- Focus on the image(s), not the texts and watermarks (if any). If it is a screen capture from a television/TV show, movie, game, etc., then make your own caption. You don't have to use its theme, but it helps.
- Use SayWhat to add dialog bubbles and texts to http://image.shutterstock.com/display_pic_with_logo/140095/140095,1291126283,1/stock-photo-statue-of-labour-ants-civilization-living-million-years-because-of-working-66243364.jpg and http://images.fineartamerica.com/images-medium/ant-kevin-middleton.jpg ...
- You can do silly doctoring and captioning like I CAN HAS CHEEZ BURGER and others for it/them. Use roflbot is recommended for this case.
- Ignore the technical errors like missing legs, antennae, etc. if any.
- If there is more than one images, you do not have to caption/doctor all of them. Do one, do each one separately, or all together for captioning or doctoring.
Your overlord's cheesy caption: It is now Spring in the northern hemisphere and ants are finally waking up, but they need to exercise first to start their work.