Middle school teachers tend to rely heavily on texts that have become increasing more visual. There is little information available about the graphical demands of general middle grades' science texts. The purpose of this study was to quantify the type and quality of the graphical representations and how they interacted with the textual material in middle school science texts. Four State of Texas-approved sixth-grade science texts were examined using a researcher-developed instrument, Graphical Analysis Protocol (GAP). Results produced reasons for concern (least analytic forms of graphics dominated, one third of all graphics were decorative in nature, most graphics were static in nature, and approximately one third were not connected to the text spatially or semantically) and reasons for optimism (approximately one third of the time more analytic forms of graphics were used, two thirds served a specific cognitive function, and over half of the graphics were connected to the text and indexically referenced). Further research is warranted on a larger sample of texts from different states, publishers, grade levels, and subjects. It is also important to establish a connection to increased student learning and/or teacher preference due to higher incidences of any of the variables or principles that make up the GAP instrument. (Contains 6 tables.)