1857–1936) had developed and applied statistical methods for probability distribution and correlation between measurement in social and other
sciences. In Groningen Jan Willem Moll (1851–1933) was professor of botany
since 1890 and Jantina (‘Tine’) Tammes (1871–1947) was his assistant since
1897 (see Fig. 10.4). Jantina Tammes would become the first female professor
of genetics in 1919. Moll had studied in Amsterdam with the famous biologist
Hugo de Vries (1848–1935), who was very impressed by the use of statistics.
De Vries and Kapteyn were to meet each other in 1904 in St. Louis as the only
two Dutch participants at the Louisiana Purchase Exposition.
Kapteyn fulfilled the wish of the biologists, although he complained to
others (e.g. in letters to Gill) that it took him too much time. Eventually he
published the lecture. He described three applications. The first concerns data
from his friend, the psychologist Gerardus Heymans, and dealt with thresholds
in people’s sensations. The second was from a publication by Pearson on house
prices in England and Wales, and the third was a data-set on the size of crabs’
foreheads in Naples. In the end he showed that these distributions followed
the log-normal very well.
The article, which as mentioned was written in English, was apparently
also read outside the Netherlands, because in 1905 Pearson wrote a reaction to
Kapteyn’s work in very harsh terms. In the first place, Kapteyn had not referred
to the work of Francis Ysidro Edgeworth (1845–1926), who had developed a
comparable statistic method as Kapteyn. Furthermore, he criticized a number
of mathematical matters and especially Kapteyn’s proposition that the form
of a distribution function should contain causal information about its origin.
10 Statistics and Other Concerns
the kind of distributions concerned. The instrument has been lost. However,
there is a good picture of it in a publication by Kapteyn & Van Uven on skew
distributions [92], which has been reproduced in Fig. 10.3 (right). Recently
Florian L.R Lucas and Simon A.M. van der Salm, chemist and mathematician,
have built a replica of the machine with financial support of the Royal Natural
Sciences Society in which Kapteyn had been so active and donated it to the
University Museum Groningen. Actually, the Museum contributed equally and
Lucas and van de Salm put in money out of their own pocket. Their publication
on this contains a thorough treatment of its workings and properties of the
log-normal distribution [93].
In 1903 Kapteyn gave a lecture for the biologists at the University of Groningen with the title Skewed frequency curves in biology and statistics. The lecture
was actually given in English, at least the text of the published version was
in English. He began to say that he had been ‘asked by several students and
some other persons to discuss the statistical methods of Quetelet, Galton and
Pearson’. The Belgian Lambert Adolphe Jacques Quetelet (1796–1874) and
the British Francis Galton, the designer of the quincunx, and especially Karl
Pearson (1857–1936) had developed and applied statistical methods for probability distribution and correlation between measurement in social and other
sciences. In Groningen Jan Willem Moll (1851–1933) was professor of botany
since 1890 and Jantina (‘Tine’) Tammes (1871–1947) was his assistant since
1897 (see Fig. 10.4). Jantina Tammes would become the first female professor
of genetics in 1919. Moll had studied in Amsterdam with the famous biologist
Hugo de Vries (1848–1935), who was very impressed by the use of statistics.
De Vries and Kapteyn were to meet each other in 1904 in St. Louis as the only
two Dutch participants at the Louisiana Purchase Exposition.
Kapteyn fulfilled the wish of the biologists, although he complained to
others (e.g. in letters to Gill) that it took him too much time. Eventually he
published the lecture. He described three applications. The first concerns data
from his friend, the psychologist Gerardus Heymans, and dealt with thresholds
in people’s sensations. The second was from a publication by Pearson on house
prices in England and Wales, and the third was a data-set on the size of crabs’
foreheads in Naples. In the end he showed that these distributions followed
the log-normal very well.
The article, which as mentioned was written in English, was apparently
also read outside the Netherlands, because in 1905 Pearson wrote a reaction to
Kapteyn’s work in very harsh terms. In the first place, Kapteyn had not referred
to the work of Francis Ysidro Edgeworth (1845–1926), who had developed a
comparable statistic method as Kapteyn. Furthermore, he criticized a number
of mathematical matters and especially Kapteyn’s proposition that the form
of a distribution function should contain causal information about its origin.