It means, in one generation, the expression Δ

*q* takes the form :

Δ

*q = m* (

*q*_{m} - q_{0})

After

*'n'* generations, the frequency of the gene becomes

*q*_{n} and then we have :

As this equation depends upon five factors.

*viz., q*_{n}, q_{0}, q_{m}, m and

*n,* knowledge of any four will lead to determination of fifth factor.

**Random drift**
As earlier discussed, while the factors,

**mutation, selection** and

**migration,** are directional in influencing the gene frequencies,

**random drift** is a non-directional factor. In this connection when we consider an actual situation, we have to realize that we always take into account only a sample of a population. This sample of a population, although representative of the population, may not have exactly the same gene frequency as found in the population. Such deviations in gene frequencies will be due to sampling errors. The extent of this sampling error will also depend upon the size of sample. For instance, if only a few parents are chosen to begin a new population, the gene frequency of this new population may deviate widely from the gene frequencies of the original population, because in this case' the sampling error will be high. On the other hand, if the sample size is large, i.e., if a large number of parents are taken to begin a new population, deviation in the gene frequencies may not be so large. In either case, however, the deviation in the gene frequency can be measured with the help of standard deviation σ

*=* √

*pq/N,* where

*p* and

*q* are the original frequencies of the alleles and

*N* is the number of genes sampled. The number of genes (population size) sampled will be equal to double the number of parents or individuals sampled from original population in case of a diploid parent, because in a diploid there would be two homologous chromosomes of each type, each carrying one gene at a particular locus. Therefore, the above formula σ

*=* √

*pq/N* will take the form σ

*=* √

*pq/2N*, where

*N* is the number of actual parents. The readers should remember that while in the former formula (σ = √

*pq/N*)

*, N* is the number of genes sampled, in the latter formula (σ

*=* √

*pq*/2

*N*)

*, N* is the number of parents.

To illustrate the above let us take an example when

*p* =

*q* = 0.5 and the population is continued using 1000 individuals. Then using the formula σ

*=* √

*pq/N,* σwill be equal to √(0.5 x 0.5)/2000 = √0.25/2000 = 0.011. this would mean that the values of gene frequency of the new population will fluctuate around 0.5 ± 0.011, which means that 68% of the population derived in this manner will have gene frequencies of the two alleles ranging between 0.511 to 0.489. Contrary to this, if we consider a hypothetical case where a new population is initiated by taking two individuals then σ

= √(0.5 x 0.5)/4 = √0.0625, which means that the values will fluctuate around 0.5 ± 0.25 i.e. -0.25 to 0.75. In actual practice, therefore, the gene frequencies due to random drift may approach the two limits that is 0 and 1. This would be possible only when new population arises due to a very small sample leading to the fixation of one allele at the cost of the other. In this manner, the changes in the gene frequencies can be brought about without the existence of any directional force i.e. mutation, selection and migration, and this change in gene frequency has been called random genetic drift.