「圖經」係方志的前身，隋唐時代最為發達。本文首先輯佚五代及其以前之「圖經」佚文，輯得有名稱的《圖經》88種、佚文360條，沒有具體名稱的「圖經」佚文186條，所得甚豐，故將這些佚文分成甲、乙、丙、丁四類，佚文依唐十道州序排列。 在此基礎上，作者提出三論：一論「圖經」佚文殘存狀況及其價值，認為「圖經」爲以後總志與方志的發展打下了紮實的基礎；二論現有輯佚成果，認爲前賢做得不夠，王謨、陳運溶所輯殊少，張國淦僅輯目錄，劉緯毅舛誤太甚，《唐代文化》於「圖經」書名不尊重原文；三論「圖經」成書年代，從政區沿革的角度看該問題，多有發明。 關於「圖經」進行如此竭澤而漁式的系統輯佚與研究，尚屬首次，表達了作者的一家之言，體現了作者的輯佚觀。

"Tu-jing" was the former generation of local records. It developed enough during Sui and Tang period. First of all, this author documents the lost before 960 A.D. and he finds many lost sentences. There are 88 "Tu-jing" and 360 lost sentences which had their own names in light of this research. There are 186 lost sentences without names as well. The author divides these lost sentences into four parts, i.e. first Jia is that Sui and before Sui, second Yi is for Tang and Five Dynasties, third Bing is for no era, and the forth Ding is lost sentences without names.
The author discusses three questions. One is about remnant situation of these lost sentences and its academic value. Second is commentary about academic productions in existence. Third is to reason the time of a certain "Tu-jing" . In his opinion, "Tu-jing" had a well-knit base for local records in Chinese history. But there were many blemishes of modern academic productions. Thus he had put forward his contribution about the time of a certain "Tu-jing" through the change and maintenance of ancient administrative division.
This is the first system research out and out about "Tu-jing" before Five Dynasties.