Journal of Heredity 2004:95(1)
© 2004 The American Genetic Association 95:93-94
Obituary |
Howard Levene Remembered
University Distinguished Professor, Emeritus Department of Biology Virginia Polytechnic Institute and State University Blacksburg, VA 24061
Howard Levene, professor of mathematical statistics and genetics at Columbia University, died on July 2, 2003, at the age of 89. Upon graduating summa cum laude from New York University in 1941, Howard joined a research group working on military projects at Columbia University. He received his Ph.D. at Columbia and soon after joined the faculty.
I do not intend to assemble here a résumé of Howard's life. Rather, I intend only to recall meeting him during the academic year 1946–1947 when, released from military service, I returned to Columbia where I completed my graduate studies under Th. Dobzhansky. Howard had already collaborated with Dobzhansky on studies of Drosophila populations. In addition, his task was to teach statistics to geneticists and genetics to statisticians.
Howard's statistics class in the academic year 1946–1947 contained several future notables; Ruth Sager and Colin Pittendrigh come to mind. At that time, however, we all were young and innocent. Having convinced Howard that we wanted to learn the theory as well as the application of statistics, we watched in dismay as he jotted an upside-down capital "L" on the blackboard: the gamma function. Our ignorance prompted Howard to abandon any effort to teach us theory. One requirement for his course, a requirement paid for under the GI bill, I still possess: a Dietzgen Maniphase Multiplex slide rule, a work of art.
The material Howard taught was eminently practical. The usual calculation of the possible range of values for small observed frequencies often included zero, and even negative values. Howard provided a formula that generated the true range of possible values, a range that excludes zero if even one event has been observed. He stressed the use of sign tests when "greater than" or "less than" were easily determined but when the actual enumeration was time-consuming. Make half again as many tests as originally intended, and then ask whether +'s and -'s are equally frequent.
Howard covered the Poisson distribution but, other than Ruth Sager and others of Francis Ryan's collaborators, few of us dealt with data that lent themselves to a Poisson analysis. My own baptism took place in Cold Spring Harbor under the influence of phage and bacterial geneticists.
While studying Drosophila populations at Cold Spring Harbor, I frequently sought Howard's advice—advice that was always readily given. I did not consult him, however, when I became displeased with Dobzhansky's calculation of the frequency of chromosomes that, when homozygous, lowered the viability of their carriers. I had access to data that allowed me to estimate the contribution of different chromosomes, different cultures, and statistical errors to total variance. Electronic computers were not then available; calculations were carried out on primitive desk calculators. Faced with the horrendous task of computing 
/
for each of hundreds of cultures, I opted to compute 
for groups of cultures. Subsequently I attended a seminar given by Howard at Columbia in which he addressed the problem that had confronted me. His conclusion: the use of was the proper method of analysis. Turning to me, he added: "It is not right that an analysis that is chosen from laziness should be the correct choice!"
Of the articles that have appeared following Howard's collaboration with Dobzhansky, I rank most interesting today that which appeared in the American Naturalist (1959;93:347–353), "Possible genetic difference between the head louse and the body louse." Head and body lice differ. They differ sufficiently in morphology that some taxonomists regarded them as separate species; not all taxonomists agreed, however. When head lice are forced to live on the body (confined to cages that are attached to the forearm) they become transformed into body lice. The transformation is accompanied by high juvenile mortality. Howard and Dobzhansky suggested that the species is polymorphic at several gene loci and that the contrasting homozygotes are adopted to the two modes of existence: head and body. Their suggestion at the time was only that—a possibility. The molecular techniques that are now available make it possible to resolve this all-but-forgotten matter.
My recollections of Howard Levene and of the era during which we were young—and our many colleagues of that era were young as well—provide me with pleasure. I am both pleased and honored to contribute these recollections in memory of an instructor, colleague, and friend: Howard Levene.
|
CiteULike 
Connotea 
Del.icio.us What's this?
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||