FIRE INSURANCE. 643 FIRE INSURANCE. from damage by fire or by the water used in extinguishing lircs. Insurance companies in the United Status have established organizations with duties similar to those of the English sal- vage corps. They are variously known as fire patrols, salvage corps, and protective associa- tions. The earliest American association of that order was established in Now York in L839. Similar organizations are supported to-day by underwriters' associations in at least thirty of the larger American cit ies. Of more importance than (bis direct prevent- ive activity of insurance companies is the in- direct influence which they exert through the pressure which they bring to bear on the insured. This pressure is exerted in sonic cases through the refusal of the company to accept the risk unless certain changes arc made in the property which will lessen the danger of fire; more com- monly through inserting in the policy that cer- tain specified acts of the insured, of such a na- ture as to increase the risk, shall render the policy void, and almost universally by varying the premium rate in accordance with the number of preventive devices adopted by the insured. The introduction of automatic sprinklers into general use was directly due to such discrimina- tion by insurance companies. Fire Marshals. Finally, it is largely due to the agitation carried on by insurance companies that States and municipalities have adopted vari- ous measures for the purpose of preventing loss by fire. The more common forms of public action for this purpose are the maintenance of fire- extinguishing organizations, and the adoption of regulations concerning the storage of explosives and other dangerous commodities, and regula- tions as to the character of material to be" used in the construction of buildings. A more recent movement in the same direction is the adoption of laws for the appointment of State fire mar- shals. Local fire marshals or other similar offi- cers have long been maintained in the larger cities. The first State fire marshals were au- thorized by Massachusetts and Maryland in 1894. Three other States have followed their example — Ohio (1900), Connecticut (1901), and Washing- ton (1901). In 1902 Massachusetts transferred the duties of the fire marshal to the district police. In Mississippi the insurance commission- er acts as fire marshal. Along with the estab- lishment of the office of fire marshal has gone the adoption of laws requiring investigation into the origin of fires. Maine and Pennsylvania, as well as the States mentioned above, require such official investigation. Much light will undoubt- edly be thrown upon the causes of fires by these investigations, and more intelligent action will be possible for the purpose of diminishing their number and destruetiveness. The duty of State and municipality to adopt regulations for the purpose of minimizing fire losses is too obvious to need discussion. When the carelessness or misconduct of one individual may result in irreparable losses to others who have themselves taken all due precautions, the State is clearly justified in doing all in its power to ward off the danger. What municipalities have done is but a small part of what they might properly do, and what they may be expected to do as more enlightened views of their responsi- bility in the matter come to prevail. Technique. The Risk. The proper business Of insurance companies is not the prevention of loss by fire, but the assumption of risk. Througn an insurance policy the risk to which the prop- erty of the insured is exposed i- assumed by the insuring company. Insurance in it lell may have no effect on the amount of property de by lire, or upon the chance oi destruction to which a particular piece of property is exposed. The gain conferred by insurance upon society arises partly from the' reduction in the uncer- tainly as to how much loss will be suffered in a given period through the accumulation of a large number of risks, and partly f'r a distribution of the lo~-e^ actually suffered i ag i large mini ber of individuals. The risk which an insurance company assumes when it insures a [lieee of prop- erty against loss by fire depends upon several factors; viz. the value of the propertj insured, the probability of fire within a given time, (he probable destruetiveness of the fire if it occurs, and the -length of time for which the insurance i^ granted. The first and last of these factors are determined with comparative ea-c and ac curacy. The determination of the second and third, on the other hand, presents very great difficulty. cSo far as human intelligence is con- cerned, it is largely a matter of chance whether a particular building is visited by fire in a given year or not. The adoption of preventive meas- ures may reduce the probability, but some degree of uncertainty will always remain. And when there is added to that uncertainty the uncertainty as to the extent of the loss occasioned by the fire, it will be seen that there is a large clement of chance in the liability which the company as- sumes so far as any particular property is con- cerned. Moreover, the uncertainty is of such a nature that the chance of loss cannot be deter- mined by the most careful examination of the property. The knowledge of the conditions affect- ing the possibility of fire is too imperfect to admit of such direct determination. It is neces- sary, therefore, to rely upon the statistics of past experience to show what chance of loss a particu- lar risk brings upon the company. If statistics show that each year for a number of years fire has destroyed, upon the average, one out of every thousand buildings of a certain kind, there is one chance in a thousand that any particular building of the same character will be destroyed during the following year. The greater the num- ber of cases from which the average is calculated, the more reliable is the average — that is, the less danger is there that it has been unduly affected by temporary fluctuations in the number of tires. The experience of insurance companies them- selves furnishes the most valuable data obtain- able for calculaf tag average losses. It is ex- tremely desirable that the movement already under way to combine the experience of many companies, and thus calculate combined experi- ence tables for fire insurance, as they have al- ready been calculated for life insurance, should be carried out. But even with the help of experience, the prob- lem of calculating the risk incurred by insuring a particular building is by no means simple. In the firs! place, the probable destruetiveness of the fire must be determined from experience, as well as the probability of the occurrence of the fire. In the second place, the problem is made more complex from the fact that no two pieces of property present precisely the same fire fea-
