Prasanta Chandra Mahalanobis

Prasanta Chandra Mahalanobis was born on June 29, 1893, in Calcutta (now Kolkata), into a family closely tied to the Brahmo Samaj, the reformist religious movement that had reshaped Bengali intellectual life in the nineteenth century ^[1]. His grandfather, Gurucharan Mahalanobis, had built a successful chemist's business and was an active member of the Sadharan Brahmo Samaj, while his father, Prabodh Chandra, ran the family firm and kept the household connected to Calcutta's community of reformers, teachers, and scientists ^[2].

The boy grew up in a house where debate about education and social change was ordinary dinner conversation. He attended the Brahmo Boys School in Calcutta and then entered Presidency College, where his teachers included the physicist Jagadish Chandra Bose and the chemist Prafulla Chandra Ray, two of the most celebrated Indian scientists of the era ^[2]. He graduated with honours in physics in 1912.

The next year he sailed for England, intending to study at London University. A chance delay changed his life: while visiting Cambridge he was so taken with King's College that he sat for its entrance examination and was admitted ^[1]. At Cambridge he read mathematics and physics, took part I of the mathematics tripos and part II of the natural sciences tripos, and crossed paths with the self-taught mathematical genius Srinivasa Ramanujan, then resident at Trinity College ^[2].

Anyone asking who was Prasanta Chandra Mahalanobis before statistics claimed him would have met a promising experimental physicist. After finishing at Cambridge he was set to work under C. T. R. Wilson at the Cavendish Laboratory, but a short trip home in 1915 became permanent when the First World War disrupted travel and he accepted a teaching post in physics at Presidency College ^[1].

The decisive accident had already happened. Just before leaving Cambridge, his tutor introduced him to Biometrika, the journal of statistics founded by Karl Pearson. Mahalanobis bought a complete set of back volumes and read them on the voyage to India ^[2]. He recognized in the new discipline a set of tools that could be applied to problems all around him: anthropology, meteorology, agriculture, flood control.

Through the 1920s he worked on statistical problems in his spare time from a single room in the Physics Department at Presidency College. His analysis of anthropometric measurements of Anglo-Indians in Calcutta led him, in a paper published in 1936, to define a measure of divergence between populations now known worldwide as the Mahalanobis distance, a scale-invariant way of measuring how far a point lies from the centre of a multivariate distribution ^[3]. He also carried out influential studies of flooding in Orissa and Bengal; his 1926 analysis of the Orissa river systems anticipated conclusions later used in planning the Hirakud dam on the Mahanadi ^[2].

The most durable of Prasanta Chandra Mahalanobis achievements is institutional. On December 17, 1931, he and a small circle of colleagues formally founded the Indian Statistical Institute in Calcutta, registering it as a learned society in April 1932 ^[4]. What began as a one-room operation with a part-time staff grew into a research and training centre of international standing. In 1933 the Institute launched Sankhya, the Indian Journal of Statistics, which Mahalanobis edited for decades and which remains a respected research journal ^[4]. In 1959 Parliament recognized the ISI as an Institution of National Importance, empowering it to award its own degrees ^[4].

His second great contribution was to the theory and practice of large-scale sample surveys. Beginning with crop surveys in Bengal in the late 1930s, Mahalanobis showed that carefully designed samples could estimate agricultural output over vast areas at a fraction of the cost of a full census ^[3]. He introduced the method of interpenetrating network of samples, in which independent subsamples are collected by separate teams so that their agreement or disagreement reveals the size of survey errors ^[3]. R. A. Fisher, the leading statistician of the age and a personal friend, praised the ISI's pioneering role in this field. The National Sample Survey, established under Mahalanobis's guidance in 1950, became the Indian government's principal instrument for measuring household consumption and employment, and served as a model studied by statistical agencies abroad ^[5].

Recognition followed on both sides of the disciplinary line he straddled. He was elected a Fellow of the Royal Society of London in 1945, served as chairman of the United Nations Sub-Commission on Statistical Sampling from 1947 to 1951, and received the Padma Vibhushan, India's second highest civilian honour, in 1968 ^[2]. He was also honoured with the Weldon Medal from Oxford University in 1944 for his work in biometric science ^[2].

After independence in 1947, Prime Minister Jawaharlal Nehru drew Mahalanobis into the centre of economic policy. He became honorary statistical adviser to the Cabinet in 1949 and a member of the Planning Commission, where his influence peaked in the mid 1950s ^[5].

His two-sector growth model, published in Sankhya in 1953 and extended to four sectors in 1955, supplied the analytical frame for India's Second Five Year Plan (1956 to 1961) ^[5]. The model, which had a family resemblance to earlier work by the Soviet economist G. A. Feldman, argued that a poor country wishing to grow quickly should direct a large share of investment into heavy industry: steel, machines, and the capacity to make more machines. The strategy became known as the Nehru-Mahalanobis approach and dominated Indian planning for a generation ^[5].

The plan drew criticism then and since. Some economists argued that it neglected agriculture and consumer goods and contributed to the foreign exchange crisis of 1957, while others defended it as a coherent response to India's circumstances ^[5]. Whatever the verdict, few statisticians anywhere have exercised comparable influence over the economic direction of a large nation, and the debate over the model is itself part of any honest Prasanta Chandra Mahalanobis biography.

In 1923 Mahalanobis married Nirmal Kumari, known as Rani, the daughter of the educationist Heramba Chandra Maitra. The marriage faced initial resistance from her father, and the ceremony was held with Mahalanobis's uncle standing in for the family patriarch ^[2]. Rani became his lifelong companion and collaborator in the running of the Indian Statistical Institute; the couple had no children.

Mahalanobis kept one foot outside science all his life. He was a devoted friend and occasional secretary to the poet Rabindranath Tagore, accompanying him on journeys abroad and serving for years as secretary of Visva-Bharati, the university Tagore founded at Santiniketan ^[2]. He wrote on Tagore's work in Bengali and helped manage the poet's practical affairs, an association that gave the statistician a standing in Bengali literary culture unusual for a man of numbers.

Colleagues remembered a commanding, sometimes autocratic personality, nicknamed simply 'the Professor' at the ISI, who ran the Institute in personal style, recruited brilliant young people regardless of formal credentials, and welcomed a striking parade of foreign visitors to Calcutta, from R. A. Fisher and Norbert Wiener to J. B. S. Haldane, who joined the Institute's staff for several years ^[4].

Mahalanobis never retired in any ordinary sense. He remained Director of the Indian Statistical Institute into his late seventies, continuing to travel, lecture, and steer the Institute's expanding programmes in statistics, computing, and economics ^[4]. The ISI had acquired one of India's first electronic computers in the 1950s, a purchase he championed as essential for the data work of a planned economy ^[4].

In his final decade he developed what he called Fractile Graphical Analysis, a distribution-free method for comparing two sets of economic data, such as consumption patterns from different survey rounds, which he applied to questions of living standards in India ^[3]. He also continued to argue publicly for the role of science and statistics in national development, a theme he had pressed since the 1930s.

He died in Calcutta on June 28, 1972, one day short of his seventy-ninth birthday, still holding office at the Institute he had built ^[1]. Rani Mahalanobis survived him; their home, Amrapali, and his papers passed into the care of the ISI.

The clearest measure of his afterlife is in daily scientific practice. The Mahalanobis distance appears in classification, clustering, outlier detection, and pattern recognition across statistics, machine learning, archaeology, and finance, cited in thousands of research papers every year ^[3]. Students meet his name in textbooks long before they learn anything else about him, which makes basic Prasanta Chandra Mahalanobis facts, such as his training as a physicist and his founding of an institute in colonial Calcutta, a surprise to many.

India has fixed his memory in its official calendar. In 2007 the Government of India designated June 29, his birthday, as National Statistics Day, marked each year with events highlighting the role of statistics in planning and policy ^[6]. The Indian Statistical Institute, with its headquarters in Kolkata and centres in Delhi, Bangalore, Chennai, and Tezpur, remains among the world's leading institutions devoted to the subject, and the National Sample Survey he designed continues, in successor form, to generate the household data on which Indian policy debate depends ^[4].

Historians of science treat him as the central figure in what has been called the Indian school of statistics, a tradition that also produced C. R. Rao, who began his career under Mahalanobis at the ISI ^[2]. His combination of theory, fieldwork, institution building, and policy influence had few parallels in the twentieth century, and the questions he worked on, how to measure a society honestly and plan with the results, have not aged.