臺灣西北海岸地帶是一個族群頗為多元之處：觀音北邊連著大園，為漳州移民優勢區；新豐南端連著竹北，是泉州優勢區，其餘則以廣東嘉應、惠州佔優勢。換言之，本區域正是臺灣諸多閩客交界地區之一，有其族群、語言研究上的重要性。本文利用日治時期族群調查資料的整理與分析，並用GIS（地理資訊系統）繪製成西北海岸地區的族群分佈圖，以探究廿世紀上半臺灣西北海岸的族群分佈情形。具體結論有二：其一，廿世紀初期本區域的主要族群為福建人及廣東人，原住的南島民族幾乎難尋其蹤。福建人方面，漳州移民集中在新屋崁頭厝庄內的舊頭家厝庄，以及觀音北部的草漯等六庄，此區與大園的漳州人連成一片，惟其範圍逐步向北退縮，與粵籍在崙坪庄、下大堀、新坡庄形成一個閩粵混居的地區；泉州移民一在新豐、竹北交界的大眉庄，與其南方的竹北泉籍聯成一片；另一在本區域西南沿海地帶，從新屋蚵殼港庄向南連接新豐沿海的後湖、坑子口、新庄子庄，也與南方竹北一帶的泉州區相連。上述地區以外為廣東優勢區，其勢力在本時期朝兩個方向拓展：一是向北滲入觀音的漳州人聚落，另一則向南朝新豐新庄子庄挺進。透過1926 年的調查可以得知，觀音、新屋以來自惠州居多，新豐則以嘉應州為主。某一程度，此結論可以解釋現代語言形成之歷史過程。其二，日治時代的族群調查統計有其侷限：1901 年的調查提供臺灣最早的族群分布面貌，但對於廣東人的統計不夠精細；1926 年的鄉貫調查適合從事宏觀研究，但統計數字僅到街庄，會使得小單位聚落的族群分佈不易彰顯；至於戶口調查及《新竹州統計書》則適合長時間的變遷研究，不過漢人族群特性則有賴1901 與1926 的調查來協助觀察。

This article describes the population distribution and the changed situations in the northwest coast of Taiwan in the half of the twentieth century, through mapping the population distribution in the northwest coast that is based on the population survey during the Japanese governance in GIS (Geographic Information System). There are two specific conclusions: First, during the early twentieth century, people of the Fujian and Guangdong were the major groups of the region. For the Fujian group, people of the Zhangzhou concentrated at the Jiou-Tou-Jia-Cuo, Kan-Toucuo village, Shin-Wu Township and Tsau-Ta with other five villages, north of Guan-Yin. The boundary of this area was moved and declined to the north of Taiwan gradually. On the other hand, people of Quanzhou concentrated at the Da-Mei village, which was the across area between Hsin-Feng and Chu-Pei. The other sub-group of the Quanzhou concentrated at the area of southwestern coast. Exception of the above areas, people of Guangdong was the majority of this area. Through the survey of 1926, it can be known that most people of the Guan-Yin and Shin-Wu were from Huizhou. However, most people of Hsin-Feng came from Jia-Ying Zhou. Secondly, there were limitations of population statistics during the period of the Japanese governance. For instance, it was inaccurate statistics on the survey in 1901. The statistics of the native place survey in 1926 was limited that only revealed the information of townships; consequently, it may inappropriately provide proper information about the characteristics of groups in the area. However, the advantage of population statistics during the period of the Japanese governance will offer benefits of long-term change of population distribution in the further studies.