Despite being out of sight roots are critical for the proper functioning and well-being of plants. In an effort to improve our understanding of plant root systems, researchers have devised many techniques for studying roots. Unfortunately many of these tend to be particularly destructive, having little regard for the well-being of the plant. Studies of a destructive nature tend to be non-conductive to replication, and observation of temporal changes within the root system. Unlike many techniques before it the Minirhizotron is a relatively non-destructive method for studying root systems. The ability to repetitively observe a given set of roots, provides the opportunity to study changes in root activity through time. Although the concept is not new, the equipment used is continually evolving. A preliminary investigation was conducted using a modern root camera purchased from Bartz Technology (650 Aurora Avenue, Santa Barbara, California 93109 USA, phone: 805 965-4343 fax: 805 965-4255 E-mail: barktek@coyote.rain.org). This system proved to be very effective for observing and recording root frames with high quality images being obtained.

The camera head, and the system in action

However, our initial experience has highlighted some deficiencies in the system, which has prompted further development to refine the equipment and procedures for capturing root images.

An experiment carried out in the Massey University Orchard, focused on strategies of water uptake on two apple trees. The techniques employed were TDR and heat-pulse, with tensiometers being installed to measure the water status around the trees. Our newly-acquired root camera was used in conjunction with the existing techniques to observe the roots of the two apple trees. A total of twelve Minirhizotron tubes were installed at this site, with eight tubes around the less-confined tree, and four tubes around the other tree. Installation of the Minirhizotron tubes was carried out mid-winter, using a modified tree auger to bore the holes, with a steel frame to guide the auger.

Our initial approach to using the Minirhizotron was to make root observations on a weekly basis, supplemented by more frequent observations as irrigation treatments were applied to the soil. During the period of the experiment (November-April) the trees underwent a combination of wetting regimes, the first of which was a full irrigation on both trees. After the first period of drying, the trees were then differentially irrigated, with only one half of the trees being irrigated. Sampling involved sending the Minirhizotron camera down each tube while recording the images to a camcorder. A locking mechanism was used to prevent lateral movement of the camera, while an index system on the camera arm allowed the operator to position the camera. This system provided reasonable image registration, which is essential for accurately determining temporal changes within the rootzone. The video tape could be played back through a standard TV monitor (essential equipment during Americas Cup time). The operator was required to record to a database the number of roots that intercepted each frame, and the number of root branches within each frame.

Consistent image registration proved to be of concern and was highlighted when a time series of individual locations were viewed. It was obvious from a series of images that the registration was quite variable, and in some instances complete frame mismatches occurred due to operator error. With this in mind we decided to automate the capture process, thus reducing operator fatigue, boredom and total frustration. By replacing the existing camera arm with an electrically driven system, we are now able to precisely locate the camera within the tubes.

Using a program written in Quick Basic, we can very accurately control the positioning of the camera via a laptop computer. A small electric stepper motor coupled to a gearbox, drives a chain mounted on the camera arm. The operator is required to input the depth of travel, then execute the sequence. The camera will travel from frame to frame down the tube, momentarily pausing at each frame long enough to allow the video to record a still image of individual frames.

When the camera reaches the maximum depth as input by the operator, it will return to the top of tube. While the program monitors the location of the camera head, it also controls the on screen data to the video camera. Therefore each time the camera head is moved the location is displayed to the laptop monitor and the on screen video data is also updated. This system ensures that there are no discrepancies between actual camera location and position data recorded to the video tape. With these modifications to the field equipment complete, the accuracy of the data and the ease and speed at which we can capture images has been greatly improved.

A small portion of the total number of frames may be digitised allowing us to build up a series of images over time. When displayed using a database developed in Paradox, the viewer can observe the changes that have taken place in a single year in a very short time, somewhat like time lapse photography. This is a very useful tool for demonstrating root dynamics albeit of a very small part of the entire root system. Each image that is digitised is saved with the whole number and the depth value as part of the file name. The files for each day are saved to a directory structure based on the date for each days observations.

The software will search a directory structure loading those images that match a given set of criteria, while also entering the DOY, hole depth values automatically. Having further developed the Minirhizotron system, we are confident that many of the teething problems experienced in our preliminary study have been rectified. Bearing in mind the new improved system now has a computer involved at every stage of the process, the potential for disaster (given our track record) has probably increased tenfold!