Some clarification here up front, as it seems you're not overly familiar with EC2 and how it all ties together.
An EC2 instance is a virtual machine running the operating system of your choice, and with a fixed CPU and RAM allocation. The amount of each is dictated by the instance type (t1.micro, m1.small, c1.medium etc.) See here for more information on instance types and specifications:
http://aws.amazon.com/ec2/instance-types/
A reserved instance is a billing construct that is applied to your AWS bill. The way it works is that you pay a fixed sum up front based on a 1 or 3 year period, and this entitles you to a lower per-hour cost for your EC2 instance.
An EC2 instance type can be changed. You may stop the instance, change it (for example) from a t1.micro to a m1.large, and start it again, which upgrades the hardware available to the instance. If you wuld like a fixed IP address, you can associate an elastic IP address at no extra charge. Note that if you do NOT associate an elastic IP explicitly, your IP address will change on reboot and you will not be able to recover the address you were previously using.
If you purchase a reserved instance, to reduce the ongoing per-hour cost of running your EC2 instance and you later change the type of your instance, it is now possible to also adjust the type of your reserved instance. However, this is subject to some conditions, and AWS will not refund you any portion of the reserved instance price. See the following link for more information on how this works:
http://docs.aws.amazon.com/AWSEC2/latest/UserGuide/ri-modifying.html
Merging multiple EC2 instances is not a feature of AWS or (as far as I know) any other existing cloud provider. If you want to spin up multiple instances and cluster them, you will need to know how to do so using Red Hat Cluster Suite, or the equivalent for your platform.
I hope this helps!