In the case of a property only recently built, in which the records are complete and the engineering and construction files are available, it may not be specially difficult to determine cost, but in the case of any of the large railway systems of the United States, which are made up of the consolidation of many different roads, some of them built many years ago, some of them having gone through many changes of management, reorganization, and earlier consolidation, it is practically impossible to secure either the old financial books or the old construction records, and without these complete records it would appear to be an utter impossibility to secure the primary cost. Primary cost is but the first step. The work of building up, by securing the amount of additions and betterments that have been made from year to year, is one of appalling magnitude and of utter uncertainty and conjecture. Keeping in mind that, prior to July 1st, 1907, the railroad companies of the country did not have a system of uniform accounting, and that additions to property were charged to operating expenses to the extent of hundreds of millions of dollars; that policies were different on different roads, and under different managements of the same road, and that the accounting methods were determined by the policy of the road or management; and the further fact that the distribution of ordinary railroad accounts may be extremely reliable on one road and abounding with errors on another; it will be seen that any attempt to depend on the auditor's office for anything approximating a complete statement of cost would lead into a maze of figures which would be confusing, unreliable, and incapable of proof. Therefore, it is the writer's conclusion that, beyond such figures on recent construction, or records of cost of such special structures as are matters of particular record, it is not advisable to attempt to secure complete data as to the cost to date of a railroad. In the Michigan appraisal, original cost was secured in the case of many structures, notably the Port Huron Tunnel, and it is by no means argued that original cost should not be considered, or investigated, but it is held that such an undertaking as to secure, from the financial books of the company, an accurate or reliable statement of construction cost, plus additions and betterments, less property destroyed, of the Michigan Central Railroad, for example, would be absolutely an impossibility, particularly if the work was to be undertaken, as in the Washington appraisal, by men who were utter strangers to the property. The admission of the appraiser of Washington, that, except for a few gaps, the information was complete, is fatal, as the gaps must needs be filled by estimates, and it would appear to be better to depend on estimated figures throughout than to use what purported to be actual costs on part and estimates on the remainder.