There are numerous evidences in the early ages history, which show the use of test-based intellectual evaluation. This happened because of the practical interest. For example it is important for employers to predict the worker's ability to perform a particular kind of job or for army to test recruits in order to assign them to suitable tasks. For example, Wainer  reports the tests usage in China around 2200 B.C. These tests were organized according to a principle that a small set of personal skills measurements could tell how the person was going to perform when hired to work. However despite these facts are known, the ancient examples can't be considered too much scientific, because there's no evidence that any kind of a serious reliability or validity evaluation had ever been performed before the dawn of the 20th century.
The first widely accepted scientific intelligence assessment method was developed by Alfred Binet (Figure 6) in 1905 in cooperation with Theodore Simon. It was designed to identify the subnormal children, who could potentially experience troubles in normal schools due to delays in mental development. The battery was constructed of about 30 tasks to model the performance in real life, e.g. naming objects on a picture, memorizing rows of numbers, counting coins, etc. As the researchers were checking this test on children, they noticed that the performances of subnormal children could be compared with the performance for much younger normal children, which allowed them to propose, so called, "mental age" scale.
The mental age scale was worked out by administering test tasks to a large selection of children and assigning the particular biological age to the amount of tasks, which 75% of children of this age could solve successfully. For example, if 75% of five years old children succeeded with a certain amount of tasks, this amount characterizes mental age of a five years old. When a child passes a test, s/he would get a score of some mental age according to the amount of correctly solved tasks. It was empirically stated, that a child, having a score of two years below the biological age is very likely to have problems with normal school education. The results of this test proved appeared in a good correlation with other types of assessments, e.g. teacher's judgments.
One of the main directions of the intelligence testing evolution was driven by the necessity to generalize Binet's system to have a measurement tool for other age groups as well and to simplify its practical usage. Wilhelm Sterm (1912) introduced the term of IQ (Intellectual Quotient) — a relation between the mental and biological age taken in per cents. This gave a possibility to compare mental abilities between different people and made it less dependent on absolute values. Lewis Terman, (1916) who worked at Stanford University, adapted the test for American school children and expanded it with the use of additional tasks for assessing the "superior adults", which was called the Sanford-Binet edition. The World War I has boosted the need for intelligence assessment because of the military demands. Two sets of tests called Army Alpha and Army Beta were developed (1917) in USA with the main objective to simplify the large scale testing of recruits (mainly to decrease the testing time and the involvement of specially trained psychologists for test administration).
Ivan Tugoy, 2003
And if you are able to read in Russian, please also check www.effecton.ru