Dairy Cattle Rumen Bolus Developments with Special Regard to the Applicable Artificial Intelligence (AI) Methods.

It is a well-known worldwide trend to increase the number of animals on dairy farms and to reduce human labor costs. At the same time, there is a growing need to ensure economical animal husbandry and animal welfare. One way to resolve the two conflicting demands is to continuously monitor the animals. In this article, rumen bolus sensor techniques are reviewed, as they can provide lifelong monitoring due to their implementation. The applied sensory modalities are reviewed also using data transmission and data-processing techniques. During the processing of the literature, we have given priority to artificial intelligence methods, the application of which can represent a significant development in this field. Recommendations are also given regarding the applicable hardware and data analysis technologies. Data processing is executed on at least four levels from measurement to integrated analysis. We concluded that significant results can be achieved in this field only if the modern tools of computer science and intelligent data analysis are used at all levels.

A TECHNIQUE FOR DEPLOYMENT OF RUMEN BOLUS TRANSMITTERS IN FREE-RANGING MOOSE ( ALCES ALCES).

Recent uses for rumen boluses, such as mortality implant transmitters (MITs), in wildlife have made it necessary to adapt deployment techniques developed for livestock. In 29 and 30 attempts to place MITs in Minnesota free-ranging moose ( Alces alces) in 2013 and 2014, respectively, success was achieved 83% and 63% of the time. In 2014, new methods for MIT deployment were evaluated in captive moose in Alaska. Mandible measurements provided guidance for selection of an appropriate-sized bolus applicator. A Schulze mouth gag was used to aid insertion of the applicator, and canola oil was used to lubricate the bolus to facilitate swallowing. Time to first swallow and time to continuous swallow following sedative reversal was measured to gauge appropriate timing for bolus administration. Using the adapted technique with trained personnel, success rates for MIT deployment were 100% (10/10) for captive moose and 88% (21/24) for free-ranging moose in Minnesota in 2015.

Implementing electronic identification for performance recording in sheep: I. Manual versus semiautomatic and automatic recording systems in dairy and meat farms.

With the aim of assessing the secondary benefits of using electronic identification (e-ID) in sheep farms, we compared the use of manual (M), semiautomatic (SA), and automatic (AU) data-collection systems for performance recording (i.e., milk, lambing, and weight) in 3 experiments. Ewes were identified with visual ear tags and electronic rumen boluses. The M system consisted of visual ear tags, on-paper data recording, and manual data uploading to a computer; the use of a personal digital assistant (PDA) for data recording and data uploading was also done in M. The SA system used a handheld reader (HHR) for e-ID, data recording, and uploading. Both PDA and HHR used Bluetooth for uploading. The AU system was only used for body weight recording and consisted of e-ID, data recording in an electronic scale, and data uploading. In experiment 1, M and SA milk-recording systems were compared in a flock of 48 dairy ewes. Ewes were milked once- (×1, n=24) or twice- (×2, n=24) daily in a 2 × 12 milking parlor and processed in groups of 24. Milk yield (1.21 ± 0.04 L/d, on average) was 36% lower in ×1 than ×2 ewes and milk recording time correlated positively with milk yield (R(2)=0.71). Data transfer was markedly faster for PDA and HHR than for M. As a result, overall milk recording time was faster in SA (×1=12.1 ± 0.6 min/24 ewes; ×2=22.1 ± 0.9 min/24 ewes) than M (×1=14.9 ± 0.6 min/24 ewes; ×2=27.9 ± 1.0 min/24 ewes). No differences between PDA and HHR were detected. Time savings, with regard to M, were greater for ×2 than for ×1 (5.6 ± 0.2 vs. 2.8 ± 0.1 min per 24 ewes, respectively), but similar for PDA and HHR. Data transfer errors averaged 3.6% in M, whereas no errors were found in either SA system. In experiment 2, 73 dairy and 80 meat ewes were monitored at lambing using M and SA. Overall time for lambing recording was greater in M than SA in dairy (1.67 ± 0.06 vs. 0.87 ± 0.04 min/ewe) and meat (1.30 ± 0.03 vs. 0.73 ± 0.03 min/ewe) ewes. Recording errors were greater in dairy (9.6%) than in meat (1.9%) ewes. Data uploading errors only occurred in M (4.9%). In experiment 3, 120 dairy and 120 meat ewes were weighed using M and AU systems. In both flocks, mean BW recording and data uploading times, as well as overall BW recording time (0.63 ± 0.02 and 0.25 ± 0.01 min/ewe, respectively) were greater in M than in AU, and uploading errors only occurred in M (8.8%). In conclusion, HHR and PDA systems were time-effective for performance recording, both saving time and improving data accuracy. Working load and time for ewe identification were faster in HHR but it did not affect the performance recording time. The PDA was the fastest device for data download. Further research will evaluate the costs of implementing e-ID for performance recording and other uses in sheep farms.