chine design occupies the greater part of our drawingroom time. The strength of materials and the mechanics of materials are taken as parallel subjects; we would prefer, of course, to have these precede the machine design, but as our curriculum is outlined we cannot do so. Concerning the assignment of work, I would not have you understand that these problems as given in the paper are now used; they merely represent typical problems and should be modified from year to year. Our method of assignment is usually to refer to some machine as given in a standard catalog, and after stating a few conditions, ask the student to design a machine after that general plan, making such detailed modifications as he prefers.

SOME EXAMINATION DATA.

BY R. D. BOHANNAN,

Professor of Mathematics, Ohio State University.

The freshman class in engineering at the Ohio State University takes college algebra the first term, plane trigonometry the second term, and analytical geometry the third term. The sophomores take a mild dose of both differential and integral calculus the first term, differential calculus the second, and integral calculus the third. Each class meets five times a week for fifty-five-minute recitations.

To each student at the beginning of each term printed lesson slips for the term are handed. These lesson slips designate the bookwork. In this way all men in any class have the same daily work and the same term's work, no matter who the instructor is. There were this last year twelve sections of freshmen and seven of sophomores, with an average of about twenty-five to a section. The lesson slips specify on what days the recitation hour will be turned into an examination. Four such "midterms" are given in each term. There is held at the end of each term a "final" examination, lasting four hours.

To determine a student's standing at the end of a term, his average standing in the four midterms is added to his standing in the final and the sum divided by two. The midterms are thus given the same weight as the final.

Examination papers are made out by an examina

tion committee, consisting of three teachers. One man goes off this committee each year and a new man goes on. All sections of the same class reciting at the same hour are set the same paper. All sections have examinations on the same day and of about equal difficulty. Midterms for any section are graded by the instructor of that section, and the marked papers handed back to their authors as early as practicable. Frequently the paper is discussed at the next recitation following the examination. Appeals from markings are considered by the instructor or by the department secretary.

Final examinations are all graded by the examination committee and not by the instructor. The final mark for the term's work for a freshman is either "passed" or "failed." No "conditions" are given freshmen. Those who fail in algebra the first term may repeat the work the second term. Those who fail in trigonometry the second term may repeat it in the third term. Those who fail in analytical geometry in the third term may repeat it in the summer school, or the next fall term, or the next spring term.

By recent faculty action, no final examinations, at any other time than the regular recitation hour, may be set any class at the end of the winter term. This raised with us the question as to whether all finals might not be omitted and the average of our four midterms be taken as a fair test.

We examined all standings for the last two years. The result for last year was almost exactly the same as for this year, which is as follows for freshmen. The first pair of columns is what the finals would

have done for students in each of the twelve sections if we had trusted to midterms alone for the grades. The second pair of columns shows in the same way what the midterms would have done had we trusted to finals alone.

The men affected were, as a rule, "on the fence." The finals shifted 35 men, failing 27 who would have passed on midterms alone, and passing 8 who would have failed. Net loss, 9 men. If we had trusted to finals alone, 28 men would have shifted, 27 of them going from the passing side to the failing list. The number of men concerned was 273.

The winter and spring terms were as follows, the trailers being accidentally omitted from my notes at hand in the third term. There were three sections of them. The winter term relates to 256 men and the spring to 195.

Thus in 722 decisions, the effect of finals would have been to reverse 85 decisions that would have been made on midterms alone, with a net change of 15 men less in the passing list. If we had trusted to finals alone, the midterms would have reversed 59 decisions, passing 57 who would have failed.

What conclusion you draw from this depends on what view you have of the "man on the fence."

* Trailers omitted.

With