time keep in touch with different branches of their subject and avoid the danger of overspecialization too early. A laboratory should therefore be organized in departments with an intradepartmental arrangement under which a young man who develops the ability to carry out his own work may be able take up work on his own initiative, still retaining his position in the department and carrying on his work under the general supervision of the chief of his department. There will always be a tendency in the departmental organization for men to desire to split away from the department to which they are attached and become semi-independent in the laboratory, and this tendency must be resisted in the organization and by the director of the laboratory. At the same time, it is important that too rigid a control should not be exercised so that men feel that they are prevented from exercising their own initiative.

A laboratory for a specific industry will generally tend to be of what has been called the "convergent" type, that is, one in which all the different sections of the laboratory representing different branches of scientific work have their energies directed towards the solution of problems relating to the same subject. The problems of such a laboratory will, therefore, all be inter-related and the work of the laboratory will be directed towards one common end.

The organization of such a convergent laboratory has been discussed in a former paper. It is shown there that charts could be prepared illustrating the organization which would be available for almost any convergent laboratory, so that, if we have to work out the organization of a research laboratory which is to study any inter-related group of problems, we can do it by the construction of similar charts. Thus, we may arrange a chart showing the derivation of the branches of the subject considered from the sections of pure science involved. We can place on one side biological, physical and chemical problems, subdividing each section so that each one represents work capable of being handled by one man in the laboratory. It will now be possible to draw a new chart, showing on the circumference the different sections of the laboratory for which accommodation, apparatus and men must be provided, and showing the relation of these sections to the problem as a whole. Having worked this out, it is easy to find the amount of space and the number of men which will be required, or which the funds available will allow for each part of the work.

Now, before applying these charts for laboratory organization to a specific industry, let us look at the question of the 2" The Production of Scientific Knowledge," Science, 1917, p. 519.

physical organization of the laboratory itself: the building and scientific equipment, the cost of building, and the cost of the

maintenance and operation. It may be mentioned here that when a laboratory is under consideration by the executive of a company, the matter which usually concerns his mind are these physical details, and he is often greatly concerned with the planning and cost of the building and equipment, a matter which, as will be shown later, is quite secondary to the internal organization of the laboratory as regards effect on the work or even from a financial point of view.

The laboratory should be housed in a convenient, special building. It is very advisable that all research work under the same general direction should be conducted under the same roof, since only in this way can good cooperation between the departments be obtained, and the facilities and organization of the whole department be available to all the workers. In technical research, where it is often necessary to install model plants on a small scale, this cannot always be carried out, but, as far as possible, a research laboratory should be a real building and not merely the name for a number of scattered departments at some distance from each other.

It is a mistake for a factory to house a research laboratory in some abandoned building designed for other purposes. The annual cost of research work, as will be shown later, is very high in comparison with the cost of the building itself. The greater part of that expenditure is on the salaries of the men carrying out the work, and any inconveniences or disadvantages

which may be caused by their working conditions and surroundings can easily depress the production to an extent which renders such economics very unprofitable. The cost of the research man, in fact, is so high that it is worth while to provide him with the very best facilities for carrying out his work, since, provided money is not actually wasted on useless ornaments, these facilities will always be inexpensive in comparison with the total expenditure on the work.

Research laboratories are almost always too small, and it is really desirable that, in designing such a laboratory, some system of construction should be chosen in which expansion can be obtained by the duplication of units. This is, of course, a very difficult thing to arrange, especially in the details of the laboratory, but, nevertheless, it should certainly be aimed at by the architect, since whatever the size of the laboratory when it is designed, it is safe to prophesy that within a very few years expansion will be necessary, and if direct expansion is not possible, this will take the form of detached groups of men working in other places, an inconvenient and uneconomical arrangement.

The cost of moving in research work is not always realized. The cost of moving into a new building will be approximately half the total cost of the building, since the men will actually not be working again at full speed in less than six months, and, as a general rule, the annual expenditure is equal to the cost of the building and equipment. It is important, therefore, in designing a laboratory to arrange, if possible, that expansion may take place without any considerable rearrangement. An aid to this is to make the internal divisions of a laboratory movable as far as is possible, and while the laboratory itself should be of fireproof construction, it will be convenient to make partitions of composition board and wood wherever the fire risk does not prohibit this. In this way, rooms can easily be subdivided, combined or rearranged.

Everything that has been said as to the necessity for the provision of a satisfactory building applies also to the question of equipment, but with even greater force. It is an economic error to allow expensive men to be short of the apparatus which they require for their work. As a general rule, men will not ask for apparatus which they do not need. There are a very few men who might be considered to be apparatus collectors, and who seem to have a real anxiety to surround themselves. with all forms of scientific apparatus, whether they have any use for them or not; but with the exception of these men, who are limited in number, it may be taken that when a research

worker asks for apparatus he needs it, and must have it in some form or other to continue his work.

The total cost of equipment for a physical laboratory represents about two months' cost of operation, and, if economies are to be made, it is clear they should be made in limiting the amount of work undertaken and the consequent cost of operation, rather than in depriving the employed workers of the necessary tools for their work.

From various sources of published information, as well as from personal experience, it is possible to form an estimate of the cost of a research laboratory per scientific worker employed, taking the term "scientific worker" to cover all graduate men working in the laboratory. It might seem that there would be very great variation in the cost, but, provided that we confine ourselves to laboratories of the physical and chemical type, there is a surprising agreement between the different figures, which show that cost of building and equipment for a laboratory will be between $3,000 and $4,000 per man; it may be taken therefore that the first cost of a laboratory will be about $3,500 per scientific worker employed. From the same sources the annual cost of maintenance of such a research laboratory appears to be slightly lower than the first cost. Probably $3,300 per man would be a fair estimate of the cost of maintenance, and of this we may take 60 per cent. as representing salaries and wakes and the other 40 per cent. all other expenses.

Let us attempt to apply the principles which have been laid down for the design of an industrial research laboratory, applicable to a specific industry, in such a form that they would be available for the directorate of the industry to understand to what they are committing themselves in establishing a research laboratory, and how to proceed in order to do so.

We may select as an example of a specific industry one of a technical manufacturing type dealing with engineering processes, handling chemicals, and also involving certain biological considerations; such an industry, for instance, as textile dyeing or the manufacture of leather goods. Exactly the same principles, however, would apply to industries of quite a different kind. Thus, an industry in which there are no biological considerations will not require some branches of a laboratory; it may need to substitute others in their place. For some industries, physics is of no importance and chemistry is of far more importance.

Let us, however, in order to be specific, consider the question of a plant whose business consists in the dyeing of textiles. Let us suppose that the industry is making a turn-over of $1,000,000 a year, of which 10 per cent. is net profit, and that the

directors have decided that, in order to improve their product and extend their business, possibly to diminish costs, they will at the outset undertake an expenditure of $15,000 a year on scientific research. Now, let us consider what they can do for this.

In the first place, we can decide at once how many men they can get. On the basis of $3,000 per man, they should be able to get five men for $15,000, but with very few workers in the laboratory, the cost per man will be somewhat higher, and it will be safe to assume that only four men can be obtained for the $10,000 available for salaries. The cost of the building will be about $10,000 and equipment about $5,000. Taking the basis of $2 per square foot for building as a rough approximation, we shall have a building with 5,000 square feet of floor space, or, dividing this into three floors, a building about 40 feet square. The work of the laboratory may be analyzed according to the chart shown in the figure. Dividing into the three main divisions of chemistry, physics and biology, we shall get the following sections for the work: In chemistry, we shall require an analyst and dye chemist who must understand organic chemistry, and a colloid chemist who will study the relation between the fiber and the dyes. In physics, we shall have work to do on the testing of the strength of materials and especially on colorimetry and the measurement of absorption. In biology, we shall require a man who understands the vegetable and animal fibers, their structure and their bio-chemical properties. We shall also require work on the staining of fibers and photomicrography. This will give us the chart shown.

Now, we can not hope, of course, to represent all these departments by separate men, since we can afford to have only four men, and in addition to the departments shown we must have one practical dyer having actual works experience. Our men may be grouped somewhat as follows: Our organic chemist can look after analytical chemistry as well; that is, we must get a man having experience in organic chemistry and some good knowledge of dyes, who can specialize in the study of dyestuffs and on their analysis, but who also can do what routine analytical chemistry it becomes essential for the laboratory to carry out. We may expect our colloid chemist to be a bio-chemist and to take care of the microscopy. The physicist may understand colorimetry and, at the same time, know enough general physics to be able to look after questions involving the strength of materials. We have thus accounted for three of our four skilled men, and the fourth must be the practical dyer, who should also be the director of the laboratory and should have a good training in dye chemistry and general chemistry,