When copying message content from an external source (such as a Word document), it is possible that your message can contain UTF or UCS characters without being obvious. An example of this is when paragraph formatting is used such as a "Tab" or "whitespace" characters. To combat this, the web portal will replace the problem characters with their equivalent GSM character, ensuring the message goes out as GSM and thus not inadvertently trigger a 70 character limit on the whole message.
PLEASE NOTE: It is not always possible to map UTF characters to a plain text equivalent, so you may wish to check the list of convertible characters before constructing your messages.
A full table of replacements is attached to this article