to his professional success in the long run, and not merely serve to drill him to perform a little better the duties of assistant in such engineering work as is liable to fall to his lot just after his graduation. Indeed, no greater mistake can be made in the engineering education of a young man that to sacrifice a thorough grounding in the scientific principles that underlie his profession for the sake of such a drill.

It is so seldom that a thesis which forms one of the exceptions referred to ought to be allowed, that no further reference will here be made to them; the remainder of the paper will be devoted to such theses as do involve original investigation. Probably the two largest classes of theses of this kind are: first, those involving experimental work; and second, designs. The number of subjects of the first class that can be chosen is enormous, as there are so very many matters in engineering that badly need experimental investigation. The only difficulties that can arise about adopting any subject which it it is desirable to investigate are the questions whether the time at the disposal of the student will admit of his making any substantial progress in the investigation of the subject proposed, and whether the facilities are available to do good and thorough work.

The investigation may be of such a character that it can be made by means of apparatus already existing in, or that can be constructed for, the laboratory, or it may require facilities that can only be had outside of the laboratory, as for instance, when a series of tests is made of some large engine, as an engine of some

steamer, or that of some industrial establishment. In either case the thesis is entirely under the control of the professor in charge, and he ought of course to insist upon careful and thorough work, and to exercise the supervision necessary to secure it.

In the second case, however, a careful and thorough performance of the work often requires that certain privileges should be granted by the owners of the plant, as the work is very liable to interfere more or less seriously with the profitable work of the plant for the time being, or with the convenience of the owners. In such cases the student should not be allowed to undertake the subject unless a perfectly clear understanding is first had with the owners, and it is certain that all the privileges will be unhesitatingly granted which are needed in order that the work may be properly done. As a rule these privileges can only be secured if the owners are interested in the work, and consider that their own immediate advantage will be subserved by knowing the results. Hence it is that there is almost always more or less difficulty about performing this kind of thesis work outside of the laboratory, and it is therefore very important that the laboratory should be equipped as fully as possible with apparatus of such a character as to permit of the investigation of live engineering questions.

As to the various kinds of experimental investigations that can be taken up, those that continually press upon us for solution are so numerous that it seems hardly necessary to attempt any enumeration of them. In questions of the strength of materials

alone, there is an enormous amount of investigation that should be made. We need more light (which can only be derived from experiment) upon the strength of iron or steel columns; the strength of riveted joints either for a tensile, or a transverse load; the strength of timber in various forms; the strength of fly wheels, etc., etc., including a list far too long to make any attempt at enumerating them. The steam engine furnishes an enormous series of questions which need investigation, besides a host of questions in regard to the behavior of steam under different circumstances. Then, besides problems involving the strength of materials and steam, there are a great many other matters that need to be investigated experimentally: as, rope transmission; belt transmission; efficiency of different machines, as of pulley blocks, of jackscrews, etc., etc. Such problems as the above arise in general engineering work, but in mining engineering, in electrical engineering, and in chemical engineering, etc., we have a host of other questions demanding investigation by means of experiment.

Let us consider next theses of the second class, viz: Designs. If it is proposed that, for thesis work, the student shall make a design, whether it ought to be allowed or not depends upon what kind of a design it is. If it is one which does not demand investigation on the part of the student, but merely requires him to follow precedent, and to go through the mechanical work of making certain usual cut-and-dried calculations, besides making certain drawings, then the subject is not one that will develop the powers of the

student for investigation, and it ought not to be allowed. If, on the other hand, the design is so extensive that the student cannot work out all the details, then again the design ought not to be allowed. His work would not be thorough, as he would attend to the generalities and neglect the details. Thoroughness is one of the most important principles underlying any original investigation, and this should be impressed on the mind of the student.

On the other hand, designs should be allowed which do develop the student's powers of investigation, where he is called upon to make a careful investigation to determine the conditions which his design must fulfill, and to make further investigations to design it so that it shall fulfill all these conditions; where he calculates the dimensions of all the parts so that they shall have the requisite strength and stiffness, and does not rest content to determine them by some empirical rule or with no better reason than custom.

Often a thesis involves both experimental work and design, as for instance, when an experimental investigation requires a considerable amount of design to devise the machinery with which to conduct the experiments, or when, in order to make the design, it becomes necessary to perform certain experiments in order to obtain data from which to design the whole or certain portions.

There are other investigations suitable for theses which do not involve experimental work, and cannot be classed as designs, but no attempt to enumerate or classify them will be made here. In the opinion of

the writer, the principles here laid down will apply to all, viz: that only such subjects should be allowed as will serve a purpose in the professional instruction of the student, and that, if the thesis is to accomplish this object in the proper manner, it should involve original investigation. It should also be noted that the introduction of laboratories into engineering schools has very much facilitated an adherence to these principles.

Now let us consider a little more in detail the process by which the thesis is evolved. First

comes the choice of a subject. In some schools subjects are assigned, and in others the student is expected to take the initiative in proposing one. The writer prefers the latter method. Now, suppose the student is informed that he must write a thesis, that this thesis should involve investigation, and that a subject must be determined upon by a certain time, which shall meet the approval of the professor. With all such general informatien as the teacher can give, the student has at best a rather vague idea of what a thesis should be, and when he proposes his first list of subjects, it is an even chance whether there will be one of them suitable to adopt or whether all will have to be rejected as unsuitable, and some suggestions made of other subjects from which to chose. When the subject has once been chosen the teacher has to give the student some hints and directions, as to how to find the literature of his subject, and also as to what are the really most important questions for him to investigate at first in regard to it. Then the teacher