in the specifications are to be rigidly insisted upon, while others will not be specifically enforced, this intention or state of mind of the engineer should be revealed in the specifications themselves. In other words, the contractor should know in advance how the specifications are to be interpreted, so far as it is possible to give this information in the specifications themselves.

(e) The last named requirement demands that the engineer should be so familiar with all the details of the work described, from actual experience, that he is able to know in advance what his decision will be in the various contingencies which may arise during the progress of the work. His foresight in this particular must be complete and distinct, which can only be the case when the engineer who writes the specifications has had considerable experience in carrying out practically the same kind of work.

(f) In choosing units of measure, in describing the work, to be used for determining compensation, only specific and definite units should be chosen, or they should be so defined as to admit of no double meaning. For instance, to say that mortar shall be composed of one part cement to two parts of sand comes very far from defining a particular ratio of ingredients. If the words "by weight" be added, it still fails to define, inasmuch as wet sand is much heavier than dry; or if the words "by measure" be used, this also fails to define, since cement may be measured in the original package, where it is thoroughly compacted, or it may be dumped and measured in a fluffy condition, in which it occupies some 50 per cent. more volume.

(g) The engineer should be familiar with the ordinary. methods employed by different kinds of mechanics and should so design his work as to obtain satisfactory results without requiring a much higher grade of work than is customary by the mechanics who will be called upon to execute it. It is practically impossible with the most thorough supervision and

inspection to get mechanics to vary their ordinary practice materially. The failure to recognize this fact often leads engineers to specify methods or results which are practically unattainable, and this leads either to continual violation of the specifications and its accompanying irritations and delays, or to an abandonment on the part of the engineer of the strict interpretation of his own specifications.

(h) In the matter of materials it is customary to specify not the very highest and best the market affords, but such a grade of material as would be satisfactory in service, and which can be supplied by the standard manufacturers of that particular product. In this way the engineer gets the benefit of a wide competition, and of a correspondingly low price. The minimum requirements for materials which serves as a criterion of rejection determines very largely the cost of the work. If, therefore, the engineer in preparing his design bases his calculations upon what might be commonly known as good or firstclass materials, with a minimum limit fairly below this generally recognized first-class grade, he will usually obtain a material practically as good as the market affords, without being obliged to pay an extravagant price for it, and without suffering from the delays and troubles caused by the rejection of a large proportion of the material furnished. To base a contract on the very highest tests known of a given material, and to require this extraordinary quality for all the material furnished is extremely unwise.

(i) If possible to avoid it, it is best not to specify a particular manufactured product or proprietary article by name. If this is done at all, more than one such name should be given if possible, and others admitted if shown to be as good as these to the satisfaction of the engineer. To limit the materials or articles specified to that of a single manufacturer subjects the engineer to invidious criticism and suspicion, and it is always wise to avoid even the appearance of evil.

(j) It is not uncommon to specify that the materials furnished shall be of well known brands, or the products of factories or works of established reputation. Similarly it is sometimes specified that the contractor himself must show a familiarity with the work he proposes to perform. Also in the matter of stone, for instance, that it shall be taken from quarries which have been in long use, the stone from which by actual use is known to have good weathering qualities.

(k) Before writing the specification, the engineer must also have a clearly defined notion as to the amount of responsibility which is to be placed upon the contractor. If the engineer prescribes the plan, the materials, and the methods to be employed, he, of course, assumes all responsibility so far as these are concerned, and he can not in justice make the contractor responsible for his own faults. He should always be ready and willing to take upon himself such responsibilities pertaining to the design as properly belong to the designer. For instance, if the contractor is held responsible for the results, he must be given a considerable latitude as to methods, and if the engineer prescribes the plan, the materials, and all the methods to be employed, he can not hold the contractor for the results, or for the successful operation of the project, beyond the simple faithful performance of the work prescribed. In general the entire responsibility for the successful operation of the work should rest upon the engineer. It is only in case of extra-hazardous undertakings, which are largely of the nature of an experiment, in which no well defined plan is outlined, that the contractor is left comparatively free both as to plan and execution. It is only in such cases that the engineer is warranted in relieving himself of all responsibility, and placing the same wholly upon the contractor. There are cases in which contractors have, by experience, acquired peculiar ability to perform certain kinds of hazardous work, and who are willing to undertake the same under a guarantee of suc

cessful execution, under which circumstances, any engineer, even of high repute in his profession, would be warranted in letting a contract and putting the entire responsibility upon the

contractor.

(1) A strong reason for enforcing specifications literally and rigidly, instead of accepting some other material or method which possibly may be "just as good" is to be found in the relation the engineer and owner hold to other contractors who bid upon the work. It is to be presumed that these other parties have based their estimates on a strict compliance with the specifications, and it is possible that the lowest bidder has presumed on his being able to substitute cheaper materials or methods for those specified. If he is allowed to do this after the contract is let, it is evident that the other bidders have been discriminated against to their disadvantage and under a species of fraud, which should not receive the encouragement of either the engineer or of the owner. It is difficult, therefore, to see how a cheaper grade of work can properly be accepted in lieu of that specified, even though it be "just as good," without encouraging this practice of presuming upon a cheaper fulfillment, and also without treating the other bidders unfairly. Other things being equal, therefore, it is best to rigidly enforce a contract, even though a cheaper material or method might, in the opinion of the engineer, be employed with equally good results. Or, if a cheaper compliance is allowed, a corresponding reduction in price should be insisted upon.

The above are some of the numerous controlling ideas which the engineer should have clearly in mind in the writing of a set of engineering specifications. He must know in the first place exactly what he wants, and then try to so describe it that others can not mistake his meaning. The general and detail plans are usually made before the specifications are written, and the engineer has these before him in writing the specifications, and makes liberal reference to them. Since

they are also a part of the specifications, he has the advantage of a double language in which to present his ideas, and, if he does not succeed in making clear to the proposed contractors exactly what is to be done, he should feel that he alone is to blame for any misunderstanding.

116. Specifications Accompanying Complete Detail Plans. As described in article 17, we have in general three classes of engineering specifications, describing the work itself, namely:

Specifications accompanying complete detail plans. Specifications accompanying a general plan only. Specifications unaccompanied by any plan, and commonly known as General Specifications.

When the specifications are accompanied by complete detail plans, these plans are prepared before the specifications are drawn, and in this case the specific descriptive specifications are largely composed of descriptions of the materials to be used, the methods of manufacture and erection, and of the results to be accomplished, or of the tests to which the finished product is to be subjected. Such complete detail plans are always necessary when a particular and definite plan is to be carried out. When the work is to be let in open competition, and to the lowest bidder it is also usually best to prepare complete detail plans in order to insure that the bidders will all estimate on exactly the same thing, and further, to insure that the final product will be satisfactory. In this case the detail plans must be duplicated by some of the various methods of copying drawings, and in very large and important work frequently by photo-lithographing, and these copies of the plans submitted along with the specifications to the various parties wishing to bid upon the work. If the bidders are expected to be those residing in the immediate vicinity, it is not so necessary to duplicate these drawings, all the bidders being asked to examine the drawings in the office of the