Significant figures comprises of digits, which are known reliably and one last digit in the sequence, which is not known reliably. We take an example of the measurement of length by a vernier scale. The measurement of a piece of rod is reported as 5.37 cm. This value comprises of three digits, “5”, “3” and “7”. All three digits are significant as the same are measured by the instrument. The value indicates, however, that last digit is “uncertain”. We know that least count of vernier scale is
10

4
10

4
m i.e
10

2
10

2
cm i.e 0.01 cm. There is a possibility of error, which is equal to half the least count i.e. 0.005 cm. The reported value may, therefore, lie between 5.365 cm and 5.375 cm.
In the figure, elements of measurement of a vernier scale are shown. The reading on the main scale (upper scale in the figure) is taken for the zero of vernier scale. The same is shown with an arrow on the left. This reading is "5.3". We can see that zero is between "5.3" and "5.4". In order to read the value between this interval, we look for the division of vernier scale, which exactly matches with the division mark on the main scale. Since 10 divisions of vernier scale is equal to 9 divisions on main scale, it ensures that one pair of marks will match. In this case, the seventh (7) reading on the vernier scale is the best match. Hence, the final reading is "5.37".
In this example, the last measurement constitutes the suspect reading. On repeated attempts, we may measure different values like "5.35" or "5.38".
Rules to identify significant figures
There are certain rules to identify significant figures in the reported value :
Rule 1 : In order to formulate this rule, we consider the value of measured length as "5.02 cm". Can we drop any of the nonzero digits? No. This will change the magnitude of length. The rule number 1 : All nonzero digits are significant figures.
Rule 2 : Now, can we drop “0” lying in between non zeros “5” and “2” in the value considered above? Dropping “0” will change the value as measured. Hence, we can not drop "0". Does the decimal matter? No. Here, “0” and “decimal" both fall between nonzeros. It does not change the fact that "0" is part of the reported magnitude of the quantity. The rule number 2 : All zeros between any two nonzeros are significant, irrespective of the placement of decimal point.
Rule 3 : Let us, now, express the given value in micrometer. The value would be 0.000502 micrometer. Should expressing a value in different unit change significant figures. Changing significant figures will amount to changing precision and changing list count of the measuring instrument. We can not change least count of an instrument – a physical reality  by mathematical manipulation. Therefore, rule number 3 : if the value is less than 1, then zeros between decimal point and first nonzero digit are not significant.
Rule 4 : We shall change the example value again to illustrate other rule for identifying significant figures. Let the length measured be 12.3 m. It is equal to 123 decimeter or 1230 cm or 12300 millimeter. Look closely. We have introduced one zero, while expressing the value in centimeter and two zeros, while expressing the value in millimeter. If we consider the trailing zero as significant, then it will again amount to changing precision, which is not possible. The value of 1230 cm, therefore has only three significant figures as originally measured. Therefore, Rule number 4 : The trailing zeros in a nondecimal number are not treated as significant numbers.
Rule 5 : We shall again change the example value to illustrate yet another characteristic of significant number. Let the measurement be exactly 50 cm. We need to distinguish this trailing “0”, which is the result of measurement  from the “0” in earlier case, which was introduced as a result of unit conversion. We need to have a mechanism to distinguish between two types of trailing zeros. Therefore, this rule and the one earlier i.e. 4 are rather a convention  not rules. Trailing zeros appearing due to measurement are reported with decimal point and treated as significant numbers. The rule number 5 is : The trailing zeros in a decimal number are significant.
The question, now, is how to write a measurement of 50 cm in accordance with rule 5, so that it has decimal point to indicate that zeros are significant. We make use of scientific notation, which expresses a value in the powers of 10. Hence, we write different experiment values as given here,
50
c
m
=
5.0
X
10
1
c
m
50
c
m
=
5.0
X
10
1
c
m
This representation shows that the value has two significant figures. Similarly, consider measurements of 500 cm and 3240 cm as measured by an instrument. Our representation is required to reflect that these values have "3" and "4" significant figures respectively. We do this by representing them in scientific notation as :
500
c
m
=
5.00
X
10
2
c
m
500
c
m
=
5.00
X
10
2
c
m
3240
c
m
=
3.240
X
10
2
c
m
3240
c
m
=
3.240
X
10
2
c
m
In this manner, we maintain the number of significant numbers, in case measurement value involves trailing zeros.
From the discussion above, we observe following important aspects of significant figures :
 Changing units do not change significant figures.
 Representation of a value in scientific form, having power of 10, does not change significant figures of the value.
 We should not append zeros unnecessarily as the same would destroy the meaning of the value with respect to error involved in the measurement.