Floral color variation across life history and geography in Mimulus ringens (Phrymaceae)

With these data, we explore the possibility that both floral color and life history have shifted together in a recently described, genetically distinct group within the species Mimulus ringens. Using a large, range-wide citizen science dataset, we test for geographic trends in flower color and flowering time. We combine this with greenhouse studies in populations of known life history to test for differences in flower color with life history. We show that darker-flowered plants are more common at higher latitudes, that annual-like populations have darker flowers, and that flowering time varies with latitude only in the subset of populations that have lighter flowers. This suggests that annual-like populations (with the earlier flowering time typical of this life history) are restricted to the northern part of the species range, and likely arose there. 

Methods
Greenhouse study 
            We assessed flower color in two ways: anthocyanin extractions and photographic analyses. We obtained 22 individual plants from six populations from the Sobel Laboratory of SUNY Binghamton University. Three populations had been previously identified as perennials and three populations as annual-like. Flowers were collected from September to November 2021.
            Plants were vernalized for six weeks at 4ºC and moved on August 9 2021 to the greenhouse in the Loyola Science Center at the University of Scranton, Scranton, Pennsylvania, USA, where they were kept in 16 hour days (28ºC) and 8 hour nights (18ºC) until flowering was complete. Water levels were monitored daily and plants were bottom-watered when their trays appeared dry. If an individual’s leaves turned brown, the plant was immediately repotted to regulate soil pH. Plants were sprayed with Neem Oil when spider mites appeared.
            Flowers were sampled every other day throughout the flowering season using the freshest fully-opened flowers from each plant. During collection, we prioritized an even sampling and attempted to collect at least one fresh flower from each plant each sampling day. However, as plants flowered at different times, we collected fewer flowers from some plants. In addition, on days when only a few plants produced fresh flowers, we would collect up to four flowers from those plants on an individual day. A flower was removed with forceps and placed in a test tube holder. A unique label was created for each flower, which included plant identification number, sample number from that plant, collector’s initials, and date. 
            We then photographed each flower from the side and the front using a Canon EOS Rebel XSi digital camera fitted with a macro lens, flash, and flash diffuser (attached to the manual focus) and mounted at 46.8 cm above the base of a copy stand. We set the camera to an AV setting with the white balance set to flash, F-stop of 10, ISO of 100, and picture style of “faithful.” Each photograph contained the flower of interest, the entirety of the unique label, and a corner of an X-Rite Classic Mini color checker (the white box, the first 3 grayscale boxes, and the first ~10mm of the scale). We used a Canon RS-60ES Remote Switch to take each photograph.
            After photographing each flower, we used a standard hole punch to remove a circle of tissue from the right lower lip of each flower with the standard corolla pigment (without any throat color). We extracted the anthocyanins in 1mL of methanol/1%HCl, and stored it for 12-24 hours. After centrifuging each sample, we removed 500mL of supernatant to evaluate absorbance. We used a Genesys 20 Thermo Scientific spectrophotometer to analyze absorbance at 520nm. Higher absorbance values indicate darker flowers, with more anthocyanin.
            To analyze the color from each photograph, we used the program FIJI, an ImageJ application. To standardize each image, we used the third grayscale square from the right, white side of the color checker. We used the first analyzed image as our standard. We loaded the image, then selected image > adjust > color balance > all colors. Using the “auto” balance function, we balanced the image’s color, saved the image as a new file, and used analyze > measure to assess the mean gray value of the square for standardizing. As the mean gray value was 183.493, we then standardized each image so that the third grayscale square had a mean gray value between 183 and 184. 
            After saving each standardized image as a new file, we created a region of interest (ROI) that encompassed the entirety of the flower petals in the image. We then performed analyze > color histogram to obtain red, green, and blue color values. We then split the image into stacks of hue, saturation, and brightness. For each stack, we obtained the mean gray value.

基于本数据集，我们探讨了新近描述的、遗传上独特的条纹沟酸浆（Mimulus ringens）类群中，花色与生活史协同演化的可能性。本研究借助覆盖整个物种分布区的大型公民科学数据集，检验了花色与开花时间的地理分布格局；结合已知生活史的种群开展温室栽培实验，进一步验证不同生活史类群间的花色差异。
研究结果表明：花色较深的植株在高纬度地区更为常见；类一年生种群的花色更深；且仅在花色较浅的种群子集内，开花时间才随纬度发生变化。这提示类一年生种群（具备该生活史典型的早开花特性）仅分布于该物种分布区的北部，且极有可能起源于此。

## 材料与方法
### 温室栽培实验
我们通过两种方式评估花色：花青素提取法与图像分析法。本研究从纽约州立大学宾汉姆顿分校索贝尔实验室的6个种群中获取了22株个体，其中3个种群此前被鉴定为多年生种群，另外3个为类一年生种群。花材采集于2021年9月至11月期间。

供试植株于4℃下春化6周，2021年8月9日被转移至美国宾夕法尼亚州斯克兰顿大学洛约拉科学中心的温室。温室环境设置为16小时光照（28℃）、8小时黑暗（18℃），直至植株完全开花。每日监测基质湿度，当托盘内基质干燥时采用盆底灌溉方式补水。若单株叶片出现枯黄，则立即换盆以调节土壤pH。当出现红蜘蛛时，对植株喷施印楝油（Neem Oil）。

在整个开花季内，每隔一日进行一次花材采样，采集每株植株上最新完全开放的花朵。采样过程中优先保证采样均匀，计划每日从每株植株上采集至少1朵新鲜花材。但由于不同植株的开花时间存在差异，部分植株的采集花材数量较少。此外，在仅有少量植株开放新鲜花朵的采样日，单株单日最多可采集4朵花。使用镊子取下花朵，放置于试管架上。为每朵花制作唯一标识标签，标签内容包含植株编号、该植株的采样序号、采集者姓名缩写以及采样日期。

随后使用搭载微距镜头、闪光灯及闪光灯柔光罩（安装于手动对焦环上）的佳能EOS Rebel XSi数码相机，从侧面与正面拍摄每朵花。相机安装于翻拍架底座上方46.8cm处。拍摄参数设置为AV档，白平衡设为闪光灯模式，光圈F10，ISO感光度100，照片风格设为“忠实”模式。每张照片需包含目标花朵、完整的唯一标识标签以及爱色丽（X-Rite）Classic Mini色卡的一角（包含白色色块、前3个灰度色块以及标尺的前约10mm区域）。使用佳能RS-60ES遥控器进行拍摄。

完成花朵拍摄后，使用标准打孔器从每朵花的右侧下唇（仅包含标准花冠色素区域，不包含喉部颜色）取下圆形花瓣组织。将组织置于1mL甲醇/1%HCl溶液中提取花青素，静置储存12-24小时。对每份样品离心后，取500mL上清液用于吸光度检测。使用赛默飞世尔科技Genesys 20分光光度计测定520nm处的吸光度。吸光度值越高，代表花朵颜色越深，花青素含量越高。

为分析每张照片的颜色信息，我们使用FIJI软件（ImageJ的衍生应用程序）。以色卡右侧的第三个灰度色块作为图像标准化的参照。以第一张分析的图像作为基准标准：导入图像后，依次选择「图像>调整>色彩平衡>所有颜色」，使用“自动平衡”功能校正图像色彩，将校正后的图像另存为新文件，并通过「分析>测量」工具计算该灰度色块的平均灰度值以用于标准化。由于基准图像的平均灰度值为183.493，因此将所有待标准化图像的第三个灰度色块的平均灰度值调整至183至184之间。

将每张标准化后的图像另存为新文件后，在图像中绘制覆盖所有花瓣区域的感兴趣区域（ROI，Region of Interest）。随后执行「分析>颜色直方图」操作，获取红、绿、蓝三色的颜色数值。将图像拆分为色相、饱和度及明度三个通道栈，分别计算每个通道栈的平均灰度值。