Overflow and underflow of values of various data types is a very common occurence in Java programs. This is usually because the beginners dont' pay proper attention to the default values of various data types. If we are creating a byte type variable and assigning it a value, we should be aware that the value will be treated as an int and hence a potential overflow condition.
In Java the overflow and underflow are more serious because there is no warning or exception raised by the JVM when such a condition occurs. Some developers argue that the program should either crash or raise exception in such case but the decision for adding such behavior is in the hands of creators of programming language. By looking at a problem in your program, you can't straightway tell that an overflow or underflow condition has occured. It is only after debugging that we come to know of the real cause.

Overflow in int

As int data type is 32 bit in Java, any value that surpasses 32 bits gets rolled over. In numerical terms, it means that after incrementing 1 on Integer.MAX_VALUE (2147483647), the returned value will be -2147483648. In fact you don't need to remember these values and the constants Integer.MIN_VALUE and Integer.MAX_VALUE can be used.

Underflow of int

Underflow is the opposite of overflow. While we reach the upper limit in case of overflow, we reach the lower limit in case of underflow. Thus after decrementing 1 from Integer.MIN_VALUE, we reach Integer.MAX_VALUE. Here we have rolled over from the lowest value of int to the maximum value.

For non-integer based data types, the overflow and underflow result in INFINITY and ZERO values. You may try the following lines to verify this:

float f = 3.4028235E38f * 20f;
System.out.println(f);

Note: As with int data type, we have wrappers for all primitive data types. So we can easily see the upper and lower limit of each data type by looking at the MAX_VALUE and MIN_VALUE constants in these wrapper classes.